DDO Named Gear Planner

[FURYous]

What do I need to edit in the wiki to get the builder to recognize minor artifacts? a lot of them are not getting flagged properly in the builder.

[black_pearl2001]

Thanks for all your work. This is an excellent tool.

Trying to manually add crafted items/challenge crafted items but save button not active.

Do I need to create a file to save too first?

Edit: I went back to planner after posting this and trying to use save/load buttons caused application to crash. Re-opened and load button again caused crash. Tried multiple times with same result and refresh gear set button seemed to have no effect. However, when I selected item slot in management to re-create item all data was still there.

[Pfhoenix]

What do I need to edit in the wiki to get the builder to recognize minor artifacts? a lot of them are not getting flagged properly in the builder.

Unfortunately, minor artifacts is a set list, as there's no standard representation of that in the wiki. Which items are new minor artifacts?

[Pfhoenix]

Thanks for all your work. This is an excellent tool.

Trying to manually add crafted items/challenge crafted items but save button not active.

Do I need to create a file to save too first?

Edit: I went back to planner after posting this and trying to use save/load buttons caused application to crash. Re-opened and load button again caused crash. Tried multiple times with same result and refresh gear set button seemed to have no effect. However, when I selected item slot in management to re-create item all data was still there.

I found a bug with loading the file (not sure how long that's been there). Bug fix will be in the next release, along with updated minor artifacts.

[FURYous]

I found a bug with loading the file (not sure how long that's been there). Bug fix will be in the next release, along with updated minor artifacts.

Any ETA on the next update?

[FURYous]

Unfortunately, minor artifacts is a set list, as there's no standard representation of that in the wiki. Which items are new minor artifacts?

This seems to be a complete list Minor Artifacts

[Pfhoenix]

OP has been updated for the release of v0.8.11, which includes a bug fix and updated item data.

[Oliphant]

Awesome!!! So happy to see an update. I shot you $50 a while back for some pizzas because you deserve it, hope they were yummy

One upgrade that would be nice:

• Gearsets show in red where there is stacking conflicts, which is great.
• Would be useful if the stats on the gear items themselves show red when there is conflict in the gear set so I see right away if an item is doing what I expect.

Another change that would be interesting:

• Currently this only yielding sets where criteria is maxed at each priority stage.
• There is interesting solution space where the first priority is not absolutely maxed but merely close to absolutely maxed.
• Would be nice if you could specify a tolerance range for being sub-optimal at each stage.
• Like say first priority is CHA, but I specify a tolerance of 1 not 0, so I get all the top CHA sets in the first swipe but also all the sets with 1 below top.
• Now second step has more items to work with and a tolerance setting here could allow more items to pass to next step.
• This could matter if you have to make really hard sacrifices on lower priorities for the last drops of power associated with your higher priorities.
• Also would enable community to find more unique flavors.

[Oliphant]

Just restocked 5 pizzas (plainspeak), need that brain food

[Pfhoenix]

Just restocked 5 pizzas (plainspeak), need that brain food

Thanks, I really appreciate it!

[Pfhoenix]

Awesome!!! So happy to see an update. I shot you $50 a while back for some pizzas because you deserve it, hope they were yummy

One upgrade that would be nice:

• Gearsets show in red where there is stacking conflicts, which is great.
• Would be useful if the stats on the gear items themselves show red when there is conflict in the gear set so I see right away if an item is doing what I expect.

Another change that would be interesting:

• Currently this only yielding sets where criteria is maxed at each priority stage.
• There is interesting solution space where the first priority is not absolutely maxed but merely close to absolutely maxed.
• Would be nice if you could specify a tolerance range for being sub-optimal at each stage.
• Like say first priority is CHA, but I specify a tolerance of 1 not 0, so I get all the top CHA sets in the first swipe but also all the sets with 1 below top.
• Now second step has more items to work with and a tolerance setting here could allow more items to pass to next step.
• This could matter if you have to make really hard sacrifices on lower priorities for the last drops of power associated with your higher priorities.
• Also would enable community to find more unique flavors.

