Sunday, April 24, 2016

The Secret Ingredient to Brew Charge Talents Analysis

As of last Wednesday's patch, we got two big things: Healing Elixirs became a charge-based heal that auto-casts at low health, and Secret Ingredients got its proc rate buffed from 30% to 50%. I've heard that Healing Elixirs has become a decent niche talent now for fights where the other two cooldowns on that row but I'll leave specifics on that to the testers proper. Today I did a good chunk of work on comparing Secret Ingredients to the other talents on its row, as well as picking no talent whatsoever as previous looks showed that with artifact perks it was actually a net loss before now. This was of course before all the brew cooldown changes as well as haste CDR being added, so with the latest change I thought it would be a good time to revisit it! Shoutout to Xiaojin for getting me the SI data for this analysis.

Once again, the amount of data I have here is a LOT so I won't be showing many pictures. You can again find the raw data on my spreadsheet. Quick reference for format: All the numbers are assuming a combat period of 10 minutes with scripted / "perfect" use of abilities and averaged proc chances. Green highlights are when each specific talent option (LB, BoB, no talent) will always be better, teal highlights show when another option is only temporarily better. Things to keep in mind are that this is not meant to be a measurement of specifically when SI becomes less valuable than other options, merely to find whether it does, the general ballpark, and trends across all brew gen talents. As well, the numbers shown are theorycrafted, which means you should take them with a grain of salt. Things won't line up perfectly like this in game. That is especially true of BoB and it's effectiveness, as right now its numbers assume you waste 10% of the energy regenerated and can use all charges. This is very optimistic.

To get right into it, I actually have a graph for you guys this time!

So to talk about this graph for a sec, this shows the rough number of charges gained as you get more haste, your tier 19 4-set, and ranks in Face Palm. Each high point in the slopes represents when you have three ranks in FP and your 4set, with the new low being when you increase your haste by 5 but drop those perks.

Now obviously, Black Ox Brew is killing it right now, as it should. It's an active ability which requires very good timing on the part of the player to maximize its effectiveness. Light Brewing, king before the nerf to the base brew cooldown, is still fairly solid and only becomes ever so more competitive the more haste we have. The last two are where it gets tricky for brew generation. Secret Ingredients is actually very marginally worthwhile compared to no talent at all, but only until you get both your 4-set and 3 ranks in Face Palm, where no talent ever so slightly takes the lead. Again I do want to stress that this is theorycrafted data, which does make specifics like that imperfect and not wholly absolute. Being said, the general patterns and trends should translate into the game. So let's dive into the data and see what's going on here.

For starters, the biggest thing to notice is that you're getting a decent chunk more Keg Smashes. The problem there is that they are at the cost of Tiger Palms. Because both share the resources of energy and GCD's, you only get so many and more of one comes at the cost of the other. What's happening though is that the extra Keg Smashes are pretty significant (about 50% more at base haste), these come at a very significant cost of Tiger Palms. And because there's less Tiger Palms, the talent is inherently throttling itself when the other options don't have such mechanics. Looking at the differences between the 30% proc rate (which is also on the sheet) and the 50% proc rate, we're seeing a good number more Keg Smashes with the new iteration but again, much fewer Tiger Palms.

Ultimately, the 20% increase on the proc rate only results in 1 or so more charges gained across a 10 minute fight with no interruptions and as perfect a cooldown usage as possible. More or less, in regards to brew charge generation, it's negligible. Being said, that's a lot more damage opportunity, but as shown in the graph, at a fairly steep cost to defensive power. At three ranks of Potent Kick (+1.5 seconds total on ISB uptime), if you want to just be able to maintain ISB fully (which I don't think Blizzard is intending), you need 80 brew charges. BoB can do this at 0 haste and no ranks in FP / 4-set. LB can begin to do this at 10% haste, 3 ranks of FP, and 4-set (though it dips down afterwards). SI can do it at 25% haste, 0 ranks in FP, and with the 4-set (again dipping down afterwards). Those are some pretty far apart numbers.

Realistically, the issues with SI at least from a defensive standpoint really stem from the opportunity cost that KS shares with TP, very specifically because in the current iteration of the spec, TP is very important to maintain because the brew charge artifact perk, and set bonus really only focus on buffing TP. As shown on the graph, these things make SI relatively less valuable compared to even baseline and really the talent just throttles itself because of said opportunity cost. The more KS' you use, the fewer TP's, which means fewer procs. It ultimately seems to balance out the brew generation regardless of proc rate.

