Various Stats on Various Raid Bosses

[Vanshilar]

This is a collection of experimentally determined stats on various raid bosses. This is not meant to be a compendium (i.e. complete set), since there are some that I can't access and some that there isn't that much interest for (to me, anyway). Rather than fill the missing stats with speculation, I will only put stats that have been experimentally tested, although the caveat is that if the tested amount is within a range, I may suggest the most likely number within that range.

I will be updating this periodically as I get more test data, so each stat will have a date next to it (in format month/date/year) that represents the latest time data was collected and analyzed on it -- so I would recommending checking periodically to see if there are updated stats. Hopefully this will serve as a marker for whether or not each data point may be obsolete.

This post will focus on the stats themselves. The second post will describe the testing methodology used to determine CR, fortification %, DR, and spells and spell DCs. The third post will describe the testing methodology used to determine the HP for each boss.

When using the stats, please be mindful of how they are described, which is at the bottom of this post. In other words, what the format of each stat type is; for example, how the 95% confidence interval is displayed for fort % has a different meaning than the standard deviation +- of the attack.

Feel free to make any suggestions as to how this can be made more readable, since I can arrange this in pretty much any format as desired.

Arraetrikos (raid boss of "The Thirteenth Eclipse", commonly "Shroud")
Note that I will be assuming that the monster in part 4 is more or less the same as the one in part 5, with exceptions (such as HP) noting otherwise.
CR
Normal: 28 (5/04/2010)
Hard: 32 (3/18/2010)
Elite: 38 (3/25/2010)

DR
Normal: 15 (15-16) (3/12/2010)
Hard: TBD
Elite: 35 (34-35) (3/25/2010)

Fort %
Normal: 47.6% (38.2% - 57.1%) (49 resists/103 attempts) (4/30/2010)
Hard: TBD
Elite (part 4): 55.4% (42.4% - 67.6%) (31 resists/56 attempts) (3/25/2010)
Elite (part 5): 51.2% (40.6% - 61.7%) (42 resists/82 attempts) (3/25/2010)

HP (part 4)
Normal: 100k (99435 to 100491) (5/08/2010)
Hard: 120k-125k (118548 to 127723) (6/08/2010) (Thanks to Sirgog for figuring this out!)
Elite: 220k (220126 to 221777) (6/24/2010)

HP (part 5)
Normal: 150k (149184 to 151218) (5/11/2010)
Hard: TBD
Elite: 330k (328856 to 333641) (6/24/2010)

Attack
Normal: 48.6 +- 6.6 (78 attacks) (3/12/2010)
Hard: TBD
Elite: 65.4 +- 10.2 (29 attacks) (3/25/2010)

Spells
Delayed blast fireball:
200-300 (250 avg) fire damage on all difficulties, damage halved on reflex save, decreased by fire resist
Spell DC: 30 on Normal, 34 (34-39) on Elite (3/25/2010)

Meteor Swarm: 4 meteors, each consisting of:
Bludgeon component: 20-30 (25 avg) bludgeon damage on all difficulties, no save, unaffected by DR
Fire component: 60-90 (75 avg) fire damage on all difficulties, damage halved on reflex save, decreased by fire resist
Spell DC: 32 on Normal, 36 on Elite (3/12/2010)


Suulomades (raid boss of "A Vision of Destruction", commonly "VoD")
Note that for all intents and purposes, the Suulomades that shows up in part 3 of ToD should be considered a different monster, and its stats may be different than the ones given here.
CR
Normal: 28 (5/11/2010)
Hard: 32 (6/27/2010)
Elite: 38 (5/17/2010)

DR
Normal: TBD
Hard: TBD
Elite: 35 (35-35) (5/17/2010)

Fort %
Normal: 42.5% (31.8% - 53.9%) (31 resists/73 attempts) (5/11/2010)
Hard: TBD
Elite: 52.3% (40.4% - 64.0%) (34 resists/65 attempts) (5/17/2010)

HP
Normal: 150k (148148 to 152143) (5/11/2010)
Hard: 250k (249035 to 254658) (6/27/2010)
Elite: 350k (349798 to 356581) (5/25/2010)

Attack
Normal: TBD
Hard: TBD
Elite: TBD

Spells
Not yet analyzed, but I want to mention that it looks like Suulomades (unlike Arraetrikos or Horoth) does not have maximize/empower, meaning that while Arraetrikos does around 250 damage with delayed blast fireball, Suulomades only does around 100 damage.


The Jailer (sub-boss of part 1 of "The Devil you Know", commonly "ToD")
CR
Normal: 26 (4/27/2010)
Hard: 30 (5/04/2010)
Elite: 36 (5/16/2010)

DR
Normal: 20 (20-20) (5/30/2010)
Hard: 20 (20-21) (5/04/2010)
Elite: 20 (19-20) (5/12/2010)

Fort %
Normal: 0% (0.0% - 9.4%) (0 resists/37 attempts) (4/27/2010)
Hard: TBD
Elite: TBD

HP
Normal: 90k (90280 to 90406 across 4 measurements) (5/30/2010) Please see note in HP post below about the actual measurements
Hard: 120k (117187 to 121833) (5/30/2010)
Elite: 158k (157472 to 158655 across 2 measurements) (6/5/2010)

Attack
Normal: TBD
Hard: TBD
Elite: TBD

Spells
TBD


The Judge (sub-boss of part 1 of "The Devil you Know", commonly "ToD")
CR
Normal: 26 (4/27/2010)
Hard: 30 (5/04/2010)
Elite: 36 (5/16/2010)

DR
Normal: 15 (13-15) (4/27/2010)
Hard: 15 (14-16) (5/04/2010)
Elite: 20 (20-20) (5/16/2010)

Fort %
Normal: TBD
Hard: 56.0% (37.1% - 73.3%) (14 resists/25 attempts) (5/01/2010)
Elite: 44.4% (27.6% - 62.7%) (12 resists/27 attempts) (5/16/2010)

HP
Normal: 90k (90058 to 91611 across 3 measurements) (5/30/2010)
Hard: 120k (119806 to 121359 across 2 measurements) (6/5/2010)
Elite: 160k (160210 to 160653 across 2 measurements) (6/5/2010)

Attack
Normal: TBD
Hard: TBD
Elite: TBD

Spells
TBD


Nythirios (sub-boss of part 2 of "The Devil you Know", commonly "ToD")
CR
Normal: 29 (4/27/2010)
Hard: 33 (5/4/2010)
Elite: 39 (5/16/2010)

DR
Normal: 15 (14-15) (4/27/2010)
Hard: 15 (14-16) (5/04/2010)
Elite: 20 (20-20) (5/12/2010)

Fort %
Normal: TBD
Hard: TBD
Elite: TBD

HP
Normal: TBD
Hard: 160k (159396 to 160505) (7/14/2011)
Elite: 220k (219287 to 221259) (9/19/2010)

Attack
Normal: TBD
Hard: TBD
Elite: TBD

Spells
TBD

Suulomades (sub-boss of part 3 of "The Devil you Know", commonly "ToD")
Note that for all intents and purposes, the Suulomades that shows up in VoD should be considered a different monster, and its stats may be different than the ones given here.
CR
Normal: TBD
Hard: TBD
Elite: 38 (11/7/2010)

DR
Normal: TBD
Hard: TBD
Elite: TBD

Fort %
Normal: TBD
Hard: TBD
Elite: TBD

HP
Normal: TBD
Hard: TBD
Elite: 350k (349182 to 351373) (11/7/2010)

Attack
Normal: TBD
Hard: TBD
Elite: TBD

Spells
TBD

Horoth (raid boss of part 3 of "The Devil you Know", commonly "ToD")
CR
Normal: 30 (4/27/2010)
Hard: 34 (5/04/2010)
Elite: 40 (5/12/2010)

DR
Normal: 15 (14-15) (4/27/2010)
Hard: 25 (24-25) (5/04/2010)
Elite: 35 (35-35) (5/12/2010)

Fort %
Normal: 44.2% (37.8% - 50.9%) (96 resists/217 attempts) (4/27/2010)
Hard: 49.7% (42.0% - 57.4%) (79 resists/159 attempts) (5/01/2010)
Elite: 50.6% (43.2% - 58.0%) (87 resists/172 attempts) (5/12/2010)

HP
Normal: 200k (198542 to 200539) (4/21/2010)
Hard: 251k (250501 to 251001) (7/14/2011)
Elite: 350k (349428 to 353867) (9/19/2010)

Attack
Normal: TBD
Hard: TBD
Elite: 104.8 +- 11.5 (350 attacks) (5/12/2010)

Spells
Chain Lightning:
Unknown electrical damage
Spell DC: 29 or higher on Elite (5/12/2010)

Delayed blast fireball:
200-300 (250 avg) fire damage (elite), damage halved on reflex save, decreased by fire resist
Spell DC: 28 or higher on Elite (5/12/2010)

Disintegrate:
Unknown untyped? damage
Spell DC: 44 or lower on Elite (5/12/2010)

Meteor Swarm: 4 meteors, each consisting of:
Bludgeon component: 20-30 (25 avg) bludgeon damage (elite), no save, unaffected by DR
Fire component: 60-90 (75 avg) fire damage (elite), damage halved on reflex save, decreased by fire resist
Spell DC: 30-36 on Elite (5/12/2010)

Shout:
Unknown sonic damage
Spell DC: 48 or lower on Elite (5/12/2010)

Sunburst:
Unknown light damage
Spell DC: 24-33 on Elite (5/12/2010)

Other Effects
Badge of Death:
38.1 +- 8.7 damage every 3-5 seconds (timing is somewhat random, seems to be either 3, 4, or 5 seconds, possibly a few 2 seconds or 6 seconds as well) (35 hits)

Honor of Fools:
39.3 +- 8.9 damage every 3-5 seconds (timing is somewhat random, seems to be either 3, 4, or 5 seconds, possibly a few 2 seconds or 6 seconds as well) (75 hits)

Medal of Cruel Pain:
37.9 +- 7.5 damage every 3-5 seconds (timing is somewhat random, seems to be either 3, 4, or 5 seconds, possibly a few 2 seconds or 6 seconds as well) (60 hits)

The statistics of these suggest that each do 6d12 damage every 2d3 seconds, which would be 39.0 +- 8.5 damage every 4 seconds.