I've considered rewriting the build discovery process to be similar to what you're describing. Instead of a forced priority chain, assign a point value per point of property and then do a complete combination discovery. The problem with this approach is that it's inherently more complicated, requires a lot more fiddling than the current system, and will take far more time and memory to do.

[Oliphant]

I think keeping current algorithm and tweaking could be relatively easier than a full rework.

Here goes my attempt at algorithm for a priority step in the process, but perhaps inefficient:

• Set tolerance, how suboptimal you allow a stat for a given priority.
• For given priority, do what you do now to find the best combos, the ones you would pass on to the next priority under current program.
• If you can just remove these combos from consideration and just rerun the same process, you should get the best possible combos remaining (probably the tricky bit since combos are concepts extending off the data). If this set of combos is within tolerance, add them to your set of combos to pass to next priority but keep doing the process of eliminating the answers as you go and rerunning until you violate tolerance.
• Pass this expanded set of combos to next priority.

Pros: Seems like least tweak needed to update the code.
Cons: Seems like reusing the code rather than exploiting understanding within the code is computationally duplicative.

In terms of complexity, if you set tolerance to 5, doesn't my process sound like it would "just" take 6 times longer or so (0 tolerance is the 1x longer case) since its just rerunning same algorithm? At least this is not growing non-linearly.

[Oliphant]

Another way that would maybe be doable but maybe also efficient:

• If you are touching combinations in your algorithm already, maybe you can create a temporary ledger that scores them all.
• At the end of a priority calculation, pull the combos from the ledger that meet tolerance.

I don't prefer my idea over your idea but thinking it might be less work. Also, prioritization is a common practice in optimizing so going another route might be wilder than expected in terms of computational issues. Your idea has more promise for being focused. My idea seems inherently prone to yielding more answers (good news/bad news).

[Pfhoenix]

Another way that would maybe be doable but maybe also efficient:

• If you are touching combinations in your algorithm already, maybe you can create a temporary ledger that scores them all.
• At the end of a priority calculation, pull the combos from the ledger that meet tolerance.

I don't prefer my idea over your idea but thinking it might be less work. Also, prioritization is a common practice in optimizing so going another route might be wilder than expected in terms of computational issues. Your idea has more promise for being focused. My idea seems inherently prone to yielding more answers (good news/bad news).

What you aren't understanding is that any change requiring calculating combinations across the board is the Big Change. As soon as you say "Priority *but with this one exception*", you go from reducing the combination space every pass over a limited set to doing a complete combination check across all possible items in order to capture those exceptions. In other words, any exception to strict prioritization means complete combination calculations and massively increased time and memory usage. If it's acceptable to do that, might as well just allow full control over relative valuation of properties instead of a priority list.

[Oliphant]

What you aren't understanding is that any change requiring calculating combinations across the board is the Big Change. As soon as you say "Priority *but with this one exception*", you go from reducing the combination space every pass over a limited set to doing a complete combination check across all possible items in order to capture those exceptions. In other words, any exception to strict prioritization means complete combination calculations and massively increased time and memory usage. If it's acceptable to do that, might as well just allow full control over relative valuation of properties instead of a priority list.

Yeah. The algorithm under the hood is a black box to me at this point and algorithms are pretty counter intuitive, like even Simplex is kind of weird conceptually. I was a math major and getting nitpicky with handling the combinatorics part of it and certain kinds of coding related to it would be the part I'm more specialized in but don't know enough to develop an app part of it.

[Oliphant]

Would be nice to tag items to not try them, like a list of no touch items.

[Pfhoenix]

Would be nice to tag items to not try them, like a list of no touch items.

You can always create a slot filter to remove them from the potential item pool.

[Oliphant]

Slot filter restricts all items though, right?

[Pfhoenix]

Slot filter restricts all items though, right?

Only the ones you tell it to.

[Oliphant]

Only the ones you tell it to.

Ah, yeah. I've used the slot filter to take out niche attributes to nix a specific item before. Works pretty good although conceivably could be some collateral damage if another item shares the attribute. I'm spinning up a notebook to see if I can handle some combinatorics with a dictionary data structure approach. Dictionary can solve the speed issue if the combinatoric approach does not blow out memory usage.