Cognitive Biases in Hearthstone – Hindsight & Outcome Biases (#2)

The human brain is a wonderful thing. Sometimes, when presented with two choices, which are the same, just worded differently, it will assume that one option is better than the other. Other times, when you don’t have enough information, it will fill the gaps itself (often incorrectly). It looks for correlations, even if there aren’t any. Or leads you to situations in which something just FEELS right, even though it’s really not.

Believe it or not, but cognitive biases aren’t something rare. To put it simply, they’re common flaws in logic. Person’s own, subjective interpretation of reality. Of course, after you really start thinking about them, you realize that they make no sense. But what’s important is that they affect everyone – like you and me – in our daily lives.

In this series, I will cover some of the common cognitive biases that can affect Hearthstone players in particular. How do they work? Why do they happen? Are there any situations in which they actually make sense? Identifying them and realizing what they are is a big step in terms of becoming a better player. Plus some of them are just interesting to read about.

Check out the previous installment in the series:

Cognitive Biases in Hearthstone – Gambler’s Fallacy (#1)

What are Hindsight & Outcome Biases?

I’ve decided to write about these two together. While they aren’t exactly the same bias, they’re somewhat connected to each other. It wouldn’t make much sense to cover them separately. But first of all, let me start by explaining what those biases are.

Hindsight bias is also known as the “knew-it-all-along effect”, and that alone describes it quite well. It’s a belief that an event, usually something completely random or at least out of your control, was predictable, even though it wasn’t. But it’s often more than just believing it – the bias can actually affect that person’s recollection of the entire situation. Yes, it can alter your memories, making you think that there were some signs, it just HAD to happen, or that you indeed were convinced beforehand that this event will (or won’t) occur. In reality, there was no way you could have known. But that’s still not the most interesting part of this bias.

Funnily, it usually occurs when you weren’t sure that something is going to happen. Even if you try to predict something, you have this creeping feeling of uncertainty somewhere in the back of your head. After all, it’s something you cannot reliably predict. However, if that thing does happen, your mind tricks you into thinking that you – indeed – predicted it. You often forget about the uncertainty, which accompanied the initial thought process.

Outcome bias is slightly different, and it’s even more simple. It happens when you judge a certain decision based on the final outcome. In hindsight (hence the similarities), instead of looking at the quality of the decision itself, you tend to look at the quality of outcome it produced. Let’s say that doing a certain thing had a 90% chance of a good outcome (make it winning $100) and 10% chance of a bad outcome (losing $100). While it obviously depends on some other factors (like whether you can afford to lose that money in case that happens), it generally seems like a great decision to make.

However, if you went for it and got a bad outcome, you will tend to look back on this decision thinking that you shouldn’t have done it and that maybe it was a bad decision after all. It also works the other way around – if you make a terrible decision and it somehow works out, you might even start thinking that it was a great decision in the first place. Of course – outcome is not something that you know beforehand. At the time you’re making a choice, you can’t take it into account, which means that logically your decisions should not be based on what happened after you made them.

Examples of Hindsight & Outcome Biases

Both of those biases affect us more than you might think they do. Even in our daily lives. Next time one of your friends boasts that “he knew” that something is going to happen, that’s hindsight bias. Or when someone will talk about this poor choice he made – poor because it didn’t work, not poor because the decision itself was bad. To make you understand everything better, here are some examples of those biases at work:

Let’s say that you hurry up in the morning and you forgot your umbrella. While nothing suggests that, you think to yourself that “it might rain in the afternoon” – you are not certain of it, it’s just a thought. And if it does not rain – it stays as some random thought you had that you will forget very quickly. However, if it does indeed rain, you might think “I just knew that it’s going to rain”. That’s hindsight bias.

When you’re packing for a trip, you often have this feeling that you are going to forget something. Of course, most of the time it doesn’t happen (and if it does, get better at packing, man) – and that’s it, you usually don’t give those feelings a second thought after it turns out that you took everything. But when it finally does happen, you start thinking that you KNEW that you will forget something. That’s also hindsight bias.