HP gain on player death:
During part 3 of the raid, whenever a player dies, Horoth gains a certain amount of HP. The percentage is:
Normal: 22%
Hard: 22%
Elite: 19%

Queen Lailat (raid boss of "Zawabi's Revenge", commonly "DQ")
CR
Normal: 22 (6/20/2010)
Hard: 24 (6/20/2010)
Elite: 28 (6/20/2010)
Epic: 44 (6/20/2010)

DR
Normal: TBD
Hard: TBD
Elite: TBD
Epic: 20 (19-20) (5/07/2010)

Fort %
Normal: TBD
Hard: TBD
Elite: TBD
Epic: 0% (0.0% - 8.8%) (0 resists/40 attempts) (6/20/2010)

HP
Normal: 11k (10064 to 11099 across 2 measurements) (6/20/2010)
Hard: 15k (15244 to 15490) (6/20/2010)
Elite: 20.5k (20276 to 20645) (6/20/2010)
Epic: 97k (95756 to 97901) (9/10/2010)

AC
Normal: TBD
Hard: TBD
Elite: TBD
Epic: 66 (66-66) (9/10/2010)

Attack
Normal: TBD
Hard: TBD
Elite: TBD
Epic: TBD

Spells
TBD


Black Abbot (raid boss of "Ascension Chamber", commonly "Abbot")
CR
Normal: 24 (7/14/2011)
Hard: 24 (7/14/2011)
Elite: 24 (7/14/2011)

DR
Normal: 20 (20-21) (6/05/2010)
Hard: 20 (7/14/2011)
Elite: 20 (19-20) (6/05/2010)

Fort %
Normal: TBD (presumably 100% since undead)
Hard: TBD (presumably 100% since undead)
Elite: TBD (presumably 100% since undead)

HP
Normal: 135k (132090 to 136061) (7/14/2011)
Hard: 135k (134828 to 136011) (7/14/2011)
Elite: 135k (133570 to 135863) (7/14/2011)
Note: Tested on 7/14/2011 on elite that the total HP is unchanged when he respawns after first round of puzzles.


Attack
Normal: TBD
Hard: TBD
Elite: TBD

Spells
TBD



CR: The Challenge Rating of the boss, a rough measure of how tough the boss is.
DR: The likely Damage Reduction that the boss has (how much less damage he takes per successful hit on him if the DR isn't bypassed); numbers following in parentheses are the possible range of DR based on experimental testing.
Fort %: The tested fortification % of the boss (the chance he has to resist critical hits), followed by the 95% confidence interval (Wilson score method) that the fort% is likely to be. This means that based on the data, there is a 95% chance that the actual fort% is within the given range. The last set of numbers is the actual data of how many times the boss resisted a crit and how many crit attempts were recorded.
HP: The likely Hit Points that the boss has (the total amount of damage needed to kill the boss); numbers following in parentheses are the possible range of HP based on experimental testing.
AC: The armor class of the boss. Your modified attack roll (1d20 + BAB + other attack modifiers) must be equal or greater than this to hit the boss.
Attack: The average damage that the boss does on each physical attack. The second number is the standard deviation of the average damage. Note that the boss can readily do damage above the average + standard deviation, since one standard deviation only encapsulates 68% of the attacks. The number following this is the total number of attacks in the sample.
Spells: Each spell is listed individually. The DC of a spell is the minimum of a 1d20 roll + your save modifier (such as your reflex save). A roll of 1 is automatically a save fail and a roll of 20 is automatically a save success. To fail only when you roll a 1, your saves versus that spell should be 2 less than the DC or higher.

[Vanshilar]

This is how I experimentally determined CR, DR, fort%, and spells and spell DC of the various bosses. The stats are not determined in realtime, but I video capture some of the raids I go on with the combat log showing, then go over the combat log in detail (frame by frame) later, and record down the stuff of interest into Excel. My capture setup isn't particularly ideal; it's using the freeware Camstudio recording at 10 FPS, but has an issue with repeated frames so I more or less only get 5 FPS performance out of that. It's enough though to see the combat log and not miss any data most of the time which is suitable enough for my purposes. A screenshot of my current setup is here:

http://i898.photobucket.com/albums/ac182/Vanshilar/ToD_end.jpg

CR
The challenge rating of the boss. The higher it is, the harder the monster is supposed to be. Determining this is simple, just target the boss, and the CR is given in the focus orb.

DR
The damage reduction of the boss. This is how much less damage the monster takes per hit if the DR is not bypassed. I use two main ways of determining the DR:
1. Do damage for a while using a weapon that bypasses the DR, and then using a weapon that doesn't bypass the DR. The damage range (minimum and maximum of the damage numbers) in each case can be determined, by looking at the differences of the ranges. For example, if when using a +3 greatsword (2d6) of bypassing DR I get 60-70 damage per hit, while using a +2 falchion (2d4) of not bypassing DR, I get 39-47 damage, then I would get that the DR is 20. The main disadvantages are the number of hits needed to get the range for each weapon (for 2d6, that's a 1 in 36 chance of hitting the minimum roll of 2, and 1 in 36 chance of hitting the maximum roll of 12), and that during that time, if any buffs fall, I will get inaccurate results if those buffs aren't taken into account. The advantage of this is that if I do get the full range, the DR can be determined exactly.
2. Compare main hit damage with glancing blow damage. Glancing blow damage is a known formula of the main damage. The formula for a BAB 20 character with the THF feats is:

glancing damage = (base damage + 30) * 0.3

where the base damage is the weapon + modifier (str modifier, power attack, etc.) damage before crits, sneaks, etc. are considered. Using a weapon that does not bypass DR, I can determine the damage range of the weapon, and compare that with the damage range of the glancing blows, and work out what possible values of DR (using the above formula) they satisfy. The disadvantages of this method are that sometimes there is a range of possible DR values, and this method only applies when the DR is less than the glancing blow damage (about 30 or so). The advantages are that I get more measurements quickly, and I am guaranteed that they are under the same conditions with regards to buffs.

When a DR range is given, the actual DR is "guaranteed" to be within the range; I include all numbers that can possibly produce the observed data.

Fort%
This is the chance that the boss has of resisting a critical hit. This is also the same chance of a boss resisting a sneak attack. To determine this, I scan through the combat log for whenever I roll a critical roll. Then, I see if the roll goes through as a critical hit, or if it is resisted. I simply tally up the total number of crit attempts and successful resists.

Whether or not a crit is resisted is a binomial event; either it happens or it doesn't. However, the measurement is of something that's probabilistic. Because of this, I use a 95% confidence interval to describe the range of probable values. Rather than the more standard interval for a normal distribution, I use the Wilson score interval (which you can see a description of here: http://en.wikipedia.org/wiki/Wilson_score_interval) which is supposed to give more accurate results for smaller sample sizes and/or when the probability is far away from 50% (such as, for example, lightning strike proc rate). I'm not a statistician, so I don't know why, but I'll take them at their word that it's better than the standard one. It's all calculated via Excel anyway so as long as I got the formula right it's not a problem to use.

The 95% confidence interval is the range which, based on the data, 95% of the time the actual value will fall within the range given. However, there is still a 1 in 20 chance that the measured value will be incorrect. This is as often as...well, as rolling a 1 on a 1d20. Put another way, if I give 20 different stats each with their own 95% confidence intervals, there is a good chance (around 64%) that at least one of the stats actually lie outside the 95% confidence interval for that stat. So the 95% confidence interval is not absolute; it's possible that the actual value is outside of the given range, but the range gives an idea of the likely values that the value could be based on the collected data. Similarly, for monster which may have 0% fort, it's not so much that I can experimentally determine that they have 0% fort (I can always be unlucky), but that I can be 95% certain that their fort% is below a given amount. For example, the Jailer in the first part of ToD definitely has a certain amount of fortification:

http://i898.photobucket.com/albums/ac182/Vanshilar/ToD_test.jpg

However, it didn't show up in the videos that I've tabulated to look at fort% data (and hence the data does not have any successful crit resists). Here, the Wilson method demonstrates its advantage over the regular normal distribution's confidence interval, since the latter cannot deal with data that are all "misses". Using the Wilson score interval, however, a certain amount of possible fort% can still be estimated. For those of you with lightning 2 green steel weapons, you can see how this plays out; going for say 100 swings without a lightning strike does not mean that the strike doesn't come out at all, just that it has a pretty low chance. The normal distribution's confidence interval formula would be unable to deal with this, but the Wilson method would give that this just means lightning strike has a 0 to 3.7% chance of occurring (95% confidence, although this is possibly 97.5% confidence since the success rate can't be less than 0%) -- which fits with established estimates of its proc rate.

Spells and Spell DCs
Like collecting fort% data, there's nothing particularly fancy here, mostly just tabulating the numbers that the combat log gives. For spell DCs, I assume that they are constant across different raids (except varying by difficulty setting of the raid), i.e. the boss does not have an (occasional) buff that boosts his DCs, nor do player debuffs that decrease the boss spell DCs work on the monsters listed. This assumption basically means that I can tabulate the spell DCs across multiple raid videos done on different days and just look for the minimum save and maximum fail across all videos. I look at the numbers at which I succeed on my save and the numbers at which I fail on my save. The DC is the minimum number at which I save. For example, for Arraetrikos, the spell DC for meteor swarm is 32. This is because on 3/12, I recorded saving on a roll of 13 with a modifier of +19 (so total 32), while on 3/3, I recorded failing on a roll of 6 with a modifier of +25 (so total 31). Note that you automatically fail a save on a roll of 1, and you automatically succeed on a roll of 20. Thus, if you're looking to fail only on a roll of 1, you should have at least 2 less than the spell DC. So for a spell DC of 32, you want to have a save of 30 or higher.

[Vanshilar]

This post is to describe how I determine the total HP of each boss. I thought I'd put it as its own post since the other stats are more readily described (but just simply takes some time to go through the combat log from each raid) while this one will take a bit of explaining as to the testing methodology.

The determination essentially uses two elements: the total damage done to the boss at any given time, and how much of the HP bar in the focus orb has been taken off up to that time. The basic idea is that when the total damage dealt is known, as well as how much of the HP bar has been removed, the total HP is equal to the damage dealt divided by the percentage of the HP bar that's been removed. For a simple example, say I've done a total of 1000 damage to a boss. If the HP bar is at 90%, meaning I've taken off 10% of his total health, then I know that total HP * 10% = 1000. Multiplying both sides by 10, I will find that total HP = 10000. Simple enough.

Although I think this form of HP estimation has been tried before, to various amounts of success, I also use another measurement to further enhance the accuracy: I look at the total damage done right before a pixel is removed, as well as the total damage done right after a pixel is removed, to establish min and max bounds on what the HP can be. As I'll explain below, this is one of the keys to the success of this method.