But the ultimate question I think is whether or not Blizzard intends for this talent to be the way hardcore raiders can get back that desired trade-off of defense for damage. If it does, it seems like it's a good way to do it, though I think it could use buffs in other ways than proc rate, like perhaps reducing the energy cost of KS when procced. If not wholly intentional, then I think a rework is in order. This was on the wishlist of my friend LeBlue, but maybe have it so that you get a second charge of KS that only activates when SI procs so that you don't just have weird RNG with the cooldown resetting a couple seconds before it's up, which wastes it. This is an issue that I see becoming ever more prevalent the higher haste you get, which could explain why the ratio of KS' compared to other talent options is reduced the more haste you get on the table. This way, you can hold onto the second charge and use it while the original CD can tick away as normal. Perhaps a combination of both options would work.

Another big issue is actually the t19 4-set and to a lesser extent, the Face Palm artifact trait. The 4-set is a temporary issue but both exacerbate the problems that SI has. Another potential facet of the solution is to change how these work... give them the same effect of better brew charge generation, but in different places. Maybe on Keg Smash. Another possible avenue is changing either KS' or TP's energy cost so that they are the same baseline, rather than just on proc. Either way reduces the opportunity cost. I think the biggest thing with SI is that there's not really one thing that will likely significantly change it, if it needs changing at all in the eyes of the devs. Any meaningful changes to the power of this talent would need to be directed at manipulating or breaking the opportunity cost and negative feedback loop the talent maintains.

So right now I think the data supports discounting SI as a serious defensive talent option and pretty much relegates it to being used purely for damage if you are 100% confident in your survivability. Ultimately I don't know what the devs want from this talent, it's just my hope to give as much info to everyone in the community as possible. This might be intended, it might not. My job is mainly just to parse info, I can only speculate on design intentions. We can only just wait and see how Blizzard decides to go forward and while I understand sometimes that's a difficult thing to ask, it's the reality we have.

Thanks for reading guys, hope this has been enlightening. As usual, feel free to throw me a message on twitter if you're interested in chatting further!

~ The Brewing Scribe

Wednesday, April 13, 2016

A Staggering Revelation: A look into Purifying Brew Post-Nerf

So it took me a bit but I've finally been able to create a proper model for stagger for build 21414. I'm just going to get straight into it. There's not really going to be any pretty pictures this time around but if you're interested in the raw data you can find it here, on my theorycrafting spreadsheet. For the sake of clarity, the model starts at 0 seconds in a fight and models for attacks every 1.5 seconds, which are assumed to hit. It then compares how much damage you would take in the next 6 seconds if you either purified and dropped ISB, or maintained ISB at that point in time. A green highlight means that purifying is guaranteed to be better from that point forward and an aqua highlight means that purifying is better at that point but it dips down at a later point.

The primary goal of this comparison was largely to determine what significance, if any, these changes have on our core gameplay: managing purifying and maintaining Ironskin Brew uptime. The findings were pretty conclusive: while there is a difference and it takes longer for purifying to become the right choice if you need to give up your ISB buff, it is only a matter of about 1.5-3 or so seconds, or the equivalent of one or two boss attacks using this model. So as far as the rough math goes, purifying is still useful and our core gameplay remains more or less intact. In reality it will take longer because you're going to be doding attacks more often but ultimately there is a point where you do benefit more from purifying.

The biggest difference is that purifying now reduces about half as much damage as before as far as how much I've modeled is concerned... which isn't unexpected. There's some worrying consequences though for intuitive use of the mechanic. Around every point on the table where purifying is better, there's about a space of two or so seconds where the numbers are virtually indistinguishable between using PB and using ISB. Personally speaking, I was taught that if a game required external assistance for a player to appropriately learn a mechanic, then the design is not doing its job properly. While this phenomenon occurs for PB when it purifies 100% of the stagger pool, its much less dramatic. The time it takes to see a significant difference in the damage taken when purifying in comparison to maintaining ISB is much lower than when PB reduces the stagger pool by 50%. This  has further consequences in that players will take longer to feel out when it's better to purify, which means they will take more damage as they hold off on purifying, making the nerf even greater in magnitude than just a 50% reduction.

There is actually another worry here that in the long-term of a fight, it's an even greater nerf. Because you can't totally remove the stagger pool via gameplay, over time it will continue to increase. This means that while you can purify at any point, by the end of a fight, you WILL be taking more damage per second than at the start. This amount could possibly be not a whole lot, or it could be more depending on how much purifying is practically necessary. This has a lot of consequences for gameplay and design. If you are constantly tanking (or at the very least tanking so often that you do not have 10 seconds where you don't take damage), you are ultimately liable to just take more damage unless you are fairly liberal with your purifying and even then it just delays and suppresses this problem, rather than fixing it entirely.

