I’m trying for frostfire weights of roughly 39% haste, 29% crit, 17% mastery, 15% versatility.
On my live character I have 43% haste, 33% crit, 13% mastery, 10.5% versatility.
So not perfect, but in the ballpark of my goals.However when I try best in bags, it’s suggesting I go with this:
33.5% haste, 35.5% crit, 21.5% mastery, 9% versatility. I think it’s suggesting a low haste because it’s recommending I use the quickwick candlestick, but the 1/6 uptime of that trinket doesn’t make up for all the time haste won’t be up. I also think part of my issue is that haste is so much more important for me, so losing any haste for mastery usually isn’t worth it until I have much more haste. So even if the weights are aligning towards my goal correctly, not prioritizing haste/crit ends up being detrimental in the end.
I’m interested in your thoughts on this, thank you.
If you don’t want the haste from the trinket proc to “count” - increase the amount of haste you are going for in your custom stat settings. Or, exclude that trinket from the optimization if you don’t want to use it at all.
I think that should allow you to get to what you are trying to do.
So if you take your second snapshot and lock in your two current trinkets, you’ll see that it doesn’t recommend any changes.
There is no way to make the optimizer ignore stats gained from trinket effects when using a custom secondary stat target. It will always average them in and then do as best it can.
I’ll see if we can mess around with the gem/enchant part of the custom secondary stat target optimization a bit… but read on for why I don’t think it really matters.
One thing to keep in mind: once you enable the custom secondary stat target, the optimization gets a lot more approximate. This is because you are essentially fighting against the optimizer… you are telling it to ignore all of the calculations that it does, and simply override the secondary stats to an arbitrary ratio.
That’s fine – we made the feature specifically so people could do this – but there are still a lot of things that aren’t secondary stats that need ranking: primary stats, direct damage effects, set bonuses, etc. So it turns into as much art as science: how do we balance using our calculations to rank these extra things, yet override just the secondary stats to this custom target? The difficulty becomes: there is no objective way to relate those two goals to each other, i.e. there is no good way to say how much more damage you will do by getting closer to this custom stat target, because it is explicitly overriding our calculations that could assign such a value to it!
That is why the slider below the secondary stat pie chart exists – it is a way to nudge the optimizer more aggressively towards using our underlying model or the custom stat target as you see fit. There is no “correct” setting for that – just move it around until the optimizer does what you want.
If I take the solution from your snapshot and start changing around gems, I can find a couple solutions that are about 0.1-0.3% better. Given the fact that using a custom secondary stat target is already a bit of hand-waving, I’m OK with that result. Numerically the found solution is very close to the maximum possible numeric solution… but there’s really no way to say whether it would be better or worse in-game, so these small margins of error mean nothing.
One thing I would highly recommend when using a custom secondary stat target is to enable the Gem+Enchant Threshold. At least set it to something like 0.25%. Then it will try some more combinations of gems/enchants that involve saving you gold by using what is currently on your items.
I do usually set 0.25% threshold for gems/enchants, but sometimes I turn it off to help me target gear with stats amr thinks I need more of. Like if it makes all my gems crit/mastery, then I try to target a dungeon or something with crit/mastery gear in it.
I trust that in patchwerk type fights your weights are best. The issue is so many fights have me moving all the time. So that extra haste lets me finish casts I’d have to cancel without it. I can’t quote a number for how much increase it would be today, but when i originally swapped the weights to a custom setup, I want to say it was a 10% dps increase. My cancelled casts plummeted. I guess I could always try the default weights again and see how it goes instead of continuing to assume haste is better in most cases.
Yeah that’s one of those things that is really hard to measure, and really hard to simulate. I can see how shorter cast times would be beneficial on fights with lots of movement, so it makes sense to adjust it if you are seeing gains from that.
So I finished a few tests and the gap between default weights and my custom which had a lot haste, went down to a slim margin. I’d need to run it a bunch to figure out how much difference there is. Bottom line is that on fights where I was lucky with mechanics, I saw a dps increase. But on fights where I was unlucky with mechanics, I saw a dps decrease. And on fights where it was normal, my dps was give or take 2% of my normal dps. I’m just theorizing, but I’m guessing after X haste (10k? 15k?) you don’t really see any additional dps gain from avoiding cancelled casts. And switching to stat weights more similar to your recommendations is an improvement. Time to farm some crit/vers gear I guess lol.