To do this in DDO, it requires first figuring out how the HP bar works. There are some things that needed to be tested, namely:

1) How many pixels does the HP bar have?
2) Does the HP bar round up or round down when it comes to displaying the remaining HP?
3) Is the HP that each pixel of the HP bar represents constant, or does it change? Basically, does the HP bar scale linearly all the way down to no HP, or does the amount of HP that each pixel represents vary depending on how much HP is left?

To determine the first two, I went scrag hunting with my barbarian, with the focus orb on myself. I had a total of 552 HP unbuffed. I went into frenzy, quaffed a potion to heal myself back to full, then whacked the scrag repeatedly. Because vicious damage is only 1 per hit due to dungeon scaling, and I can see what my actual HP is, I can match up my HP to the number of pixels removed as my HP is decreased one by one. The data looked something like this:

HP Pixels removed
552 0
551 1
550 1
549 1
548 2
547 2
546 2
545 2
544 3
543 3
542 3
541 3
540 4
539 4
538 4
537 5
536 5
535 5
534 5
533 6
532 6
531 6
530 6
529 7
528 7
527 7
526 7
525 8
524 8
523 8
522 9

By "pixels removed", what I mean is how many pixels are empty (no longer red) in the focus orb's HP bar.

That the first point of damage removed a pixel means that the HP bar is displayed "rounded down", that is, it'll remove a pixel when any fraction of that pixel's worth has been removed. So even if a boss has 1000000 HP, if you do 1 damage to it, the first pixel will be removed. This explains why sometimes the focus orb will show the boss already has no health, yet it is still alive -- you've already started working on the very last pixel (thus it disappears), but you haven't finished it off yet.

To determine what fraction of the total HP each pixel represents, consider that now knowing that it rounds down, the number of pixels removed would be determined by:

Pixels removed = roundup[ (total HP - damage done) / total HP * # pixels of HP bar]

where the # pixels of HP bar is the total number of pixels that the HP bar takes up. Trying different numbers, the only number that fits the data is 148. If it were 147 or less, when my HP was at 537 it would've shown 4 or less pixels removed. If it were 149 or more, when my HP was at 526, it would've shown 8 or more pixels removed. So what this means is that each pixel of the HP bar represents 1/148 of the target's total HP.

By the way, I should mention, I have no idea if this depends on the resolution that you're playing at. This amount may be different for different resolutions.

To double-check this, I buffed up (rage and madstone), for a total of 752 HP, and redid this test. The data was:

HP Pixels removed
752 0
751 1
750 1
749 1
748 1
747 1
746 2
745 2
744 2
743 2
742 2
741 3
740 3
739 3
738 3
737 3
736 4
735 4
734 4
733 4
732 4
731 5
730 5
729 5
728 5
727 5
726 6
725 6
724 6
723 6
722 6
721 7
720 7
719 7
718 7
717 7
716 8
715 8
714 8
713 8
712 8
711 9
710 9
709 9
708 9
707 9
706 10
705 10
704 10
703 10
702 10
701 11
700 11
699 11
698 11
697 11
696 12
695 12
694 12
693 12
692 12
691 13
690 13
689 13
688 13
687 13
686 13
685 14
684 14
683 14
682 14
681 14
680 15
679 15
678 15
677 15
676 15
675 16
674 16
673 16
672 16
671 16
670 17
669 17
668 17
667 17
666 17
665 18
664 18
663 18
662 18
661 18
660 19
659 19
658 19
657 19
656 19
655 20
654 20
653 20
652 20
651 20
650 21
649 21
648 21
647 21
646 21
645 22
644 22
643 22
642 22
641 22
640 23
639 23
638 23
637 23
636 23
635 24
634 24
633 24
632 24
631 24
630 25
629 25
628 25
627 25
626 25
625 25
624 26
623 26
622 26

Again, if you plug this into Excel, you'll see that 148 is the only number that fits the data.


Now I can ask the question: If I were given the combat data (i.e. how much damage frenzy did to myself) as well as at which amount of damage were each pixel removed, would I be able to figure out how much HP I have? This is where things get fun. When the first pixel is removed, I get no usable information really (unless multiple pixels are removed at once), since any amount of damage will remove that pixel. So the first pixel is more or less useless. However, let's look at the first data set (where I know for sure that my HP is 552). After 3 total points of damage dealt, my HP bar is still at only 1 pixel removed. After 4 total points of damage dealt, however, my HP bar changed to be 2 pixels removed. Now, if at 3 points of damage dealt, the HP bar only had 1 pixel removed, this means that I had to have at least 444 HP. If I had 443 HP for example, with 3 damage done to me, I'd have 440/443 HP, and the HP bar would show 3/443 * 148 = 1.002257 pixels removed, which would be rounded up to 2 pixels removed. Thus, I have at least 444 HP. With 4 points of damage dealt, the HP bar had 2 pixels removed. From this, I can deduce that I have at most 591 HP. If I had 592 HP, then the HP bar would show 4/592 * 148 = 1 pixel removed, rather than 2. Thus, from the second pixel, I know I have 444 to 591 HP.

Continuing on with this logic, when a total of 7 damage has been done, there's only 2 pixels removed. When a total of 8 damage is done, however, my HP bar has 3 pixels removed. Doing the same calculations, this means that I have to have between 518 and 591 HP. From the 4th pixel being removed, I would know I have between 543 and 591 HP. From the 5th pixel being removed, I would know I have between 518 and 554 HP.

Now, from the 4th pixel, I knew that I had to have at least 543 HP. From the 5th pixel, I know I have at most 554 HP. This means that the cumulative minimum and maximum possible HP range (as opposed to the HP range from each individual pixel measurement) is 543 HP to 554 HP. Not too shabby when I've only done a total of 15 damage, which is 15/552 = 2.7% of the total HP.

Now the remaining issue (which I implicitly assumed to be true above), is whether or not each pixel represents the same amount of HP, an even 1/148 of the total HP. After all, it's possible that the first few pixels each represent 1/148 of the total HP (which is tested above), while the other pixels each represent say 1/100 or 1/200 of the total HP. Just testing the first few pixels wouldn't really determine if this were true.

So to test this, I decided to do the portal test. That's right, I would punch the Aussircaex Valley portal (punching since it does 1d3 base damage, so very little variation) until it died, sum up the damage and at how much damage did each pixel get removed, and figure out the portal's HP from there. Each punch did 15-17 damage; since I punched the scrags for 25-27 damage, this works out to 10 DR for the portal (I also confirmed the 10 DR by using the glancing blow method described in the previous post). Anyway, at 79 total damage done, the HP bar was still at 1 pixel removed; on the next hit, with 95 total damage done, the HP bar dropped to 2 pixels removed. Thus, from this second pixel, using the above calculations, the portal had to have between 11692 to 14059 HP. Continuing in this way, with 161 total damage done it was still at 2 pixels removed; on the next hit, with 178 damage done, the 3rd pixel was removed. Thus, from the third pixel, the portal had between 11914 and 13171 HP. Next, with 227 total damage done, it was still at 3 pixels removed; with 244 damage done however, the 4th pixel was removed. This implies that the portal has between 11199 and 12037 HP. Cumulatively, based on the 3rd and 4th pixels, the portal had to have between 11914 to 12037 HP.

Rather than bore you with the complete set of data going all the way down to the 149th pixel (of course, the HP bar only has 148 pixels, so on the 149th pixel the portal dies), I made a graph of this:


http://i898.photobucket.com/albums/ac182/Vanshilar/portal3.jpg

and a blown-up version of the area around 12000 HP:

http://i898.photobucket.com/albums/ac182/Vanshilar/portal4.jpg

The x-axis (measurement) is the nth pixel that disappeared, starting from 2 (since you don't get usable information from the 1st pixel) to 149 (at which point the portal dies). The X's represent the maximum amount of HP and the minimum amount of HP from each measurement (each pixel being removed). The line represents the cumulative max and min, that is, the maximum of all the minimum measurements (since the portal has to have at least that much HP) and the minimum of all the maximum measurements (since the portal has at most that much HP). What should be clear is that after a few measurements, the cumulative range is very small -- when the 6th pixel drops (representing 6/148 ~ 4% of the total HP), the portal's HP is determined to be between 11988 and 12024 HP, so the measurement error is basically (12024-11988)/2 / average(12024, 11988) = +- 0.15% measurement error.

By comparison, when I reached the very end, summing up all 100% of the damage done from the combat log, at 11998 damage done the portal was still alive, while on the next hit, for 12015 damage total, it died. If I were doing it this way (summing up 100% of the damage done), my measurement error would be +- 0.07%. To be able to sum up only 4% of the damage done and get a measurement error of just +- 0.15%, compared with doing 25 times more work to sum up 100% of the damage just to reduce the measurement error in half, is a lot simpler.

You can see from the graphs that the cumulative range continues to get smaller as more measurements are made, and eventually converges on the final value of 12000. That's right, at some point it becomes even more accurate than summing up all of the damage done, and does eventually converge to a single value (i.e. determining the HP exactly), in this case on the 135th pixel removed (so when the portal has around 9% of its HP left). Considering that in the second graph, each y-axis gridline represents 20 damage while the last hit (and thus the measurement error if you summed up all of the damage) was 17 damage, you can see that this method is more precise after around the 35th pixel removed, or when around 24% of the total HP has been done. Being able to determine the HP precisely by using this method is what allowed me to say that the portal has exactly 12000 HP here:

http://forums.ddo.com/showpost.php?p=2944559&postcount=106

These figures are for the portal, and the measurement error remaining after n measurements varies from case to case. However, I would expect that the same trend holds, that the measurement error decreases very quickly in the first few measurements, then the solution converges slowly thereafter. What you are looking at is basically a graph of y = 1/x, which should not be too surprising because that's precisely what I am doing, using that total HP = total damage done / (pixels removed / total pixels).

To get back to the original topic (whether each pixel represents the same fraction of the total HP), you can see that each measurement basically agrees with each other. If different pixels represented different fractions of the total health, then the measurement from each pixel should contradict each other at some point, that is, some pixel's minimum HP measurement would be higher than another pixel's maximum HP measurement. Remember the level of precision that we're talking about here -- from the 11th pixel onward, each pixel's measurement error is less than +- 1% of its measured value for the total HP. If the amount of HP that each pixel represented deviated by more than that much, it would've shown up in the graph. So the conclusion is that each pixel does indeed represent the same amount of HP.

