AC: A simplified suggestion

[Cyr]

The AC system falls apart fairly quickly in game. Basically lower AC's are pointless and while leveling very high AC's are equally pointless.

So the goals here are simple to any rework...

• Lower AC's should make a difference against all mobs at all levels of play.
• Very High AC's should make a difference against all mobs at almost all levels of play (the extreme twinkage possible at the lowest levels of play is going to be deemed not important enough to worry about).
• Create a system where grazing hits are no longer needed to insure some incoming damage in raid type fights.
• Insure that AC still does not protect against 'spike' damage...ie no static type dr stuff for AC since spike damage helps retain the danger element in game.
• Create a system where there is less concern over that last point of AC being achieved which likewise frees the developers to be less wary when making new AC gear or other buffs.
• Insure that any future increases to mob to hits or player ACs will not break this system. IE make the AC system work well within the framework of an expanding game.
• Eliminate the needs for any 'special' AC related rules as band aid fixes on a broken system.

So what do I propose?

• Eliminate grazing hits.
• Eliminate epic special rules involving to hit roles.
• Change how ALL mobs roll to hit as follows...
  
  • Mobs now have extra d20's to hit instead of base + to hit. This change would be made so that mobs have the same average to hit roll. Extra d20's would be assigned until the mob has less then a static +11 to hit.
  • Critical hits would be assigned a percentile chance of occurring based upon the mobs previous crit range and would be assigned in the proper upper range of rolls possible. This is simple for a computer to do.
  • Automatic hits and misses would be assigned as a ~5% chance based upon the d20 rolls being done to preserve that these are the top 5% of rolls or bottom 5% of rolls possible from the range being rolled. This is very simple for a computer to do. Note, I would not be opposed to elimanting these automatic misses and hits entirely as the dynamic ranges are much much larger in my proposed system then the current system making these less critical.

In practice what would this mean?

• Mobs could roll worse (much worse) then normal meaning that even the most minor investments in AC would add some protection against incoming attacks. A mob with a truly massive to hit roll currently that had let's say a 1d20 + 90 would instead have a to hit roll of 9d20 + 6 meaning anything over a 15 AC would provide some added protection versus their attacks. Even a raging barbarian can manage over a 15 AC if they bothered to do so.
• Grazing hits are not needed because mobs can roll crazy high to hit players. That same mob above could roll to hit a 186 AC. Also this lifts some of the headache from the developers when designing new items with AC in mind as there is no longer a 'magic number' that must be avoided otherwise content becomes trivial.
• Massive AC numbers do provide extra protection, but at a degraded rate as mobs are less and less likely to roll each AC to hit the higher they go over their average (unlike the previous 1d20 system where the odds of them rolling that number is exactly the same as rolling the average).
• Lower AC numbers before a mobs average to hit provide more and more protection as you approach their average to hit.
• Developers can create mobs with smaller then prior average to hits without strong worries about creating 'invincible' players. Likewise their old favorites of super high to hit mobs fit better in the overall structure of the game. This provides lots of wiggle room for content designers as they are not constantly having to worry about falling off the dice.

If anyone wants practical examples of any math I should be able to provide it.

[Monkey-Boy]

If anyone wants practical examples of any math I should be able to provide it.

Please. One syllable words if you can.

[Cyr]

Please. One syllable words if you can.

Let us take a troll who in the current system has a 1d20+14 to hit with an average to hit of 25.5 = 10.5 + 14. That troll threatens a crit on a natural roll of 19 to 20, hits automatically on a nat 20, and misses automatically on a nat 1.

The proposed system changes this as follows...

• The troll now has a to hit of 2d20+4 with an average to hit of 25 = 10.5*2 + 4.
• The troll threatens a crit on a to hit of 44 to 39 with a few specific rolls of 38 also threatening a crit to make the exact same 10% chance for the troll to threaten a crit.
• Likewise the troll automatically misses on a to hit of 6 to 10 with a few rolls of 11 also being an automatic miss to make the exact same 5% chance for the troll to automatically miss.
• The same dynamic works for the trolls automatic to hit happening on a roll of 44 to 40 with a few roles of 39 also automatically hitting to make exactly a 5% chance for the troll to automatically hit.

