Custom Sims / Client DL

Hey all - Just returning to WoW. I used to use AMR exclusively to test out custom dps rotations and tanking scenarios, just curious where the simulation page went. I can access Best In Bags etc, but that’s not really what I’m looking for. I can’t seem to find it in any of the menus. Is this no longer supported?

1 Like

I just answered a similar question (see below).

I know this transition won’t be particularly useful for experimenting with rotations - that is the main benefit of a simulator. Our site is mostly focused on picking good gear, although our development has shown that we are able to gain a lot of insight into rotations through our mathematical models. I hope to share as much of the under-the-hood calcs as I can during the expansion.

Do I understand correctly that there will no longer be a tool where I can simulate my “custom” rotation?

1 Like

Not in the near future. We have put our effort into making the optimizer customizable instead, and modeling the mathematical relationship between all the talents/abilities of the specs. I’m not sure what you were customizing in particular, but we can probably help you get a setup in the optimizer that picks the gear you are looking for.

The gear you want is significantly more rotation-agnostic than one would think, after we’ve gone through the calculations.

The community loves simulation though, so simc is always available to mess around with custom rotations, although without the graphical interface for setting up rotations :frowning:

I used this to test “my custom” rotations. It was very useful to test my hypotheses. My rotation for holi and proto paladin was very different from the default. It was also a very handy tool for testing rotation nuances. Since then I for example have a WA that tells me the best way to use SoTR and WoG.

Finding the optimal gear was a nice addition to a great simulator.

2 Likes

Honestly one of my favourite things was making rotations / optimising my amr results. I used to love making custom gearing strategies (they often were significantly different). I get why you guys are moving even more away from this though as I’m sure most users just want bib and done. I understand simc exists but it really it is a pain to use compared to amr and doesn’t really support tanking very well.

Is the new mathematical model going to account for custom rotations? For example I’m not a huge fan of spamming sck (which is accepted as the “best” thing to do) on my Brewmaster. Because of this I have been using a custom rotation to Sim for quite some time… If the model assumes I’m spamming sck the talents (more so than gear) would definitely not be weighted appropriately for me.

Anyhow I look forward to seeing what the new system you’ve devised is like! Thanks for the hard work!

1 Like

I’m in the same boat. I’m not a fan of just assuming everyone uses the “Wowhead” rotation which it sounds like may be the direction they are going. We shall see!

1 Like

We can theoretically add options to not use certain spells - some would be harder to support than others. As far as talent choices - I think it would also be quite easy to exclude certain talents if you don’t want to consider them (once we get talent picking features up on the site).

What I mean is I usually use my custom rotation to help me decide if a talent is worth taking. For example - if I use the “standard” Brewmaster rotation charred passions will do a lot more damage than it will the way I actually play. It’s not that I want to exclude charred passions it’s that I want to see if it’s worth taking the way I play or if dragonfire will edge it out.

1 Like

Dang. I am playing DPS this tier, so there’s less nuance than when I was tanking. But your sim page / client was just vastly superior to dealing with SimC. And customizing / tinkering with rotations and specs was 90% of the fun. I guess I’ll have to wait and see what I think. Thanks!

Edit: Although, if I ever swap back to tanking, I’m really going to miss modeling my own dps playstyle and defensive usage. :frowning: This site helped me get mostly all 99s and 100s every tier and be as beefy as possible while doing it.

3 Likes

Hey, I get it. For sure. I love our simulator! I also thought our simulator was a big step forward in simulation compared to simc - we designed it specifically to advance simulation and bring it to more players/theorycrafters. What we did not anticipate was that the folks who make simc would have a vice-grip hold on the theorycraft community and also have zero interest in using a different simulator, even if it has better features.

Regardless, the new talent trees do not lend themselves to analysis via simulation. There are way too many relevant talent builds to ever test well via simulation. We need something faster. Once we get that done, we can assess whether or not there is time/demand to update the simulator for dragonflight, at least for tanks/healers who have no alternative.

3 Likes

The simulator is definitely a fun project and a huge step forward for people who want to customize their own rotations. It’s too bad that most theorycrafters were too stubborn to try it or even admit it.

For now we want to focus our efforts on getting the Dragonflight gear optimizations really dialed in, and releasing some new features to help people choose talents. A talent optimization feature has been requested for some time, and it’s even more relevant with Dragonflight. It’s also going to be extremely difficult to implement… but we’re trying! We hope to have an alpha test of it for all users to try soon.

At first we just dismissed talent optimization as intractable. It takes a LOT of time and computing power to generate the simulation data required just for ranking gear. Now throw in talents… a typical spec has over 500 million relevant talent combinations! And talent combinations are in general more difficult to rank than gear combinations…

So then we said… well what would we have to do to make it even possible? Firstly, make it about 10,000 times faster to “simulate” a setup of talents+gear (that’s not a number for rhetorical effect, that’s about how much faster it would need to be). And the only way to do that… is to not use a simulator! So we have been working day and night to create a new kind of model that is extremely fast yet still accurate enough to rank things.