Now that I've gone through the method's logic, the remaining thing is just how to collect the data. What I need is the total damage done to the monster right before and right after a pixel is removed. The more pixels with which this is done, the more accurate the measurement is. Now, it's possible to have everyone attack the boss at once, and then piece together everyone's combat log together later, and thus figure it out that way. Obviously, though, it's much easier to have just one person attack. Since the beginning is a known point (i.e. at the beginning of the fight, the boss is for sure at 100% health, so the HP bar is full), it makes sense then to have the one person start off the fight by attacking the monster, without anyone else also attacking.

Let's see...so when is it that you have one person and only one person attacking the boss at the beginning of the fight? That's right, this is precisely what a hate tank does, to be the only one to damage the boss at the beginning of the fight and thus establish the aggro on him. This also perfectly matches the conditions needed to test out the HP of the boss.

Unfortunately, although my main is a capped human pure barbarian (and would otherwise be a very good candidate for a tank due to human healing amp), he doesn't have any incite armor at the moment, and as a gimpy non-min/maxed 28-point barb with just madstones but no min 2 nor epic SoS and with average twitch skills, he's fairly mediocre by endgame standards currently in the raids that he's gone on, not particularly suitable to be the main tank on raids other than normal (and normal only because it's safer for the other guys to throttle their DPS and pull back when needed, or hand-off the aggro to the main tank, etc.). However, hopefully once I win the Dragontouched lottery and get some incite on my armor, I'll be able to tank more and figure out more of the different HP that the bosses have.


Notes to people who might want to use this method:
This method is generally applicable to anyone that wants to determine any monster's HP accurately. However, unless everyone has the same size focus orb regardless of screen real estate and resolution, you should probably test out how much pixels the HP bar takes up for your setup. Usually, a screenshot + MS paint may not be accurate, the more accurate way is to take a known HP amount (i.e. yourself) and do a similar test as above.

Also, note that the combat log may be slightly delayed. However, I've noticed that the dice roll shows up somewhat earlier than the combat log sometimes. So when you're looking for the hit at which a pixel dropped, you should be looking at the dice roll, not necessarily at the combat log (although you will have to use the combat log to see how much damage that hit did).

From experience, you should decide if the monster you want to test is a high-HP monster (like a boss), in which case you are testing basically how many hits does it take for a pixel to go down, or if the monster you want to test is a low-HP monster, in which case you are testing how many pixels does each hit remove. Either way, you should try to avoid the situation where each hit removes about 1 pixel or so. Consider it this way. If it takes you 5-6 hits on average to remove a pixel (so it's a relatively high-HP monster), you know that it's around 5 or 6 hits per pixel removed (not 4 and not 7), so each pixel measurement's error range is around say +- 10% (since the total range is 1 pixel out of 5 so 1/5 = 20%, then halved due to +-). If it takes you 2-3 hits to remove a pixel, however, then each pixel measurement's error range is around +- 25%, and the cumulative measurement will converge slowly. So you should try to either remove a lot of the HP bar at once (like say at least 3% per hit), or do a lot of weak hits where it takes many hits to remove the HP bar by 1 pixel (such as punching). For what it's worth, I prefer using the method where it takes many hits to remove a pixel, since I can tell directly from the video whenever the HP bar shortened by a pixel, while to count the number of pixels removed I need to screenshot the video into MS Paint both before and after the hit to count the number of pixels. But for lower HP monsters, it may be more feasible to do it by removing many pixels at once.

If you're having everyone else in the group wait for you to collect this data, even in elite raids I think you should only need around 20 seconds (or 30 seconds at most) of solo time to get an estimate within a percent or two. Basically, you just need to get past that knee in the error curve, to the point where further measurements converge slowly, which occur after the first few measurements. You should be able to do enough damage in 20-30 seconds to remove 3 pixels or more from the HP bar, which represents at least 2 usable measurements. For normal ToD for example, I got the HP estimate within +- 0.5% in 15 seconds, 3 of which was spent moving and not hitting. (Although your experiences may vary, but that was a particularly good run.) I think the only exception would be something like Epic Velah where, since the dragon's fire breath is fairly timed at around 40 seconds and the dragon has a lot of HP, you might as well as solo the first round (and can have everyone be doing final buffs behind the pillar or something while they're waiting for you until after the first fire breath).

As for what weapon to use, if you're using the many hits to remove one pixel method (i.e. vs raid bosses), one with low crit multipliers (and low chance of crits) is preferable. The reason is that although it lowers your DPS, the measurement error is pretty much based on how much damage you did on the hit that dropped an HP bar pixel, so if you critted on that hit, or had a lightning strike proc, etc., then the measurement errors will be larger. In theory what this means is that a good way to test would be to go all-out DPS for the first 20 seconds or so (so you build up a lot of total damage), then just punch the thing until a pixel goes down, so that the damage done on the hit that removes a pixel is very small compared with your total damage, and hence you get a very accurate measurement. I haven't tried this out yet though because I'm sure it'll raise some eyebrows for my groupmates to see that I'm punching the guy to test his HP. Of course, if you go the all-out DPS route, then you may get more measurements in the same time frame, which may also reduce the cumulative measurement range, depending on your luck.

Also consider that for monsters, it's possible that they may have somewhat different amounts of HP each time. I'm not talking about dungeon scaling here, what I mean is that for example, the spiders you encounter in Heyton's Rest may have somewhat different amounts of HP each time you go in. Testing the first ToD orthon that you encounter, for example, on normal (soloing), the first time he had 3650 HP while the second time he had 3550 HP. So there is a certain amount of variation in how much HP regular monsters may have.

Even bosses seems to differ somewhat on HP, though not by much. Making repeated measurements on the Jailer in part 1 of ToD normal, I get measurements of 90280 to 92055, 89688 to 90773, 87986 to 90773, and a pretty precise measurement of 90095 to 90406 thanks to a healer from Over Raided who was willing to help out and keep me alive as long as possible. Although those measurements congregate around 90k HP, some of them exclude it from possibility. So either the HP will vary somewhat, or the true number of HP is not a nice round number (for example, the true amount may be something like 90350), or there are inaccuracies in the testing method (for example, the damage that shows up on the HP bar may be out of sync with when an attack is displayed via the dice roll). Although this doesn't make too big of a difference (we will not likely notice nor care if the Jailer has 90000 HP or 90350 HP), it is something to note.

I figure this is one way to test out how dungeon scaling may vary by group composition, since HP is one thing that is known to scale by dungeon scaling. For what it's worth, based on testing the HP of the first orthon you encounter at the beginning of ToD on Elite, there wasn't any difference (within the measurement error) in solo and full group situations, so at least dungeon scaling doesn't exist for elite raids (or more narrowly, at least for HP for ToD elite, I'll have to get around to testing out the other settings sometime).


Thanks go to the folks over at the Over Raided, Epic Fail, and The Second Coming guilds and assorted PUGgers on Orien for letting me do these tests, even though I'm sure (for example) that having everyone play dodgeball in Shroud part 5 while I chase after the boss for 30 seconds was probably not what they expected of a Shroud run.

[Rasczak]

Are you going to be putting in Harry's Hold Monster, poison effects and disease effects as well?

Will have to double check on Suulo as he has ennervate, lightning and the chain effects but I don't know the stats for it.

[sirgog]

For Fortification: Can you please rename 'Fort' to 'Fortification' to make it unambiguous that you aren't referring to Fortitude Save? I was perplexed for a while there.

For HP: Take the following group into a raid: One melee that does not twitch fight at all, and that wields an Epic SoS, a bard, and (up to) ten healers. 'Hero' the boss. Have the melee count their (unresisted) critical hits, and the healers count the number of swings (possibly by using time - here not twitching makes things easier). Don't use burst damage increasing effects like Madstone Boots.

You'll need to be certain to minimize the impact of any boss self-healing mechanisms, such as the Gnolls in Shroud part 4. A cleric/FvS heavy group could do this with Implosion.

A melee like that should be able to deal a predictable amount of damage per non-crit and per crit. This data will enable you to quite accurately estimate boss HP.


Oh and on spell DC - Higher level spells have higher DCs, unless the boss uses Heighten Spell.

[Vanshilar]

Are you going to be putting in Harry's Hold Monster, poison effects and disease effects as well?

Will have to double check on Suulo as he has ennervate, lightning and the chain effects but I don't know the stats for it.

For the most part (testing HP being the major exception), I don't go out of my way to test stats during raids, since, you know, there's 11 other people that are affected if one of the melees die. I don't think it'll fly really well if I tell the healers "so, I will be using 0% fort for this raid so that I can see how hard he crits for" etc. I mean the 11 other people are still looking to complete the raid. So most likely, the spell DCs will be stuff that's testable when fully buffed, such as damage spells. If the combat log simply says you're immune without giving a saves roll on it, then I won't be able to test it for the most part.


For Fortification: Can you please rename 'Fort' to 'Fortification' to make it unambiguous that you aren't referring to Fortitude Save? I was perplexed for a while there.

For HP: Take the following group into a raid: One melee that does not twitch fight at all, and that wields an Epic SoS, a bard, and (up to) ten healers. 'Hero' the boss. Have the melee count their (unresisted) critical hits, and the healers count the number of swings (possibly by using time - here not twitching makes things easier). Don't use burst damage increasing effects like Madstone Boots.

You'll need to be certain to minimize the impact of any boss self-healing mechanisms, such as the Gnolls in Shroud part 4. A cleric/FvS heavy group could do this with Implosion.

A melee like that should be able to deal a predictable amount of damage per non-crit and per crit. This data will enable you to quite accurately estimate boss HP.


Oh and on spell DC - Higher level spells have higher DCs, unless the boss uses Heighten Spell.

Okay I updated the fortification thing, adding a description at the bottom. Changed it to "fort %" though, hopefully it's obviously that that's fortification since fortitude save does not have a percentage.

I already have an HP testing system, I basically only need around 20-30 seconds to determine the HP to within a percent or two, I just need to write it up.

My question on spell DC is whether or not there are ways that it can change from fight to fight. For example, if the casters can do touch of idiocy or something and lower the caster level of the boss (and hence the spell DC). Currently I'm assuming that it's always the same (though scaling by difficulty), but I don't know if that's really the case.

