Should Blizzard Adjust Draw Rates of Over-performing Cards?

This was in response to Zeddy's video today: https://www.youtube.com/watch?v=xBmmACAz8j8 I just felt like sharing my response here because I really have no life XD

The nerfing method they are oblivious to is tweaking the draw rate of an overpowered card. Take example of many broken cards that have gained a mana cost and have fell into oblivion, Spiteful Summoner an example. The better solution is to simply adjust the rate it can be drawn from deck. Right now every card has a 1/X chance of being drawn X being deck size. I believe that for over-performers they should occur less, thus carrying decks less consistently.

This already sort of exists in the game in Arena where outlying classes get worse drops and weaker ones get better drafts. I don't think it's too much of a stretch to apply this to the general game where, sure, you can add Secret Passage but just know that it occurs just 4%* less in a draw than your other cards will. There can even be written an algorithm where the more a card is performing inside of these constraints it automatically adjusts itself a % or two accordingly to the point where the problem takes care of itself.

A hyperbolic example would be if there came to be a card that read, "1 mana: deal 30 damage to the enemy hero" the card would eventually be so reduced in draw rate you had a 1% chance of even drawing it, making it a pretty much useless card (not counting tutoring ofc). By this means, Skull of Guldan doesnt have to be nerfed into oblivion it just doesn't show up quite as much.

As far as I am concerned this is the natural remedy to this ongoing problem and they certainly have the brains to whip up a method of implementing this and everybody wins: people who want to play the broken cards can still play them in all their glory and people who hate them get to see them less often.

* number was originally 15% which admittedly is way too high

This would also be a good idea for keeping a balanced meta- just readjust draw rates! All the high impact cards are on the bottom of the deck now...

Sorry, I don't think that idea would work. Hearthstone is also a game of deck building and card collection, and using strong cards in a well-thought-out deck should be rewarded. Good deckbuilders should on average beat bad ones. An automatic system penalizing cards with high win rates would warp the meta, by prolonging games, and adapted deck building to have a guaranteed early game by including some cheap cards of middling quality and lots of high impact cards for the bottom of the deck. Which could actually lead to some new and interesting decks abusing the meta!

For a while, if those decks would get too strong, the algorithm would adjust their win rate. This constant adjustments would make most decks to be of similar strength, which would result in a different game where quality decks mean little.

Last point: transparency. A system that works against (some) players would not be well received, especially if it's hard to understand. 

Even if this can be done transparently it will never happen simply because it makes games more about highrolling rather than skill. Team5 have stated on multiple occasions that they would prefer cards to not have that 'draw this and win' quality, and assigning draw rates to certain overpowered cards will not change the fact that games are won because one player managed to get lucky. To use your example of a 1% draw rate of a 30 damage dealing card. Will you, or anyone else in fact, be content understanding that you've done everything right in the game only to lose because your opponent managed to draw that 1 percenter card?

It will also accomplish very little at the end of the day. One card rarely makes the whole deck. Decks that rely on specific cards to win are usually combo decks which already suffers from draw rng and consistency as a whole. And then there are decks that plays cheap costed cards which are otherwise average in strength and value, but are more consistent than other decks. These are your average tempo aggro deck. No adjustments are needed. If you want your deck to be more consistent in tempo, then play a deck with more 1-2 drops in it. If a specific set of cards is the key, then the deck needs more draw cards to draw them. Why make silly adjustments on backend when we can do it ourselves to fit the purpose of the deck.

Besides which, how does one really define 'over-performing'? Most play win rates of cards tend to be around 10-15% more than the lowest win rate cards in a deck, and even then these rarely tell the whole story, since reactive cards don't usually have high win rates. Something like Mindrender Illucia can easily win a game at times, but her winrate is abysmal. Cards like Envoy Rustwix have middling win rates but can hardly be considered unimpactful. And these win rates tend to change according to deck build, and the pilot behind the wheel. High legend players are always going to be better at handling cards than normal players like us; which set of data should team5 then use when considering this exercise?

Reply obviously did not grasp the concept of post judging by text. It doesn't put the cards at the bottom of the deck. You have to understand statistics being that if you reduce a card's draw rate by even a % you affect its win rate ever so slightly. Do your research and look at draw win rates at hsreplay. 

The general idea is if a card is leading to many wins nip it in the bud by adjusting it's frequency. I wrote 15% above but more realistically it would be smaller, like 5%. We can't expect every card Bliz releases to be perfectly balanced. That's impossible. So do the obvious thing and at least make them show up just a little less often.