Now let's say we have Bob the gimped AC build. He has a 15 AC.
The troll in the old system would either critcally miss Bob or hit making the troll hit Bob 95% of the time. In the proposed system Bob is missed more often...

• The troll still automatically misses Bob 5% of the time.
• However, the troll also misses Bob merely based upon his AC 8.75% of the time.
• This means that Bob is now missed 13.75% of the time compared to 5% of the time before which is a considerable improvement for poor gimped AC Bob who is suffering ~2.75 times less damage then before. Not bad for a really modest investment.

On the flip side let's say Jimbo has a pimped AC of 35. The troll in the current system only hits Jimbo on a natural 20 (or 5% of the time) and who knows how much based upon grazing hits. Jimbo in the new system though is hit much more often, but never grazed as follows.

• Jimbo is missed 86.25% of the time or hit 13.75% of the time. Jimbo is taking ~2.75 more damage excluding grazing hits and crits.
• Jimbo has to bump his AC upto 39 before being only missed 5% of the time. Note this is a 4 point shift in AC for a 8.75% reduction in chance to be hit which in the old system if it was on the dice would have been 20%. These lesser returns the closer you get to the 5% threshold gets more and more prononunced the higher the mob average to hits are. This is a big part of the future proofing of this system change since a wider and wider range of viable top end ACs are on the dice as mob average to hit goes up (instead of only an 18 range in the current system).

So that is a pretty basic, but realistic explanation using numbers. If you want any further details please ask about specific aspects or examples and I can see what I can do.

[Stonen]

Great system!

It has all the elements I would want:
- Minimal effort in AC still matters.
- Every point in the middle of the mobs to hit matters most.
- High end still AC can be hit, but always has good chance to be missed.

[Monkey-Boy]

Let's talk bosses and bigger numbers . . .

So normal Horoth's 1d20 + 57 to-hit (right?) would be changed to a 2d20+47?

Currently a 77 AC is only hit on a 20 . . . now if he rolls well enough he'll be able to hit up to an 87 AC target . . . but his average is still the same? But the low 80s tanks would take more damage on normal as there aren't as many grazing hits.

Elite Horoth is 1d20+73 . . . if that's changed to 2d20 +63 I think a 90 AC tank will actually take less damage because he's not getting grazed on a 13. Unless I got the math wrong.

So please tell me I'm not understanding this right . . . it's looking like your idea will yield more difficult normal content and easier elite . . . and no change on epic where AC needs the most revision.

I strongly suspect from the lack of developer feedback on this subject that the "new" defense system has already been figured out anyway. it would be nice to know what they are thinking so we can feedback on that and not have to guess as much as we are.

[Krago]

The diminishing returns system works great as long as the increase to AC is not exceptionally difficuly to obtain.

The bump from 35 to 39 is not nearly as difficuly as say from 90 to 94 where those mobs would be rolling 9d20s in your system.

[Cyr]

Let's talk bosses and bigger numbers . . .

So normal Horoth's 1d20 + 57 to-hit (right?) would be changed to a 2d20+47?

Currently a 77 AC is only hit on a 20 . . . now if he rolls well enough he'll be able to hit up to an 87 AC target . . . but his average is still the same? But the low 80s tanks would take more damage on normal as there aren't as many grazing hits.

Elite Horoth is 1d20+73 . . . if that's changed to 2d20 +63 I think a 90 AC tank will actually take less damage because he's not getting grazed on a 13. Unless I got the math wrong.

So please tell me I'm not understanding this right . . . it's looking like your idea will yield more difficult normal content and easier elite . . . and no change on epic where AC needs the most revision.

I strongly suspect from the lack of developer feedback on this that the "new" defense system has already been figured out anyway. it would be nice to know what they are thinking so we can feedback on that and not have to guess as much as we are.

Nope, you are missing the main component of the systme. Elite Horoth would be 7d20+10.

That would mean higher possible to hits of 150 compared to 93.

[Monkey-Boy]

Nope, you are missing the main component of the systme. Elite Horoth would be 7d20+10.

That would mean higher possible to hits of 150 compared to 93.

Thanks . . . But with an average to-hit roll of 83.5 (right?) I'm still right on this meaning a 90 AC tank will still take significantly less damage than he does now since he's not getting crazes for 50-80 on a 13-19.