I see this having a notable chance of messing with balance in the new Mythic+ dungeons which are supposed to more or less be the new Challenge Modes. You're on the clock and if you spend 10 seconds waiting for stagger to tick off every couple of packs, that's a lot of extra time that you've wasted. Or if you only wait before bosses, you're potentially liable to be chunked on trash and that's not good either. On top of that, there's some affixes that could be very bad for Brewmasters, like the constant damage of Decay. You're just piling up constant damage and even if Decay isn't affected by Stagger, that's just even more of a constant mana drain for healers. In combination I think it's fair to worry about the viability of Brewmasters in this environment when you're talking about the higher difficulties.

This is also a potential problem for hardcore progression. While I'm not going to say that the game should be balanced around a very small minority of players (sorry guys), the worry that a Brewmaster would be taking more damage at the end of a fight than at the start, compounded with the ends of fights largely being tougher on everyone really makes Brewmasters seem like a bad class to bring. That is, I think, a very strong worry.

The change to purifying I think reinforces a fairly valid concern about Brewmasters and that is that we actually have very little control over how much damage we take. Our emergency heal is pseudo-randomly procced and requires us to move around which is potentially not a valid option depending on the encounter or the point in time in an encounter. Our cooldowns are very long and realistically not useful in moment-to-moment situations if you are on an encounter with sufficiently deadly mechanics that you need to save these cooldowns for. We have no passive mitigation so damage taken is roughly equivalent to a DPS without purifying, and now purifying can't even fully get rid of the amount of stagger we have which according to feedback that I've seen does not feel good. I realize that this is in part necessary because of the paradigm change from Warlords to Legion. Tanks as a whole are supposed to rely more on healers. That being said, I thing being able to purify some damage is important for that as well. We're giant mana sponges without it. I'm just a humble theorycrafter and budding designer. I'm not a Warcraft dev, I don't have the specific experience of being a class designer on World of Warcraft. Perhaps I'm just over-sensitive to the potential of this change. That being said, I still do think that changes to what we have wouldn't be a bad idea. There are other avenues that could be explored to balance Brewmasters.

Ultimately if the big reason Purifying Brew was nerfed is because Brewmaster's capacity to reduce the amount of damage taken is too high with it (which I don't know if this is the reason, I'd honestly really like a clarification on this still), then you have to ask how to try and rectify that while bringing purifying back up to 100% of the stagger pool.

You could reduce the amount of stagger Brewmasters have, but that has a lot of worry about taking way too much burst damage for the tools we have to mitigate it and that's already a concern of players (which admittedly isn't a foolproof metric for balance). You could reduce their self healing, but there's frankly not that much there in the first place. Reducing health doesn't reduce the amount of damage you mitigate and would make it easier for Brewmasters to be burst down. You can't really reduce the amount of passive mitigation they have seeing as they have zero from being a Brewmaster.

I think that strategic buffs and nerfs would be appropriate. Perhaps have less stagger from ISB but at the same time, give some passive defense and/or some more self healing. Just doing something like this would be much easier to test across various encounters and content to see whether or not it's good balancing and I think ultimately it would be the best for everyone. I don't see messing with purifying brew as an easy method of balancing the spec. There's just so much going on under the hood and it's exceptionally difficult to realistically model the consequences of that without scripting.

To wrap it up, I think the changes to Purifying Brew are... complicated. I think there's a lot of potential "down the road" consequences for it, especially in various content where downtime is not really available. Being said, the model I made really didn't have that significant of a difference for at least gameplay, so that at least is largely okay in the sense of actually needing to purify at all. Still, the model's worrisome because there's a larger period of time where as a player it's harder to tell when purifying is useful, so there's a double-dipping of increased damage taken. Overall I would like to see some changes and I can only hope that Blizzard appreciates this post if anyone from the class team comes across it. I'd like to say that I've been fairly reasonable and objective in this feedback rather than alarmist, but ultimately that's for Blizzard to judge.

Thanks for reading guys. Sadly, I don't have alpha myself so if people would be interested in trying out the stuff I mentioned here and seeing if it's worrisome, that would be appreciated. Any more specific data we can throw Blizzard's way the better.

~The Brewing Scribe

Friday, April 1, 2016

Brewmasters Are In: How to Get Back Into Giving Feedback!

It's time to celebrate, friends! We're finally able to start playtesting Brewmasters! There were a couple changes that we've found out with the latest patch but I wanted to take the time to talk about providing feedback in as effective a way as possible. But first, a quick summary of the changes.

Firstly, we got Detox back! So that's cool! Secondly, additional uses of ISB extends the shuffle buff's duration, which is pretty great. Thirdly, ISB and PB's cooldown is now affected by haste CDR. We're expecting to see their base CD rise up pretty high (at least 3 seconnds or so) but that's fine. Right now it's a bit silly. All that being said, onto the main topic!