It isn't about the one game: it's about many games in a row. You shouldn't be able to win game after game because the overpowered cards keep showing up.

I don't like this idea. This feels like a very sneaky, and unfair thing to do. In any card game, the point of shuffling cards is so that every card has an equal chance to be at any spot in the deck. By reducing the draw rate of specific cards blizzard is effectively stacking the cards, in other words, cheating.

In games like this there are rules, and expectations. Expectations can be broken by specific cards, rules are concrete. Something like all minions must wait a turn to attack is not a rule it is an expectation since it is usually true but some minions can attack immediately. But something like all minions dying when their life reaches 0 is a rule since it is core to the fundamental nature of the mechanics of the game. Cards keeping their current text/mana cost/attack/health is an expectation since blizzard is able to change those thins at their discretion. The cards being shuffled fairly at the beginning of a match is a rule. It shouldn't be changed. And even if it was changed, I doubt it would fix the game.

If a card is OP then it is just OP regardless of whether or not you draw it. If you fight a deck and they don't draw their OP card because of this new balancing system then blizzard could have just nerfed that card so that it was unplayable and that match would have gone the same. But if your opponent does draw the card then it is still a poorly balanced card, and the knowledge that your opponent had a smaller chance to draw that card won't make the loss feel any better.

In short, adjusting the draw rate of cards is a lazy way to fix a problem, and it would just be healthier for the game overall if unbalanced cards were balanced instead of just being pushed away into a corner.

The only feasible way to implement something close to this would be to add a “prevalence: X” stat to cards (like health and attack) that increases its odds to be drawn by X times. That honestly might not even be a bad idea because it would allow for Blizzard to make cards that are integral to their archetype more likely to be drawn (for example, Lucentbar could have a very high prevalence to make Druid Heal decks actually functional ). Given that players only get 1 copy of Legendaries, that might actually be a good way to make niche payoff Legendaries a little better. A high prevalence 1 drop would also be able to help control decks reliably get a turn 1 play without sacrificing their deck quality too much. That being said, all this seems like it would be better as an Expansion keyword instead of a rule change to every single card. 

I think this is a really bad idea. You're not fixing the problem that the card is "overperforming". And even if a card has a 1% draw chance, then it doesn't change the fact that in that 1% of games, basically nothing has changed. 

And besides, a single card doesn't make the deck unless it's a combo deck or a build-around card like Baku the Mooneater. If you lose every game you don't draw the specific overpowered card in your deck, then it's a bad deck.

Also, shouldn't players be rewarded for playing the right cards in the right decks? Skull of the Man'ari and Voidlord are both awful cards on their own, but they become strong when played together in a deck like Cubelock. Isn't that the entire purpose of a deck-building game? It's the ideal example of really, really good card design: cards that are bad on their own, but when played in the right decks, the deck is much greater than the sum of its parts. Reducing the draw rate just disincentivises people from making good decks, because why play good cards if you never draw them?

TL;DR, it's a lazy solution that doesn't fix any problems, and probably introduces some as well. If a card is unbalanced, why not make the obvious choice and balance the card?

Ok, the people have spoken. It's the perfect solution to solve imbalanced cards and allow a little more variety into the game, but I guess not such a good idea. Have fun net-decking the same boring decks that involve cards that carry the win for you like the good little robots that you are : D

Reading your replies you guys seem to think I'm saying to make it the card won't show up at all. Not the case. They would show up  less often in a series of games, thus taking down the winrate of deck abusing power cards over time.

I think you’re misunderstanding the issue people have with this idea. I think it’s fair to say that this change would lead to better game balance and deck variety, but it comes at a tremendous cost. A mechanic like this is literally the most unintuitive thing I’ve heard, on par with maybe manipulating matchmaking to reduce winrate. Imagine that you really like playing one specific deck, and you learn how probable each card is to be drawn. Then the next week you play the deck again, but you have to look up the new probabilities to play the deck well. Rather than being able to play HS as a casual game whenever you have time, you instead have to either constantly look up stats or end up frustrated because certain cards don’t show up when they should. A lot of people play HS casually and enjoy coming back to the same deck again and again without having to put in tons of effort to relearn the deck every time. For that kind of player, a perfectly balanced meta and deck diversity don’t really matter because they only play a few games a week. By making this change you’d help the 1% of heavily invested players while alienating the 99% who just want to play a decently balanced game. From a player satisfaction standpoint this change has huge downsides with an upside that isn’t actually that impactful for most players.

