Hello,
I’ve always enjoyed Ask Mr. Robot, but the sims for Sub Rogue are extremely wrong. It’s asking me to stack haste while I already have more haste than what everyone (and the top 50 sub rogues) have.
I don’t believe that even last tier haste was valued so highly.
Please see this side by side picture of Simulationcraft and AMR -
I’m not sure how much the devs of this check here, but I’ve been subbed for years and am hoping the quality of the product hasn’t gone down over time.
@yellowfive & @Swol are in here more than their post counts would have you believe.
I feel that using terms like ‘wrong’ are very combative, as the only people who can say that anything about how a character is geared is Blizzard.
Unless you are a coder of the level that @yellowfive or @Swol are, it’s a bit presumptuous to call out their code. In that respect, they have always maintained that all stat’s have value - however small - if the final job gets done.
I understand that you may feel that something doesn’t feel right, but please be mindful that the above observations DO need to be borne in mind before ‘hyperbolic rage mode’ is engaged.
Pretty far from rage mode, I’m just surprised that it’s completely unusuable due to the bug or whatever is going on with Sub. I cannot use the program in any way with it asking me to basically jump up 10k haste whenever 3k haste is ideal for sub. I think saying “I’ve paid for this for years and just now it’s horribly wrong” is very valid whenever I’ve supported the project for a very long time.
Could you press the “help” link next to the big “Best in Bags” or “Best in Slot” section header (depending on which tool you were using), and then copy the generated 32-digit code here? With that we can see your exact settings, talents, etc. and comment further on why it might be choosing what it is in your case.
I’ll have to think about this one… for me to change the calculations such that haste is nearly worthless for sub rogues, I would need a lot of damage to be largely unaffected by haste. Haste mainly affects combo point generation (shadow techniques) and energy generation… for haste to be so low in value we would need to be in a situation where the additional resources are unnecessary. Obviously my current calculations are not coming to that conclusion.
Here’s the snapshot ID
abcf922be94b4319ba3bf711e797ae85
For reference, Sub rogue should be as close to zero haste as possible and I already have almost double what the top 50 sub rogues have.
Yes, I understand. Please keep in mind that the very top logs can be outliers. Getting max damage gearing exactly like those players will also require that you play like those players.
I think the reason that simulations and top logs are preferring very low haste is because they are doing almost all of their damage during shadow blades. I see some spikes in damage when secret techniques and symbols of death line up in between… but really it appears to be all about that one big damage window.
If you are not maximizing that window, the low haste gearing strategy will not be good, as you are pretty gimped in between due to low resource generation.
My calculations were taking an approach that smoothed out the damage more, which would lead to lower overall dps, but definitely a more forgiving way to play/build. I’m working on an update that will shift it to this higher skill style.
I used to play sub rogue as my main, so I know it’s all about burst damage. But it looks like it has skewed extremely heavily towards the 90 second cooldown window of shadow blades and flagellation at the top end of play. I was looking at it more from the perspective of maximizing each shadow dance and getting more shadow dance uptime, which is why I was getting a higher haste result.
90 second CD windows are the only way it works in general. I get what you’re saying, but sub is considered the highest skill ceiling spec in the game and I don’t think anyone is going to spec into sub for an “easy” playstyle. I’ve used your program to get 95+% parses in 7/8 of the LoU mythic fights as a dev evoker (along with other tools of course) and there was no discrepancy.
I’ve been going between assassination and sub and outside of the 90 second window, sub is absolute dogwater no matter how you do your stats. Right now it’s only good in 10s+, more likely 13+ now as of this reset because mobs need to live long enough for the 1.5min CD timings to make sense. If the 90s timings didn’t make sense, you just wouldn’t bring a sub rogue.
I would even say that outside of 90s I doubt that haste has much value, every other sim tool reflects haste being useless (this is with 4k haste)
Yes, I hear what you are saying. I’m working on some code that centers around the 90 second cooldown windows for these popular builds. I suspect the result will be low haste, as haste doesn’t do much for you while shadow blades is up and you are swimming in combo points. It isn’t super hard to write some code for the one most popular build to get an answer that aligns with how most people are playing that build.
The difficulty comes in also supporting any talent build that people can input into the optimizer. For instance, what if someone doesn’t take flagellation? Or they don’t take dark shadow? That is going to take a huge multiplier out of the 90 second cooldown window. Is it still best to center the entire model around it? What if they take gloomblade also? That shifts damage to times outside of shadow dance. I have to support all of these possibilities. Are those builds going to do as much damage? No. But, our website isn’t about telling people how to play the game. Our website is about telling people how to maximize the potential of the build they prefer to play. One thing that we do, which I don’t think anyone else really does, is truly give a fair shake to ANY build. You would be surprised at how many off-meta builds there are that can get to within 5% of the max theoretical build.
That stat weight output you are posting from simc doesn’t really apply to our conversation very much. Especially this last one you posted where you put some haste on the gear - if you add haste to your character in a model of the game (simc) that doesn’t want haste anyway, the “stat weight” for it will obviously go down. Those numbers from simc, in plain english, are telling you this: for the next fairly small chunk of stats (1 item, maybe 2) I add to my gear, what would be the best to add? If they already want basically zero haste, they’ll never want you to add haste. Putting more haste on the gear you simulate will not make haste look better. Those numbers from simc are NOT answering the question: how good are the stats on the gear I simulated?