[Rasczak]

For the most part (testing HP being the major exception), I don't go out of my way to test stats during raids, since, you know, there's 11 other people that are affected if one of the melees die. I don't think it'll fly really well if I tell the healers "so, I will be using 0% fort for this raid so that I can see how hard he crits for" etc. I mean the 11 other people are still looking to complete the raid. So most likely, the spell DCs will be stuff that's testable when fully buffed, such as damage spells. If the combat log simply says you're immune without giving a saves roll on it, then I won't be able to test it for the most part.



Sorry you misunderstood. I meant under Harry spells and effects there needs to be those 3 things as it is important for someone reading this to know a disease item is important as well as FoM and Poison proof for part 4 & 5. Also in VoD that DW is needed to avoid the neg levels ;) I obviously don't expect you to wipe a party for science :p

[Vanshilar]

Sorry you misunderstood. I meant under Harry spells and effects there needs to be those 3 things as it is important for someone reading this to know a disease item is important as well as FoM and Poison proof for part 4 & 5. Also in VoD that DW is needed to avoid the neg levels ;) I obviously don't expect you to wipe a party for science :p

Well currently I'm intending it to be just regarding the stuff I've been able to test. Those things are more for strategy guides (i.e. what buffs you should have and why), so I don't have plans to put them in this thread right now, maybe in the future.

Oh yeah, posted how to determine monster HP, so with those I've also updated the stat post regarding the HP of the bosses tested so far.

[sirgog]

Love this thread.

A couple of further things to consider testing:

1) Reflex Save and presence/lack of Evasion
This is important for casters to have an idea of, as certain spells are only efficient against low Reflex save foes.
Suggested test method: Turn off all Metamagics on a Sorc. Spam a single damage spell and count successes and failures of the boss's saves. (A similar test could be done for Fort saves with Disintegrate).
This will, among other things, tell us whether Polar Ray or Chain Lightning is superior against the Pit Fiend for a range of save DCs (I suspect the latter is only better once save DCs are in the low 40s, and I'm not sure that's possible at the moment).

2) Melee attack extremes of damage
A boss that hits for an average 44 damage per hit might be hitting for 23+2d20, or 30+4d6, or 42+1d3, or something else. In the first case, wilder damage spikes will occur to players, making higher HP more important.

This is harder to test, and would probably be best done by writing down all incoming damage numbers for several hundred attacks, and analyzing the outlying points.

3) Attack bonus (before debuffs)
Easy to test. Take a toon that has a very very high potential AC and wears armored bracers. Unequip bracers and +1, +2, ..., +8 ones in your inventory. Keep upping your AC 1 point at a time until you start seeing grazing hits occuring.


4) Major attack cooldowns
We know Velah's breath is on a 30 second cooldown and Horoth's Blasphemy is on a 60 second one. What about Harry's DBF, Meteor Swarm and Mass Hold Monster?

[Bronko]

This is a collection of experimentally determined stats on various raid bosses.

Well done sir. +1 rep for you.

[Khelden]

Will you add Velah's stats for Normal to Epic? It would be really great :)

[sirgog]

Will you add Velah's stats for Normal to Epic? It would be really great :)

Velah - Fortification - 0 (All difficulties).

Not sure how he'd go about calculating Velah's stats - she has significant regeneration, which plays havoc with HP calcluations. One possible way is to get a melee that can survive her fire breath (triple Fire absorption weapon) to beat on her for two minutes straight without using any twitching or short term buffs. The breath can be survived in this way, even on Epic.

The amount of damage that melee has done would be estimated by setting them against the Stormreaver with a greensteel blank of the same weapon type. From the Reaver's change in HP, you can determine the melee toon's DPS, and thus the amount of damage that has been done to Velah, and from how long it takes her to recover to full HP, you can calculate Velah's regeneration.

Alternately, you could get a less accurate (but easier to obtain) estimate by having a Sorc unload spells on her until they have done a lot of damage (perhaps 25000 hp worth), then stopping until she is at full HP again. This requires quick calculations, but doesn't require wasting a GS weapon and three large scales.

A couple more stats (these were once accurate but may have changed)

Arraetrikos: AC 39/45/51
Lailat: SR 44 (Epic), AC 65 (Epic)

[MrCow]

My question on spell DC is whether or not there are ways that it can change from fight to fight. For example, if the casters can do touch of idiocy or something and lower the caster level of the boss (and hence the spell DC). Currently I'm assuming that it's always the same (though scaling by difficulty), but I don't know if that's really the case.

The answer to this is "it depends".

On all standard spells a player can cast, yes, the critter's DC will change if they have a gain or loss in the stat that they are casting spells from (so things like the Avatar of Cunning in Offering of Blood casting Fox's Cunning will net a +2 DC to his spellcasting).

On non-standard spells it may or may not rise or fall. For instance, lowering WIS on an Ice Flenser will cause it to have a lower DC on Cone of Cold, but not on Field of Ice Razors. Side note, Ice Flensers are very odd, as they cast Arcane Spells (often INT) as Spell-Like Abilities (often CHA), but from WIS. :)

On abilities more often than not a monster will not have an adjusted DC. Things like a monster's Overrun, Trip, Mummy Despair, Minor Wounds Aura, and so forth are static. I think I filed that bug over two and a half years ago. It has yet to be corrected.

[sirgog]

On the assumption that Harry is the same monster in parts 4 and 5:

I'm not sure about this. His AI is clearly different (he focuses much more on melee in part 5 than in part 4), he has at least one new ability in part 5 (the 'teleport someone to my feet' ability), and the graphics of him alternating between normal and double size indicate that he's under the effects of an Enlarge Monster, Righteous Might or some similar spell.

I wouldn't be surprised if he has +4 or even +8 Str in part 5, which would grant extra melee damage. The best way to test would be to record extreme values of Harry's melee attack damage in both parts to try to determine exactly what his damage formula is. (I'm guessing on Normal he deals 29+3d12 damage per attack, it's the most likely fit given that you've measured a mean of ~48.5 and a standard deviation of 6.6). 21+5d10 and 17+7d8 are also plausible given your observed standard deviation, but feel like unlikely numbers of dice.

After testing which of these formulae fit, you could then see if you get any hits in part 5 that just don't fit the formula at all.

[Vanshilar]

Currently I only have a melee character, so my range of stuff to test is limited by that. So I probably won't be able to test reflex/evasion for the foreseeable future, since it looks like that's a caster thing.

Yes the attack damage numbers that I posted are based on copying down how much damage a boss did from the combat log for however many attacks I got hit with, and that's how the standard deviation comes about as well. For an estimate of the upper range of the attack damage, you can add 3 times the standard deviation to the average damage. For Horoth elite for example, I got 104.8 +- 11.5 over a total of 350 attacks. Indeed, the five highest attack damage numbers that I saw across the 350 attacks were 129, 130, 132, 133, 136, which fits the three standard deviations method of 139.3 fairly closely. It's hoped that with the attack damage averages and standard deviations, people can get a fairly good idea of what to prepare for in terms of the physical damage done by a raid boss.

Keep in mind that these bosses can also hit twice per swing. Currently I'd estimate it at around 70% of the time for Horoth elite, but I'm running into some trouble analyzing it completely due to the relatively slow capture rate of my video capture software, meaning that it's somewhat difficult to tell if a given attack is actually just a follow-up of the previous attack, or if it's a new attack (in sampling theory, the problem that I'm having is known as aliasing, due to the sampling frequency being fairly low compared with the thing being measured). Hence I haven't posted about the rate of double attacks yet.

Given that my current main is a barb, it is unlikely that I will be testing any monster's attack bonuses via AC for the foreseeable future. Except maybe those rats in Korthos.

I will eventually be seeing if there's cooldowns on spells, or more specifically, I will be giving the minimum amount of time I see between a cast and the next cast of each spell. But currently tabulating a monster's attacks and spells takes the longest of any of the different stats (several hours per video), compared with any of the other stats (CR = around 15 seconds, HP = around 5 minutes, DR = around 10 minutes, fort% = around 15-20 minutes per video), so it will likely be few and far in between.

I will likely add in Velah's stats for epic at some point. Probably not for the other difficulties, because since I've gotten the pack, I've done nothing but epic for VoN 6. I would like to get an estimate of the HP of course, but currently it looks like it will be a while before they'll let me solo Velah for the first round of the fight.


The answer to this is "it depends".

On all standard spells a player can cast, yes, the critter's DC will change if they have a gain or loss in the stat that they are casting spells from (so things like the Avatar of Cunning in Offering of Blood casting Fox's Cunning will net a +2 DC to his spellcasting).

My question is more of whether or not the spell DC of a raid boss can change between fights, i.e. maybe in this one raid a player can make the DC lower while in another the boss might cast some self-buff that increases the DC. The reason is that I narrow down the range of spell DC's by tabulating my save successes and failures across multiple videos, so quite oftentimes, the results are values gleaned from different runs on the same boss to get (for example) the spell DC for meteor swarm for Arraetrikos. If the DC can change between fights (such as through player debuffs), then I might have to rethink this method of testing. Of course, since I'm mainly focusing on raid bosses, these are purple-named guys, so whether or not a debuff works depends also on whether or not it works on purple-named.

Yes it's likely that Arraetrikos in part 4 and part 5 are two different monsters, the different HP if nothing else makes that clear. However, it's also a matter of whether or not the stats between the two monsters (such as fort% or DR) are different. For the moment I am assuming that they are the same unless I test differently, and noting which stats are different. For what it's worth, I have a separate file for part 4 and part 5 for him because some of the stats are indeed different. Eventually if I get around to testing the different things (such as attack damage) I may split up his listing into two different monsters.

[sirgog]

Bump for the best thread on the forums.

[Xeraphim]

I came here looking for what Elemental Resistances he has, as he seems immune to small amounts of cold and fire, yet electric gets through fine.

[sirgog]

I came here looking for what Elemental Resistances he has, as he seems immune to small amounts of cold and fire, yet electric gets through fine.

Arraetrikos seems to have 30 Cold resistance, immunity to fire, and no electrical resistance.

[sirgog]

I'm going to try to talk some guildies into testing Arraetrikos Elite in the next hour. Primarily testing his hitpoints.

[sirgog]

Got video of me heroing Arraetrikos Hard (not elite as I needed to mostly PUG it) - analyzing HP results now.

Part 4 looks to be 120k or 125k based on fall of second pixel. My combat log was placed poorly, however, so some of the damage numbers (particularly later on) were illegible - I'll try to redo it later without that.

[sirgog]

Going to check my Shroud Hard part 4 and 5 videos to see if I can get you any useful information about how hard Harry hits there.

(Edit - no useful info. Due to Harry's AI, most of the video has him aggroed on someone else, spamming DBF or Meteor Swarm).