Quote From Live4vrRdieTryn

Ok, the people have spoken. It's the perfect solution to solve imbalanced cards and allow a little more variety into the game, but I guess not such a good idea. Have fun net-decking the same boring decks that involve cards that carry the win for you like the good little robots that you are : D

 

Reading your replies you guys seem to think I'm saying to make it the card won't show up at all. Not the case. They would show up  less often in a series of games, thus taking down the winrate of deck abusing power cards over time.

Way to take criticism like a champ (not).

Clearly it's not the perfect solution if everyone in this thread disagrees with it, eh?

It's just such a convoluted and unfun way to try and solve a problem that literally every card game, could even say every game bar something like chess, has to deal with. I don't even know where to begin honestly.

Heartstone thrives off of its simplicity and how appealing it is for casual players. Do you really think hidden, altered draw rates based on a card's strength are going to help in that? Even more so when you consider that most decks that rely on one card strategies are often Timmy decks, which are preferred by the casual audience (big priest and Tickatus warlock for example). They'd literally be breaking down the formula that has made Hearthstone the juggernaut of digital card games that it is today.

And it's even more ridiculous when you consider that their balancing, and I never thought I'd say this, has actually gotten a lot better recently. With both the current and previous expansion we ended on a very varied and, I'd say, good meta. Blizzard have actually been doing really well in this regard, excluding evolve shaman.

Above meta talk obviously applies to standard. Wild has never received attention so it's no wonder the meta over there is as bad as it is.

Quote From Live4vrRdieTryn

Ok, the people have spoken. It's the perfect solution to solve imbalanced cards and allow a little more variety into the game, but I guess not such a good idea. Have fun net-decking the same boring decks that involve cards that carry the win for you like the good little robots that you are : D

 

Reading your replies you guys seem to think I'm saying to make it the card won't show up at all. Not the case. They would show up  less often in a series of games, thus taking down the winrate of deck abusing power cards over time.

Your tone here reminds me of your recent thread where you insisted (without proof) that the pity timer was broken, and everyone told you that you were probably wrong based on all the existing data, so you got mad at everyone.

Your idea of changing card draw odds is not some "nerfing method everyone is oblivious to" - it's just a terrible idea, and that's why people are criticizing it.

The first major problem is a problem of implementation. How do you determine whether a card is OP? There are lots of ways you might imagine doing this, since "OP" is somewhat subjective, but you seem to think the answer to that is to judge based on the "drawn win rate." That's simple and feels good, but it ignores the fact that "drawn win rate" will be impacted by this algorithm because "drawn win rate" depends a lot on - you guessed it - drawing the cards in question. In that way, this is sort of a classic case of Goodhart's Law (when a measure becomes a target, it ceases to be a good measure). Frankly, there's no way to determine the power level of a card that won't suffer from this problem precisely because you're manipulating draw probabilities. If a card is good its draw odds will drop, and once its odds drop it will be worse, which will in turn drive its odds up. What you end up with is just a mess of probabilities that go up and down arbitrarily.

The second major problem is transparency. Today you know that if you have N cards left in your deck, you have a 1/N chance of drawing any given card. This proposal would make it impossible to know what the odds were of drawing a card because those odds would be constantly changing based on the current prevalence of a given card across all Hearthstone games. Even if Blizzard was explicit about how they calculated these values, you wouldn't have the data needed to determine the true odds. But knowing the odds of drawing specific cards is a big part of "playing to your outs," (that is, it's critical to the skill ceiling of Hearthstone - and CCGs generally, for that matter). If your choices are to try and draw into some burst damage or trade on board and stall your opponent, you need to know the odds of drawing your burn cards in order to make a decision that maximizes odds of winning.

The third issue relates to "cast when drawn" cards. They have no "drawn win rate" in any meaningful sense, but how they're distributed in your deck is super important. For example, this algorithm could result in ruining the Soul Demon Hunter archetype by driving an unusually high number of Soul Fragments to be drawn before you can play the payoff cards. In Wild, you could end up with strange play patterns around The Darkness because of how those cards would interact with your opponent's existing draw odds. And Bomb Warrior would be crazy under this new system.

Normal people: "Draw RNG can be so polarizing that it sometimes feels like the game is downright rigged."

This guy: "BLIZZARD SHOULD ACTIVELY RIG THE GAME TO MANIPULATE WINRATES."

Obsequious disagree with whatever new idea is being brought up replies. You guys already proved you didn't understand my post by your replies so I'm just gonna quit trying to convince you. Have fun seeing the same cards over and over again and playing with just 5% of your collection. Say what you want about me, but that fact says a lot about you.