Learning game design has been an amazing experience in really understanding and appreciating feedback for games. A good part of our first class is actually about learning how to give and receive feedback because it's so absolutely vital to the success of games. I wanted to take some time and go through all the things that in my opinion will enable the best, clearest feedback possible so that as a community we can communicate our worries and concerns with Blizzard with as little white noise as possible. To that end, I made a little list of things that I think we should all keep in mind when providing feedback.

1) Avoid buzzwords

People really love to use words like "clunky" or "janky" when describing gameplay they don't like. The problem with these words is that they are actually really bad at getting at the issue with the gameplay and communicating the problem. What makes it feel clunky? What's the specific problem? If a rotation feels clunky, is it because there's not enough resources? too many resources? There's not an inherent feel of flow in the gameplay? Why? Cooldowns don't match up with other cooldowns / resources? Too long of cooldowns? Too short of cooldowns? Too many cooldowns? Too few cooldowns? Answering those questions for yourself and then communicating feedback based on those answers means that Blizzard gets a clearer idea of what could be improved, rather than just getting a general feel of dislike. This actually goes into point number 2, which is:

2) Communicate what you don't like and why, not how you feel about it.

Being able to say what aspects of gameplay don't feel good and why that is provides much more meaningful information for Blizzard than the mere dislike of those things. Saying you don't like how our self-healing is doesn't really give Blizzard anything to go off of. That could mean a lot of things from the potency of heals to how easy it is to heal to how frequently you heal to how you heal. This is somewhat like number 1, but the core of it is to communicate what the problem is rather than tiptoeing around it by describing how you feel about it. Being direct by thinking about and communicating what you specifically don't like and being able to back that up with reasoning rather than just saying you don't like it will be beneficial for the spec by the end of testing.

3) Focus more on fun, not balance

Right now we're still in alpha which means Blizzard isn't yet done with implementing core content. They are still very much in an iterative phase. So right now, I'd say it's generally better to focus more on whether the spec feels engaging and fun to play than balance. You can make a point of balance, don't get me wrong, but that is mainly what beta is for, when everything's more or less penciled in. We have plenty of time to worry about balancing, since that can happen even after beta... but we do have a more limited timeframe to provide feedback concerning how the spec functions at its core mechanics.

That being said I would argue that mechanical balance (for example, our level 45 talent tier) is also important. Most of what I say here is primarily directed towards feedback on our rotation and our play. Providing feedback on the mechanics laid out in our talents and our artifact traits is perfectly valid. I'm mainly talking about fine-tuning of things like overall damage and defensive power. Like I said though, that's still important stuff, but at least for me, any feedback that I'm going to be giving for a while is probably going to be more related to either mechanics or overall feel of play.

4) Think about the spec now, and not what it was.

I realize this is probably one of the hardest things to do as a tester. Especially for Brewmasters who have had more or less a really fun spec for the last four years. And I realize that this is actually even harder because of something called anchoring bias, where we inadvertently place a greater importance on initial information (in this case, the old version of the spec) when making decisions. However, whether we like it or not, Blizzard has decided to change how the spec works. Feedback in the line of "I like the old version better" is probably just going to be met with "Well... we're changing it... so I mean... I don't think we can really do much about that, sorry." At this point it's a bit cynical to say it but saying that the old spec is better as one's feedback is kind of a waste of energy and a waste of bandwidth for feedback that could be improving what is here now.

Ultimately, if we're going to get something resembling what is currently on the alpha in Legion proper, our priority should be looking at the spec for what it is - not what it was - and ensuring that this is the best it can be. That should be our goal. This isn't to say that the old one is worse and the new one is better, that is purely opinion and I am not touching that, but we should all try and focus on making what is there now the best it can be.It doesn't mean you are wrong to feel the way you do nor that your opinion isn't valuable. It doesn't mean that you can't provide feedback using the old model in some ways but with a focus on improving whats there in the build. But the goal shouldn't be to have feedback that is more or less saying that the old version of the spec was better and that they should just revert it. I sincerely doubt that it would happen. That's just the ugly truth of it.

I realize that's a hard thing to ask of a lot of you guys because you really do love the current Brewmaster. I do too. But it is what it is, and we just gotta roll with the punches (pun intended). As a designer you have to learn this really quick and I think the same lesson applies for testers as well.

Hopefully this has been useful for some of you guys! I'm sure some of this is just reiteration but I don't see much harm in reinforcing good attitudes towards providing feedback. The better we communicate with Blizzard, the easier they will understand what our concerns are and what could be improved, thus speeding up the iterative process and making it much more likely that the final product will be of a higher caliber.

Thanks for reading and like always, if you ever want to chat I'm on Twitter or you can reach me in the comments section below!

See ya next time,

~ The Brewing Scribe