[grodon9999]

if there's anyway to figure out their to-hits that would be great as well.

[sirdanile]

Going to check my Shroud Hard part 4 and 5 videos to see if I can get you any useful information about how hard Harry hits there.

(Edit - no useful info. Due to Harry's AI, most of the video has him aggroed on someone else, spamming DBF or Meteor Swarm).

You can potentially fix this by having everyone surround him like normal but not swing, you'll keep aggro most of the time because you're the only one dealing damage and even if he does go after someone else as long as they don't die he should stay put, it would also help you determine how often he mass holds people.

[mediocresurgeon]

I'm impressed. A+ and +rep for your effort and thoughts on this! I'd love to see this information added to the DDOwiki.

I'm interested in finding useful attack speed data for raid bosses so that I can get an accurate estimation of how much DPS my Guard effects add. Finding raid boss elemental resistances would also be a boon. What strategy would you employ to find this information?

[Varthalos]

Do we know the ACs of all raid bosses or any raid bosses, i've been looking for this info all over?

[QuintonReece]

Without regard to accuracy(I cant comment on that) which may be fine..... that must have taken a lot of time and effort. I commend you.

[sirgog]

Do we know the ACs of all raid bosses or any raid bosses, i've been looking for this info all over?

Harry 39/45/51 (this is from ages ago, might have changed)
Lailat Epic - 65

[mediocresurgeon]

(Combat): You roll to confirm a critical on Nythirios. You roll a 5 (+57): Critical hit resisted!

[sirgog]

(Combat): You roll to confirm a critical on Nythirios. You roll a 5 (+57): Critical hit resisted!

That tells us one of two things:

Either Nytharios has an AC of 63 or higher (this is definitely not the case), or he has a non-zero Fortification stat.

We need a lot more testing, however, to close in on whether that Fort is 25%, 50%, 75% or 100%.

[Vanshilar]

Sorry, I've been busy of late so haven't really updated much in the way of boss stats. But a few things:

* Figuring out their to-hit would require a character with a substantial AC. Given that my primary characters are a barbarian and a str-based rogue tempest, I probably won't be able to get the AC to get useful results. So I likely won't be doing their to-hit any time soon.
* For Shroud part 5, the way I did it was actually to ask everyone to run away from him as much as possible and avoid his meteor swarms etc., while I ran after him trying to hit him. When people surround him, it's hard to know conclusively if anyone actually attacked; it's also hard to know if (for example) people had some guard effects that damaged him. It's more conclusive (unless someone was ranging him and deliberately messing up the results) if people were far away from him when you do the testing.
* I do have some attack speed data for some of the bosses. However, I haven't posted them because I'm still working my way through them and trying to figure out what's going on. Basically any time you're the main tank you'll get a bunch of good data (along with for all their other stuff like spells and such). Obviously this is easier for quests like VoD and ToD than Shroud.
* I will be posting some AC data soon. Previously I only had a barbarian that was raid-ready; naturally, a capped barbarian with a raid strength in the 60s is not ideal for testing monster AC. However, now that I also have a rogue tempest that's doing raids, I'll be able to get some AC data. I do have a question though -- other than destruction and improved destruction, are there other things that we can cast on raid bosses that'll affect their AC? Those I can figure out via examining them, since they show up on the effects, but I don't know if there are other things that I'd need to take into account.
* I already have some videos of Nythirios, it's just a matter of going through them and getting enough data that's worth posting. So far though the 95% confidence interval on his fort% is pretty large (i.e. sample size too small, I haven't gone through enough videos yet) but once I go through more videos I'll get to posting them. So far though it looks like he'll also be 50% fort.

[mediocresurgeon]

I believe Exhaustion lowers the Dex of a monster by 6, resulting in a drop of 3 AC. However, this is a calculated value only--I'm not sure that lowering the Dex of a monster affects its AC at all, since I've never tested to see if this was implemented correctly.

Exhaustion also lowers the Str of the monster by 6, which gives a -3 to attack rolls (I can confirm that this is works). Exhaustion also lowers the attack speed of the monster--important to note if you are testing attack speeds (Exhaustion is a common raid debuff on my server).

Stats for Epic Lailat and Epic Velah would be and awesome addition, though you might want to wait until Update 5 to test their +to-hit (rumor has it that these values are changing to make AC more viable on Epic).

[sirgog]

Sorry, I've been busy of late so haven't really updated much in the way of boss stats. But a few things:

* I will be posting some AC data soon. Previously I only had a barbarian that was raid-ready; naturally, a capped barbarian with a raid strength in the 60s is not ideal for testing monster AC. However, now that I also have a rogue tempest that's doing raids, I'll be able to get some AC data. I do have a question though -- other than destruction and improved destruction, are there other things that we can cast on raid bosses that'll affect their AC? Those I can figure out via examining them, since they show up on the effects, but I don't know if there are other things that I'd need to take into account.


Sundering Ooze (or its guard) affects AC, as does Waves of Exhaustion.

[Vanshilar]

Sundering Ooze (or its guard) affects AC, as does Waves of Exhaustion.

Do you know if Sunder will show up on the examine window? Waves of Exhaustion does show up (as "Exhausted"), so I can account for that (and test to see if that does affect AC, once I have more data).

By the way I've taken the server downtime time to update some of the results. Some of the stats for the Black Abbot are now added, along with some stats for Queen Lailat (which were added a week or so ago). Whenever I get around to figuring out how to use manyshot on my tempest rogue, flag him for DQ, and get a group willing to sit around at the beginning of epic DQ, I'll hopefully get some HP data on the marilith.

Also, although I'm not quite sure yet how to put it in, it looks like the shroud portals are all 12000 HP and DR 10, and they do not increase by difficulty. Far as I know, there are 15 portals to destroy, making it a total of 180k HP (at 100% fort and 10 DR, unbypassable I think) in a shroud run, regardless of difficulty. Compare this with Arraetrikos part 4 (100k HP on normal and 220k HP on elite) and part 5 (150k HP on normal and 330k HP on elite) which is a total of 250k HP on normal and 550k HP on elite (at 50% fort and 15 DR or 35 DR if not bypassed).

Edit: I guess the thing to notice (as much as I hesitate to point it out) is that it seems like the Black Abbot's stats don't scale up by difficulty. Although I haven't examined his spells to see if they scale yet.

[MrCow]

Do you know if Sunder will show up on the examine window?

Just about every AC dropping aliment will show up on the examination panel, including Sunder. DEX damage (via things like maladroit and Contagion: Shakes) is among the only thing which won't show.

[sirgog]

Also, although I'm not quite sure yet how to put it in, it looks like the shroud portals are all 12000 HP and DR 10, and they do not increase by difficulty. Far as I know, there are 15 portals to destroy, making it a total of 180k HP (at 100% fort and 10 DR, unbypassable I think) in a shroud run, regardless of difficulty. Compare this with Arraetrikos part 4 (100k HP on normal and 220k HP on elite) and part 5 (150k HP on normal and 330k HP on elite) which is a total of 250k HP on normal and 550k HP on elite (at 50% fort and 15 DR or 35 DR if not bypassed).



On other foes in the Shroud, Anur-Shub (the Fire Elemental) seems to be 10k HP on Normal (a crit Polar Ray for 2600 took off almost exactly 25% of his health bar, although I didn't do a pixel count) and so did Kas'Quik (based on ten ticks of a crit firewall). On Hard, Anur-Shub seems to have 20k, but this is an even less exact approximation.

I'm not going to generalize and guess that they all have 10k/20k HP, however, as some of the foe types represented in the Lieutenants (Earth Elementals and Orthons) traditionally have very high HP compared to other foes of similar CRs.

[sirgog]

Also, that info you have on Shroud Elite makes me question the accuracy of my tests in Shroud Hard. It would surprize me if there's such a tiny difference between Normal and Hard, and an enormous one from Hard to Elite.

[RS-Makk]

This is great stuff. Thanks for doing the work.

Something I find pleasantly surprising is that the %fortification seems to stay about the same as difficulty goes up, and that it seems to be around 50% for most.

[sirgog]

This is great stuff. Thanks for doing the work.

Something I find pleasantly surprising is that the %fortification seems to stay about the same as difficulty goes up, and that it seems to be around 50% for most.

That wasn't always the case, I believe.

I think it started at 50/75/75 on the Devils, but Turbine changed this (when they also dropped the DR on hard/elite from 30/60 to 25/35) because some classes (particularly Rogues) are weak against higher fort.

[Backley]

Since Update 5 nerfed everyone's DPS, did Harry and the other bosses get a reduction in HP? Would be interesting to find out.

[sirgog]

Horoth question: What difficulty made you suspect Badges do 6d12 damage per 2d3 seconds? Are they the same on all difficulties, or is that only from testing on one particular difficulty?

Arraetrikos/Horoth question: How often do these bosses seem to 'double strike'?

[Vanshilar]

Also, that info you have on Shroud Elite makes me question the accuracy of my tests in Shroud Hard. It would surprize me if there's such a tiny difference between Normal and Hard, and an enormous one from Hard to Elite.

Doing this pixel method to measure HP means that if others also damaged the target, then you will underestimate the amount of HP that the target has. In other words, say you did 1000 damage to get to the 2nd pixel, so that with 148 pixels, he should have around 1000*148 = 148000 HP. But unbeknownst to you, a groupmate did 200 damage while you did your 1000 damage, so in reality, it took 1200 damage to get to the 2nd pixel, so he would actually have 1200 * 148 = 177600 HP. Your measurement would be lower than the actual amount.

Anyway, it may have happened, it may have not. That's one limitation of the method, in that it doesn't account for if other people were also damage the target. When I test Arraetrikos's HP in part 4, I have everyone staying far away, so it's pretty clear that others aren't damaging him (unless they're ranging, but hopefully other groupmates would see and let me know that someone's doing that).


Since Update 5 nerfed everyone's DPS, did Harry and the other bosses get a reduction in HP? Would be interesting to find out.

It looks like monster HP are still the same. Testing post-Update 5, the Jailer in part 1 of ToD has 88258 to 95163 on normal, which matches with the 90k amount from pre-Update 5, and 146668 to 160579 HP on elite, which matches with around 160k from pre-Update 5.


Horoth question: What difficulty made you suspect Badges do 6d12 damage per 2d3 seconds? Are they the same on all difficulties, or is that only from testing on one particular difficulty?

Arraetrikos/Horoth question: How often do these bosses seem to 'double strike'?