I read your replies and they aren't even in response to what I'm proposing. Every one is just jumping down my throat at something I didn't even say. Just a lot of TLDR to make up for no point at all except, "i dont like the sound of this idea".

Then explain your idea. If eight people are willing to engage you in a discussion - and all in a civil tone - you may want to actually answer them and clarify your assumptions and ideas before dismissing everyone rudely.

I think people were thinking that I was saying that cards with huge win-when-drawn rates should basically have little chance at all of appearing when put in a deck, cards like Nitroboost Poison, Secret Passage, and Skull of Gul'dan. Admittedly I arbitrarily used 15% in my example which maybe lent to those extreme opinions (i'll change it now).

I think a better way would be if a card has a winrate when drawn of 54% it should be drawn 4% less, or even 2%. That reduction is very small you must understand and would only be a factor in games over time. It'd be a soft nerf that could be implemented universally instead of having to sit through meta after meta where one new card appears that everybody hates unless they are abusing it.

This way people that want to play it can still play it and people that don't like to see it see it a little less over time. It's a concept to lessen the state of being forced to "join them or lose them" which is pretty much every meta where a few cards are released a little overtuned and we are captive for sometimes months losing to them before they are nerfed, which usually destroys the card alltogether.

Again, this strategy already has success in Arena where outperforming classes have less likely drafts of higher winrate cards.

To my computational physicist's mind I admit I do find it a very elegant solution that lets you tune a card/deck's power on a continuous scale without being subject to the discrete jumps of changing small integers (e.g. a card's cost). But underneath it's clean exterior I don't think it is actually any simpler or more effective than conventional nerfs are.

For one thing, we'd still have to wait just as long for changes to be made. It's not like the devs are going to tweak these numbers every day. No, they're going to wait a while and see how they perform. And when the numbers are tweaked, do players get to dust cards for full refund? A card is objectively weaker if its draw rate is reduced after all.

It also has a negative effect on player agency: Players cannot do the calculation of how likely they will be to draw a specific card, because the proper calculation will have to include the draw rates of every card still in the deck. It's an easy calculation for a computer, but too laborious for a human to do in the middle of their turn.

It's also not as helpful to deck diversity as you might think. Unless you want to maintain multiple different draw rates for single cards based on what deck they are in (a very difficult thing to define btw), you'll end up reducing the power of decks you weren't trying to nerf, with much the same effect as a conventional nerf would have had.

I could go on and raise similar concerns to other people, but I'll stop here. As I said, I recognise it is an elegant solution, and I know it could be made to work in principle, but in practice I don't think the benefits outweigh the complications. That's especially true when I believe the players themselves are the bigger problem than the cards most of the time. No matter what you do, Spikes will latch onto a small handful of decks and create a meta that sooner or later you'll get fed up with. Occasionally a Tier S deck comes along and you can blame the HS team for the meta, but most of the time the number of good decks is much larger than the number of common ones, and you can only blame the players for that.

hard to overstate how bad this idea is.

Quote From Live4vrRdieTryn

Again, this strategy already has success in Arena where outperforming classes have less likely drafts of higher winrate cards.

I think I've already responded pretty fully to your comments about how this would work, but this probably merits further discussion.

This comparison is very bad. All of the systems in place for Arena that do this exist to make drafting more consequential (e.g. the bucket system that uses win rate and pick rate data to group cards and avoid "instant pick" situations where two out of three cards are trash). 

Presumably you mean to draw a parallel with your proposal and Blizzard's tinkering with the odds of seeing certain cards in these drafting pools. But there's no proper corollary for that in constructed because the contents of the deck are not determined randomly*. The closest you could get would be something like "At the start of the game, for each card C, you have some probability P(C) that C gets replaced with a random card, where P(C) depends on the win rate of C." But that's obviously terrible.

It makes sense to use these kinds of win rate statistics to ensure a balanced playing field for drafting because players want to feel like their choices matter when they draft their decks; playing Arena should never feel like you're just playing Hearthstone with a random pile of 30 cards. This is basically the same reason Blizzard prints lots of different tech cards to deploy in constructed dependent on the meta - they want you to feel like your deck-building choices matter, and that smart choices will yield better outcomes when you finally play the game. It makes no sense to do this for the game itself, as it would simply punish players for using good cards and make their deck-building/drafting choices less impactful by virtue of making it harder for players to draw the cards they chose to play.

*Technically you could have the deck building randomly fill in your deck, but you'd still be choosing to do that, so it's not exactly random.

Or better yet they have more then a singular pay-off to what you're doing, or have the thing you're doing be something that is inherently good and that you don't necessarily need a pay-off for.