Would it be possible to leverage some of the great work being done by the community over at wowsims · GitHub to support a robust solution for MoP? Very ignorant to what goes into AMR backend so please tell me if I’m out of line. Given we have generally reliable tools in place for modeling damage outcomes that are accepted as accurate, is it possible there is a path that doesn’t reinvent the wheel but is also seen as comprehensive?
The problem is that simulation is too slow to be the backend for a gear optimizer. We tried that for a while (Legion, BfA, Shadowlands) and the results were decent, but it was difficult to maintain. Trinket rankings remained just as difficult to tweak and maintain. The noise inherent to simulation results in needing to create massive amounts of data to get a good model to run the site off of.
Simulating on demand seems like a good idea… people use sites like raidbots and simulate a curated subset of setups, whether that be talents, a few items swaps, etc - but you can’t really do much more than a few dozen or a 100 combos - and even that is quite slow. It is very hard to even know which setups to simulate. Simulating on demand is not a good approach for what we do. Features like BiB and the Upgrade Finder wouldn’t work with simulation on demand.
Ultimately we abandoned simulation because it isn’t the right approach for our goals, which are to be able to gear any character, with any build, with any subset of items available, in 1-2 seconds.
2 Likes
Hey, sorry if I’m bothering you, but are there any updates regarding MoP? I guess you won’t update the site for the last phase of Cata.
We are working on 11.1 right now. Once that is out the door we’ll spend some time looking at MoP to see what kind of implementation we could do for it. Until there are some test servers up, it’s hard to get a good picture of the time it may take.
1 Like