[Monkey-Boy]

Also how would debuffs work? WoE is standard for the devil raids (some use Str-sapping as well) and the double-curse is critical in N/H Lord of blades.

Would that double curse remove 8 from the total to-hit value?

[Cyr]

The diminishing returns system works great as long as the increase to AC is not exceptionally difficuly to obtain.

The bump from 35 to 39 is not nearly as difficuly as say from 90 to 94 where those mobs would be rolling 9d20s in your system.

I guess you would have to give me a practical example of why this would work worse then it does currently.

Do you think that an AC tank is not going to bother getting AC for elite horoth for example?

[Systern]

Let's talk bosses and bigger numbers . . .

So normal Horoth's 1d20 + 57 to-hit (right?) would be changed to a 2d20+47?

Currently a 77 AC is only hit on a 20 . . . now if he rolls well enough he'll be able to hit up to an 87 AC target . . . but his average is still the same? But the low 80s tanks would take more damage on normal as there aren't as many grazing hits.

Elite Horoth is 1d20+73 . . . if that's changed to 2d20 +63 I think a 90 AC tank will actually take less damage because he's not getting grazed on a 13. Unless I got the math wrong.

So please tell me I'm not understanding this right . . . it's looking like your idea will yield more difficult normal content and easier elite . . . and no change on epic where AC needs the most revision.

I strongly suspect from the lack of developer feedback on this subject that the "new" defense system has already been figured out anyway. it would be nice to know what they are thinking so we can feedback on that and not have to guess as much as we are.

You forgot about this part

Change how ALL mobs roll to hit as follows...
Mobs now have extra d20's to hit instead of base + to hit. This change would be made so that mobs have the same average to hit roll. Extra d20's would be assigned until the mob has less then a static +11 to hit.

So Elite Horoth would go from 1d20 + 73 to 7d20 + 10. This changes the range from 74-93 to 17-150 for his to-hit rolls. Both ranges average 83.5.



My question is what happens when your AC is in the range of the mobs 5% crit range? In your Troll example, what happens when I have 42 AC or more? Are the crits absorbed like fortification? What happens when a level 20 goes farming for muckbanes, when total AC eclipses the mobs entire to-hit range. Poor little kobolds. Even with Red Dungeon Alert, it's just a matter of slogging through until one acid blast takes out the entire dungeon at once.


My thought would be to not standardize on d20s. Mobs have hitdice. Players have hitdice. Use the hitdice for hitrolls to determine to-hit! Hey, that small brown spider is a 3d4 mob.. 3-12. That ogre is a 5d10 mob... 5-50. That Drow rogue is 8d6... 8-48. Casual Kobold 1d20. Elite Kobold 5d4!

Your abilities score wouldn't add a static value to the range, it would add an appropriate number of dice. But, the big problem is that this strays from D&D into GURPS or Whitewolf PNPRPGs...

This keeps the range of AC needed to be somewhat sane, since getting 140 AC seems so far out there.

[Cyr]

Also how would debuffs work? WoE is standard for the devil raids (some use Str-sapping as well) and the double-curse is critical in N/H Lord of blades.

Would that double curse remove 8 from the total to-hit value?

Debuffs would work exactly as they work currently. A strict minus modifier so they would have the same effect as an equal change in AC.

On the horoth thing...

Horoth would be rolling a 7d20+10 to hit with an average of 83.5 (correct). So a 90 AC would be 6.5 AC over that average. I would have to look up the grazing math on horoth, but my gut says that a 13-19 graze (30% of attacks) is outweighed by the increased pure hits that a tank takes under the modified system. This could heavily depend upon DR math going on also though. Let me see what I can come up with here for classic scenarios...

[HarveyMilk]

This idea has been proposed before, and again, I say /signed x3.

It's simple, and elegant.

[Monkey-Boy]

This keeps the range of AC needed to be somewhat sane, since getting 140 AC seems so far out there.

The crazy stars-aligned numbers that a 6 STR halfling can get with every buff under the sun for 2 minutes a quest are not relevant to the conversation. Those toons don't do any damage so even if they don't get hit they can't hold aggro.

The sustainable AC that a tank that actually has the ability to hold aggro with REASONABLE party buffs (bard, haste, recitation, no paladin aura unless you are the paladin, nothing silly like halfling's companion) is about 100 - that's the number that content should be balanced for.