It was on an elite ToD run where I was the main tank. I've only finished analyzing the data for that one run, so I don't know about the other difficulties yet.

Horoth seems to double strike roughly 70% of the time or so. I've worked out that his DPS is roughly 130 DPS for just the physical damage attack part (which is around 100 DPS) and the damage over time effects (which total around 30 DPS if you have all three on you).

[mediocresurgeon]

Suulomades (ToD) on Hard has evasion.

[sirgog]

Suulomades (ToD) on Hard has evasion.

Thought so. Is his Reflex save high or low?

[mediocresurgeon]

Thought so. Is his Reflex save high or low?

I don't know for sure how high. But it was high enough that the FvS who was supposed to kite him had a hard time establishing aggro, because he made his first few saves against Blade Barrier and Cometfall.

[wax_on_wax_off]

Can we see required intimidate scores for the various bosses?

+1 btw

really well explained and laid out, great read

[bobbryan2]

Spell resistance would also be a nice stat to have on hand.

[sirgog]

Any further testing been done lately?

[mediocresurgeon]

Has acid resistance been tested yet?

[sirgog]

Has acid resistance been tested yet?

I strongly suspect it to be 20 on Normal pit fiends. This is not a scientific test.

[sirgog]

One more thing:

I strongly suspect the Suulomades encountered in Tower of Despair is exactly the same monster as the one in A Vision of Destruction, except that in ToD he benefits from the Shavarath environmental effects (effectively +4 to-hit and +4 to saves).

This is because his CR is the same in both encounters, and whilst I have not measured his hitpoints, 150k seems about right on Normal - he takes about as long to kill as Horoth does, but Horoth has more people attacking him.

[Vanshilar]

Any further testing been done lately?

I did do some AC testing but they were pretty much all invalidated due to incorrect testing procedure (namely: the examine screen doesn't update until I close/reopen it, so I can't be sure if I accounted properly for AC-modifying effects like armor destruction), so they need to be redone. I also did some epic raid boss HP testing but wanted to retest before I post the results.

Otherwise though, lately I've been on a road trip and preparing for graduate school in the fall so have been doing somewhat less testing; also, I've been looking at renown so have been more busy with that as well. Once things have sort of settled down I'll get back to doing more boss stat testing -- although no ETA on that yet since I'm doing TA, RA, and quals this term.


Can we see required intimidate scores for the various bosses?


Spell resistance would also be a nice stat to have on hand.

I don't have a character with high enough intimidate to test that, and I don't have any casters so won't be able to test spell resistance either. Anyone who does though can post what their test results were if they have actual data.


Has acid resistance been tested yet?

So far, all I can say is that I've yet to notice any natural 20 acid damage (4d6 or 4 to 24) come up when i roll a 20 on my min 2. However, I don't have a large enough sample size yet to say if that's just unlucky rolls or if it's below the max. Suffice it to say though, it looks like the acid part of a min 2 is pretty much useless against devil bosses, since thus far they've all done 0 damage.


I strongly suspect the Suulomades encountered in Tower of Despair is exactly the same monster as the one in A Vision of Destruction, except that in ToD he benefits from the Shavarath environmental effects (effectively +4 to-hit and +4 to saves).

This is because his CR is the same in both encounters, and whilst I have not measured his hitpoints, 150k seems about right on Normal - he takes about as long to kill as Horoth does, but Horoth has more people attacking him.

Yeah hopefully eventually I'll be able to test the Suulomades that shows up in ToD, although it's a bit tricky to do (since I don't have a high enough intimidate to maintain aggro nor do i work well as a hate tank as a fleshie due to curse, so likely I'd have to arrange to "hand off" aggro to someone else once I've collected my data which is somewhat trickier to arrange compared with our regular ToD farm runs). Eventually I hope to also get some Nythirios data as well, although the problem with that is I still only have 15% incite which doesn't hold up that well.

[sirgog]

I did do some AC testing but they were pretty much all invalidated due to incorrect testing procedure (namely: the examine screen doesn't update until I close/reopen it, so I can't be sure if I accounted properly for AC-modifying effects like armor destruction), so they need to be redone. I also did some epic raid boss HP testing but wanted to retest before I post the results.

Armor destruction is VERY visible if a mob suffers it (they glow blue). Usually only Waves of Exhaustion will apply, and that is a -3 to AC.




Yeah hopefully eventually I'll be able to test the Suulomades that shows up in ToD, although it's a bit tricky to do (since I don't have a high enough intimidate to maintain aggro nor do i work well as a hate tank as a fleshie due to curse, so likely I'd have to arrange to "hand off" aggro to someone else once I've collected my data which is somewhat trickier to arrange compared with our regular ToD farm runs). Eventually I hope to also get some Nythirios data as well, although the problem with that is I still only have 15% incite which doesn't hold up that well.

Fleshies can hate tank Suulo just fine, especially now that the guild curse pots can be used by a raging barbarian. Nytharios, on the other hand, is likely to be very tough to get data on.

As for your Epic results - post them but say 'provisional, needs verification'. Sort of like my Hard Shroud figures.

[sirgog]

Did you get around to the Epic testing you had planned?

[Holymosher]

thank you very much sir !!! very helpful, !!!

[Ybbald]

What is the save on velah's breath on epic?

[Vanshilar]

I've updated the post with some more monster stats. Specifically, it turns out that Horoth and Suulomades in elite ToD both have 350k HP, so Suulomades taking longer is just due to having to maintain someone on Horoth and thus less DPS on Suulomades (not to mention, having to quaff curse pots a lot). Also it turns out that Epic Queen Lailat doesn't have much hit points, just under what Arraetrikos has in part 4 of Shroud on normal. Plus Queen Lailat doesn't have fortification either (although she has 20 DR). Of course, all those blades running through the platform also mean that you can die in a hurry, and also due to her attacks.

Testing AC on Queen Lailat was via a bow (while testing out the HP at the beginning at the first). I used a bow of improved destruction to reduce her AC by 8. Thereafter, a modified roll of 58 or higher (i.e. I roll an 8 with an attack modifier of +50) was able to hit her, while a modified roll of 57 or lower did only misses or grazing hits. I used examine to make sure that nothing else that affects AC was on her, and no other party member was supposed to attack anyway. My understanding is that matching the AC means a hit; let me know if this is not the case.


What is the save on velah's breath on epic?

I actually have quite a few videos of my barbarian standing in Velah's way, but since I also have quite a few VoN videos, it may take me a while to figure out which is which. Of course, a barbarian doesn't exactly have high reflex saves, so hopefully I was able to save on a roll lower than 19, in order to get some figures.

[sirgog]

I'm pretty certain than Lailat's AC is 64 on Epic.

I've hit her with a total roll of 61 while she is subject to Waves of Exhaustion (-6 Dex, which means -3 AC) and have missed on a 60.

That said, the epic Lailat encounter has been changed a couple of times.


In addition, one thing with the raid boss melee attacks. Suulo and each of the Fiends have three different melee attack animations - a first claw (slashing damage), a second claw (bludgeoning) and a bite. The bites do more damage than the claw attacks.

Also, I've tested (by tanking Suulo without fortification in a guild 'naked' run) and found that on VoD Normal, his claw crits do double damage (if they bypass fortification) and proc his healing curse (regardless of whether they bypass fortification). In addition, as sometimes I'm too slow to change equipment, I can confirm that in ToD hard, Suulo's melee attacks also crit for only double damage. Whoops.

[sirgog]

Further tests re: raid boss critical hits in melee:

Arraetrikos crits for double damage with both claws and bites on both Normal and Elite. Normal Horoth is the same, as is Epic Lailat.

Epic Velah - well, she crits HARD. The ~220-250 point crits I observed in there were probably triple damage, but could have been quadruple too.

[Bolo_Grubb]

interesting read

+1

[Wizzly_Bear]

/subscribe

[wax_on_wax_off]

Anyone know how many times/second various raid bosses attack and if it changes based on difficulty?

I'm just about to make my second GS item for my tank character and I can't decide between air guard or lightning strike guard. Air guard is a lot cheaper and adds some good utility for soloing while lightning strike guard will add hate while tanking. However, the amount of hate that it adds is dependent on the amount of times that it procs in a boss fight, I'm wondering if anyone knows what the odds are ...

Take Sully, Horoth or Harry for instance. As a null hypothesis perhaps we can assume that have 20 BAB and have the same attack speed as a TWF'ing character (maybe with an offhand claw). That would result in 86.7 attacks/minute with some percentage of offhand attacks (according to your thread on attack speeds). If that was the case, assuming no offhand attacks, a lightning strike guard would give about 16 DPS (86.7/60*550*.02). That's quite considerable.

Anyone know for sure attack speeds or hits/minute?

[Shade]

Anyone know how many times/second various raid bosses attack and if it changes based on difficulty?


Their attack rates don't scale with difficulty.

What does however: Is there double strike chance. (I call it that now, tho they actaully had this long before it was called double strike for players_

Normal arraetrikos has maybe a 50% double strike chance.
Where as elite has over 75%.

Nearly every raid boss double strikes at a high rate. Even as early as tempest spines Sorjek.

Lailat is more like having superior multi weapon fighting. Each swing always does 6 hits regarless of difficulty. Thus her special triple strike animation does 18 hits.

[Shade]

Epic Velah - well, she crits HARD. The ~220-250 point crits I observed in there were probably triple damage, but could have been quadruple too.

Not neccasarily. If you were testing on epic, she can do 125 point non crit hits, so 250 is only a x2.

I've fought a few times forgetting ot put my hvy fort on, noticed mid battle lotsa crits.. Then just never put it on anyways since i was at no risk of dying with my 900hp to see how bad it might be.. She doesnt really crit too ften.

Just has a ultra wide damage range.
I believe its 2 differnet ones, thus the variance.
EG: Common Claw attacks maybe 70-100
While less comon bite attacks are 100-130

Just tends not to be too challenging evne at 250+ pnt crits, as she attacks extremely slowly, and has a very small double strike chance compared to pitfiends much faster attack sped and ultra high double strike chance.
Rough estimates would be:
Velah can get 2 hits off during an quickened mass heal cast anim.. With maybe 5-10% double strike.
Horoth can get off 5 swings in that time.. each with a good chance to doublestrike (50%+).

[Emili]

Suulomades (ToD) on Hard has evasion. All the purple devils have evasion on hard/elite settings.

Has acid resistance been tested yet?
All the three named purple devils have 30 resist to acid and cold but goes up to near 50 on elite.


