Getting some very strange enchant and gem recommendations. Vitality on boots for instance, which is 4 health and mana per 5, when +12 agi exists. Same holds true for gems, which seem to favor 8ap + 2mp5 over 16AP or 8agi. Strangest one is probably Glyph of the Outcast for head enchant.
The stat weights seem to far overvalue mp5 and int. The fact that 2mp5 is weighted greater than 8AP makes zero sense just with the eye test. I understand that when OOM, less damage is done, but the model seems incredibly sensitive. The above snapshot is using default buffs and debuffs. If I add improved mana spring totem (which I likely won’t have in hunter group) which adds a total of 3mp5, it drastically changes the recommendations. (Snapshot: 24594ac45a8148a6b8b3c39fdaf0a866).
Happy to test more. Appreciate the effort thus far, love the tool, just looks like there are still a few kinks. If you can expose a bit more of the sim logic, could probably help identify where the issue resides.
You are correct that the model is quite sensitive to running out of mana. If you run out of mana and can’t use your shots, your DPS will go down drastically. The model finds that stats to get you more mana can become much more efficient stats. I’ve reviewed the calculations and I’m not sure that there is an issue with this advice. We ran into the same thing in classic and I checked it over many times there as well. Maybe other folks haven’t considered the limiting factor of mana closely enough in their calculations?
For people who want to ignore the mana items, I recommend setting a shorter fight length. Or, add in lots of mana buffs/consumables (as you found).
The model has a bunch of logic to get the most damage from your mana pool. If you are mana-limited, it figures out which shots do the most damage/mana and culls out less efficient shots. Usually this means it ends up only using steady shot because both multi shot and arcane shot generally do less damage/mana spent.
That is fair. I played around more, and even without “sim” results, I can see adding dark runes and mana pots has the expected effect of prioritizing damage stats. Thanks for looking into it.
They don’t appear to be in the list, and although gems exist with the same stats, it has the unfortunate effect of causing them to be replaced when using BiB or Upgrade Finder when they should be valued higher than the gem which replaces them.