We’ll keep people posted, but so far it is working well. And it will have a ton of side benefits that we were locked out of doing when basing everything on simulation, like giving users way more options to customize the optimizer on the fly, being way more responsive to patches and game changes, etc.

1 Like

Yeah, the push-back from some of the Discord communities on Tanking sims always struck me as odd.
The tools were there to adjust your playstyle, health pot usage, defensives, etc. and give you “better than napkin math” data on the relationship between survivability, stats, talents, and decision making.

The problem seemed to me to be more of an administrative one - Moderating a community of tanks who might need help unpacking all of the data. Because while “just sim yourself” is an easy mantra to stick with in DPS forums due to the relative ease of modeling dps, tanking is more of a pass-fail ordeal that inherently off-loads some amount of work to the healers. 90% of DPS players may not even look at the sim rotation being used, hit 75% of their dps potential, and call it a day. But with tanking, you have a ton of nuance in healer gameplay, comp, and playstyle. You really need to look under the hood and understand what’s going on. Finding those gaps in your defensives / rotation where the sim had no good outs, and planning for those situations ahead of time with externals etc. Or finding new ways to improve the default logic.

I guess it’s just easier to discourage sim conversation than it is to answer 900 questions per day about how to use the tools properly.

1 Like

Oh - And not to keep bugging everyone, but will there be an overall DPS value output in the AMR tools? Or am I just missing it. I’m seeing DPS +/- delta on Best In Bags (etc), but not really understanding the rationale for hiding the overall value.

In any event, thanks for the hard work - I’ve always liked the site and appreciated the responsiveness of everyone on the forums.

This is because we can use a more formulaic model to calculate the breakdown of spells and capture the relationship between stats and talents, but predicting the absolute dps value players will get in-game is much more difficult. The mathematical approach tends to get much higher dps numbers than are achievable, but the relative values - say from taking one talent vs another, track very well with simulation data when we compare to something like simc.

If we wanted to calculate the actual dps value you could get - we’d have to start trying to guess how often you are unable to perform optimally - when a high priority spell with a 20 second cooldown becomes ready, how long does it take, on average, before you cast it? We know you are trying to use it “on cooldown” and can calculate the right gear to use based on that - but you can’t do that. Simulators are good for capturing that type of thing - and even the simulators tend to be “too good” for a lot of specs. It isn’t necessary for comparing gear and talents, though. That is why we focus on relative values instead of absolute - saves a ton of complication.

I love you guys, but I’m pretty unhappy about elimination of one of core site features.

I hope you reconsider.

2 Likes

We were able to provide a free simulator because we also used it to generate our gearing strategies and data for the gear optimizer.

The changes in dragonflight make using a simulator to generate all the data we need impractical at best and most likely impossible.

Updating the simulator would take 2 months of solid work, at least, probably 3 realistically. For us to do that without being able to generate revenue off of it… would not be viable for us. We are a two person team - we have to pay the bills!

We may be able to put the simulator into a state such that the UI is updated for dragonflight and hooked up to the talents, but then the users would have to go into the wiki and update the mechanics/rotations for dragonflight. My guess is that people wouldn’t really want to do that… but if there are some people who want to we could consider it.

1 Like

I think you underestimate the demand for this. AMR was the only acceptable place to make healer rotations comparable and test them. It is also one of the major reasons I support this site financially. I’ve literally been checking the forum every two or three days since pre patch to check if it’s coming back already.

I believe if you make it a bit easier to share the rotations with the community instead of just having the public ones and give people time they will definitely do a lot of work for you. I don’t know how much work it is on the simulator side, because your client is closed source, but APL related people will use this.

1 Like

Updating the simulator would take 2 months of solid work, at least, probably 3 realistically.

I find this surprising, since a lot of the dragonflight talents are copies or variations of stuff that has already been in-game (and therefore supported) in the past, in the form of legendaries, conduits, etc.

Part of the decision came from the fact that there was very little demand! We could see how many people were doing simulations on our site… and it was very small.

I enjoy the healing simulations - those were like my passion project. There is a good chance I will update the healers in the simulator just for fun, especially because there is no alternative tool out there for players to use.

I have completely given up on engaging the theorycraft community in using it, though. They won’t use it. Rotations were already designed to be shared - you could create a link to your rotation to share with anyone. And it was a live link too - perfect for theorycrafters. People could use a rotation curated by their favorite theorycrafter and it would be updated whenever they updated it, without needing to share a new link. The whole simulator was designed from the ground up to be a resource for theorycrafters - but other theorycrafters do not want to be associated with AMR in any way, even tangentially by using a free tool we provide.

We had always hoped for this type of setup - a crowd-sourced simulator that is basically open source, but we had almost zero takers when it came to updating the theorycraft wiki and implementing specs. All of the game mechanics were completely open-source in the theorycraft wiki. The only part that wasn’t open was the code that runs the simulation - but that code almost never changes. Only the game mechanics and rotations need work/updating.

After keeping the simulator up to date for 6 years, we have a pretty good idea how long it takes. There are 38 specs. Updating the numbers/mechanics for each spec takes at least half a day, and then making a good rotation that handles all these talents takes at least a day. Best case 1 day of work per spec, realistically 1-2 days per spec.