Can we see required intimidate scores for the various bosses?

+1 btw

really well explained and laid out, great read

I do not use the skill that often and prefer to just beat on them. Usually I only click such when things become messy or when bored and toying around so maybe someone with a DoS or DDS may chime in.

These are ballpark on elite/epic ...

Harry - 62
Abbot - cannot be
Abishai - mid 60s seems to grab it in epic but never nailed down the exact point of fail
Horoth - 66
Lailat - 76 epic
Stormreaver - 52 I think been a while
Suulomades - 65
Velah - cannot be
Titan - who'd care?
Xy'zzy - 80 I believe


Further tests re: raid boss critical hits in melee:

Arraetrikos crits for double damage with both claws and bites on both Normal and Elite. Normal Horoth is the same, as is Epic Lailat.

Epic Velah - well, she crits HARD. The ~220-250 point crits I observed in there were probably triple damage, but could have been quadruple too.

Seems on elite Harry hits for 65, Horoth 105, Suulomades 60 and x2 on crits, The Devils double-strike procs seem to be nearly constant on elite difficulty and the fact they have so many attacks when they go on a melee spree adds the risk when they toss a spell between. If not for Horoth's blasphemy coupled with Badges at 30-60 per tick and his combine double-strike he'd be a chump. ;) Those three badges plus a melee spurt when the healers are on cooldowns or stunned are the risky part.

Velah's melee to me seems slow even in epic ... I do not know if it's because she so stationary - stuck in mud - or is she's busy fire spells to the back-lot or slapping someone else in the pack that I do not notice or what. She seems to have a bite for about 130 but the slash damage is more 90 to 100. x2 on a crit still she is much slower than the elite devils and I'd wager somewhere almost about half their speed.

Lailat is by far the most melee intensive of the pack, on epic she does 40 to 60 per khopesh and those things have True Chaos Unholy Burst Screaming Thundering, Anarchic Burst and they by pass epic DR then toss on that tailwhip with 30 to 40 bludgeoning. Now usually she's six attacks on a spurt but every so often deliever's up to 18 attacks at once. I wonder if Turbine would give me a feat to wield weapons with my feet? Oh well, She's still my fave I do love my khopeshes and if I had six arms I'd do likely the same. ;)

[sirgog]

Malicia, Epic Big Top:

HP is approximately 150k (details
click the quote mark here)

Testing method: Solo* her with a caster, using the quite predictable DPS of 3-stacked Niac's Biting Cold and Eladar's Electric Surge.


*Solo here means noone else goes all-out DPS on her. Other party members are present but mostly concern themselves with trash killing.

[sirgog]

Their attack rates don't scale with difficulty.

What does however: Is there double strike chance. (I call it that now, tho they actaully had this long before it was called double strike for players_

Normal arraetrikos has maybe a 50% double strike chance.
Where as elite has over 75%.

Nearly every raid boss double strikes at a high rate. Even as early as tempest spines Sorjek.

Lailat is more like having superior multi weapon fighting. Each swing always does 6 hits regarless of difficulty. Thus her special triple strike animation does 18 hits.

What you are mostly seeing here is one animation with two attacks. Take a toon with no static DR, equip an item that reduces bludgeoning damage, and check your combat log when you think you've been double-struck - you'll see something like this:

Combat: Arraetrikos hit you for 53 damage
Combat: Arraetrikos hit you for 47 damage after 3 damage was blocked by your damage reduction (piercing or slashing).

That's 53 slash damage and 50 bludgeoning damage coming from his 'claw attack' animation, which has two attacks to it.

The Pit Fiends also have a bite attack, this does more damage than the claw attacks.

[Junts]

Malicia, Epic Big Top:

HP is approximately 150k (details )

Testing method: Solo* her with a caster, using the quite predictable DPS of 3-stacked Niac's Biting Cold and Eladar's Electric Surge.


*Solo here means noone else goes all-out DPS on her. Other party members are present but mostly concern themselves with trash killing.

I might be able to persuade someone to actually let me damage her all by myself tomorrow when I'm off timer just to test.

I've been considering trying to organize a shroud where no one damages Arraetrikos in pt 4 at all during the first round just to see how far an air savant can take him in the 90 second window.

[Vanshilar]

Still busy with school stuff but I went and tested Malicia's HP and AC. It was a solo run except with a hireling to heal. The HP was tested by the pixel method, while AC was from looking at the 1d20 + attack roll modifier totals. No destruction or other AC-modifying stuff was used. It was done this morning so post-Update 9.

The results were that she has between 96792 and 97222 HP, so she likely has around 97k HP. A roll of 13(+55) was a hit, while a roll of 4(+63) was a miss, so my understanding is that means she has an AC of 68 (if I understand AC correctly, i.e. roll that ties with AC is considered a hit). So you'd need +66 or higher to have guaranteed hit on a 2, although stuff like destruction naturally makes that lower.

Note that this was more or less solo, I don't know if epics will scale with number of players (i.e. whether or not epic HP has been tested to scale). She may or may not also have slightly HP on each run, this was tested for just one run.

[sirgog]

Still busy with school stuff but I went and tested Malicia's HP and AC. It was a solo run except with a hireling to heal. The HP was tested by the pixel method, while AC was from looking at the 1d20 + attack roll modifier totals. No destruction or other AC-modifying stuff was used. It was done this morning so post-Update 9.

The results were that she has between 96792 and 97222 HP, so she likely has around 97k HP. A roll of 13(+55) was a hit, while a roll of 4(+63) was a miss, so my understanding is that means she has an AC of 68 (if I understand AC correctly, i.e. roll that ties with AC is considered a hit). So you'd need +66 or higher to have guaranteed hit on a 2, although stuff like destruction naturally makes that lower.

Note that this was more or less solo, I don't know if epics will scale with number of players (i.e. whether or not epic HP has been tested to scale). She may or may not also have slightly HP on each run, this was tested for just one run.

Those HP figures are interesting - they disagree so wildly with Junts' test that it isn't funny, and I can't see a source of error that large in his method (the only possible source of such a large error would be if he mis-timed the '4 minutes'; another source if he had a statistically unusual number of crits).

Pretty certain Epics do not scale at all.


Your understanding of AC is precisely correct. She has the highest AC in the game now (Raiyum used to have that honor at 73 but he's been nerfed hard). IIRC Turigulon (Epic Devil Assault) is next at 67, which goes alongside his DR 30/Silver and (I believe) DR 30/good (It could be 35/good as well, I seldom run that quest when not on my main, and the main is on the TR train at the moment). He is quite a bucket of HP as well.

[Backley]

From another thread:

Abbot's HP has been adjusted from 50k HP (regardless of difficulty) to 135k HP (regardless of difficulty).

Far as I know it's usually just done on elite because the HP doesn't scale, and the puzzles are identical, so you might as well as get more loot (probabilistically). I'm not sure what changes in Abbot by difficulty. The DCs maybe?

[Vanshilar]

I've been busy with school so haven't really had time to update this or some of the other threads. But anyway, in case anyone's interested, the Lord of Blades has around 860k HP (and CR 58) on epic elite, around 350k HP (and CR 38) on epic hard, and around 266k HP (and CR 36) on heroic hard. When I have some time I'll try to update the other difficulties as well.

http://i898.photobucket.com/albums/ac182/Vanshilar/ScreenShot00806a.jpg

http://i898.photobucket.com/albums/ac182/Vanshilar/ScreenShot00809a.jpg

http://i898.photobucket.com/albums/ac182/Vanshilar/ScreenShot00812a.jpg

[sirgog]

Interesting, suspected from completion times it was well over 1m on EE.

Warforged Mastery will give absolute confirmation of what HP percentages his phase changes trigger at - whether it's 5%, 7% or 10% for the final phase, and so on. The ability to have more control over phase changes could be quite strategically useful.

[Shade]

Interesting, suspected from completion times it was well over 1m on EE.

Warforged Mastery will give absolute confirmation of what HP percentages his phase changes trigger at - whether it's 5%, 7% or 10% for the final phase, and so on. The ability to have more control over phase changes could be quite strategically useful.

mm probably due to the fact he has a gillion AC on EE.. Most everyone gets grazing hits on 2-4/5/6 due to that, lowering dps a lot.

Im kinda surprised he has more hp on EH vs HH, always felt lower. Probably due to the huge boost in dps the destinies added though, that and his other very poor stats on EH.

Check EN, bet thats like equal to HN, if not lower.

Also keep in mind monster hp is rolled randomly and varies GREATLY from run to run. Though possibly less for raid bosses since they are probably like 50d500+300,000 or something.

The optional rednamed spider in ID EE for example had 50k hp one run, and 95k the next.

[Vanshilar]

A few more HP for bosses:

Velah has 138k HP on EN, 303k HP on EH, and 743k HP on EE.

Lolth's Immanence has 311k HP on EN, 566k HP on EH, and 961k HP on EE.


Warforged Mastery will give absolute confirmation of what HP percentages his phase changes trigger at - whether it's 5%, 7% or 10% for the final phase, and so on. The ability to have more control over phase changes could be quite strategically useful.

As far as I'm aware, the trigger for the last phase change has always been 10%, or at least since I started doing the raid (I was a bit late to the party though, because I didn't have enough TP for the pack when it first came out). The first run that I recorded was on October 21, 2011:

http://i898.photobucket.com/albums/ac182/Vanshilar/DDOLoBHP1.png

You can see that he had 14 pixels left when the final phase triggered, which meant he was between 14/148 = 9.46% and 15/148 = 10.1% HP at the time. Or, after I changed the focus orb to directly have HP percentages on it:

http://i898.photobucket.com/albums/ac182/Vanshilar/DDOLoBHP2.png

Again, you can see that he was at around 10% when the final phase triggered. I don't know why anyone would think differently, it's fairly straightforward to just count the number of pixels and do a bit of simple division, or to just read off of the scale in the latter case.


Also keep in mind monster hp is rolled randomly and varies GREATLY from run to run. Though possibly less for raid bosses since they are probably like 50d500+300,000 or something.

Regular monsters differ in HP quite a bit because of dungeon scaling and stuff (and that they inherently have a fairly wide range). Raid bosses on the other hand, or at least all the ones that I've tested so far, don't differ in their HP very much, something like less than 1-2%. For example, Epic Elite Lolth so far has had between 960k and 962k HP in the different tests. I don't think that difference is really worth worrying much about for raid bosses. Certainly it will change a lot for regular monsters though, so I haven't really bothered to look at regular monster HP.