Regarding non-tanks and trash AC for quests . . . this is an issue. As it stands now too much of it is gear-oriented and it would be nice if some of the basic gear was easier to get. We've seen this with Epic Cavalry Plate and the bracers of wind help . . . but it's just not enough. Perhaps making more of the AC come from PREs and APs instead of gear would be a better solution . . .

[Monkey-Boy]

Horoth would be rolling a 7d20+10 to hit with an average of 83.5 (correct). So a 90 AC would be 6.5 AC over that average. I would have to look up the grazing math on horoth, but my gut says that a 13-19 graze (30% of attacks) is outweighed by the increased pure hits that a tank takes under the modified system. This could heavily depend upon DR math going on also though. Let me see what I can come up with here for classic scenarios...

You seem far too math oriented to "go with your gut."

Me on the other hand, I'm a monkey and apes can't add very well. I'll go with mine. Mine's telling my under your system with those numbers the 90 AC tank will be nigh invulnerable which is what Turbine obviously doesn't want.

[Cyr]

You seem far too math oriented to "go with your gut."

Me on the other hand, I'm a monkey and apes can't add very well. I'll go with mine. Mine's telling my under your system with those numbers the 90 AC tank will be nigh invulnerable which is what Turbine obviously doesn't want.

It has been awhile since I have done any serious statistical analysis and I am having issues remembering how to do this using formulas instead of by hand (which is not fun for 7d20). I will probably be posting something later in the week so I can get something correct instead of estimated out.

[Systern]

As you said, 7d20 isn't fun... Let's use something a little more manageable... 4d20 (since 5d20 doesn't fit into a 32 bit int...)

4d20 gives a range of 4-80.

With 4d20, you have 160,000 permutations - possible rolls (1112, 1121, 1211, etc). (as opposed to 1.28 BILLION with 7d20)
You have 116,280 combinations (1 20, 3 1s, etc) (opposed to 390,700,800 combinations with 7d20)

Now statistically, there's only 1 way to roll a 4. 1,1,1,1. You have a 1/160,000 chance of rolling that 4 (0.00063%
chance). There are 5,340 permutations to get a 42. A 5,340/160,000 chance is 3.3375%. If you chart this out, you get a classic bell curve.


Your 5% absolute miss, and 5-15% critical range would need to cover, not the values along the X axis, but 5% of the area of the bell curve. 5% of the distribution of 4d20 covers all rolls from 4-22 (totaling 4.57188% probability, but including all 23s as well would put us at 5.53438%).

In your proposed system, AC 67+ gives you 99% invulnerability against 4d20. AC 54 gives ~85% invulnerability, and leaves that 15% range for crits.

[Xezno]

In the end i believe its just easier to increase the 1d20 dice to a 1d40 or more and decrease the base number. Its easier for the player to understand, and i fear a system with to many rolls would require to many calculations by the system.

Another problem with using the bell curve is that getting the last bits of AC into your build requires a lot of sacrefice. If getting that last AC dont matter much it would diminish the value of a lot of AC equipment. Increasing the die allready does that - by introducing the bell curve its even more so.

If you can you should allways pick the simple solution because the result more often than not turns out to be superiour.

[donblas]

Hey! Cyr your idea sounds like something i suggested on one of Sirgog's threads.

Therefore I like it!

[Krago]

I guess you would have to give me a practical example of why this would work worse then it does currently.

Do you think that an AC tank is not going to bother getting AC for elite horoth for example?

Sirgog put some nice numbers to the RoR (Rate of Return) based on AC increments in the current system. http://forums.ddo.com/showthread.php?t=365722

I believe that the mobs "to-hit" calculation needs to be changed for AC to be viable across a wider range. With your proposed system however, the viable AC range reaches to un-attainable levels.

7d20 puts the average AC threshold at 70. This is fine, but the amount of sacrifice it takes to get to that AC is substantial because of the limited amount of areas from which to draw AC, especially when they start removing some of the current stacking sources.

This is why I mentioned how much work AC tanks need to accomplish to get that extra 1-2 AC from equipment at the higher levels, when the AC starts to get into 80+. On a bell curve, the amount of work far exceeds the reduced damage seen.