The best real life example of outcome bias would probably be buying lottery tickets. Let’s say that you’re struggling in your life and need money. You spend your entire savings – $1,000 on lottery tickets and somehow you manage to win $1,000,000. You might think that it was the best decision you made in your life, but it’s actually quite the opposite. The decision itself was terrible. Your chance to win was abysmal, and the majority of the time you would lose your savings – money that you absolutely needed. That’s outcome bias – even if you make a very poor decision, if that decision produces a good outcome, your mind might fool you into thinking that the DECISION itself was good.

While it’s hopefully not something that will happen to you in real life, another way to illustrate outcome bias would be going back to the initial experiment that has proven its existence. Two psychologists – Jonathan Baron and John Hershey – conducted it in 1988. They’ve presented subjects with hypothetical situations, but the most notable one was rating the decision of a surgeon to perform a high risk surgery. The surgery had a known rate of success. Subjects were also given the outcome of the surgery and asked whether the surgeon’s decision to perform it was correct or not. Logically, the outcome should not be relevant when grading the decision, but it was quite the opposite. Majority of subjects didn’t judge it based on the success chance, but on whether it was or wasn’t successful. If the person died – it was a bad decision, and if he survived – it was a good one. It’s pretty grim, but also a good way to show how outcome bias works.

Hindsight Bias in Hearthstone

Both of those biases revolve around looking in the past and analyzing the things you thought or the choices you made. It’s not surprising that those biases are so important in Hearthstone and can affect hundreds of thousands of players every single day.

I’d say that hindsight bias is less damaging when it comes to the game itself – it’s more important in the social aspect, such as discussion forums, as well as creating Hearthstone-related content. If you want to see a perfect example of this bias in action, just tune into some tournament or at least a big streamer and read Twitch chat. The amount of people who “knew” that something is going to happen or “called it” is astounding.

Hindsight bias probably affects my line of work more than the average Hearthstone player. When analyzing the game, especially the past events, I often catch myself writing that I knew that something is going to happen. “I knew player X will get punished for being too greedy”, or “I knew that Blizzard is going to nerf this card”. In reality, however, I didn’t “know” anything. I made an educated guess, but that’s still only a guess.

Rating cards is another area in which hindsight bias is incredibly common. I’ve written card reviews ever since The Grand Tournament expansion, and I read (or watched) thousands of reviews from other players. Believe it or not, but it’s incredibly difficult to rate cards correctly. You have a very limited amount of information available. Early during the reveal season, you don’t know other cards that will also be released. Even when the full set is revealed, you still don’t know whether the meta will be fast or slow, which classes will end up being popular etc. You also don’t know what old cards might suddenly see play when combined with the new ones. There are so many variables that you can’t take into account. So, in the end, pre-expansion card reviews/ratings are simply predictions. “Educated guesses”, like I’ve mentioned before. There is no way to be sure whether a certain card will be good or not. Sometimes the highest rated cards turned out to be bad (Countess Ashmore, to not look far – I still think that the card has potential, but it’s bad in the current meta without a doubt), and the lowest rated ones turned out to be strong (such as Rin, the First Disciple, for example).

However, not so long after the expansion’s release, social medias, streams, chats, even casual discussions with your friends are simply flooded with something that you could only call hindsight bias. It does not matter that 2/3 of the predictions were wrong (and there’s nothing to feel bad about, like I’ve said, it’s really hard), those who got some obscure card right will surely tell everyone that they knew that it’s going to work. But even worse is the self-righteous attitude of some users, who will surely make comments such as “LOL! How could you possibly not know that this card is going to suck? It was so obvious!” (yes, I’ve seen it many times already) Before the expansion launches, no one is really certain how strong each card will be, especially the more controversial ones. But after the expansion, it seems like so many people knew everything all along.

But reviewing cards is not the only situation in which it happens. You should still be aware of that bias when playing the game. Remember that the majority of events in Hearthstone are random, and thus unpredictable. There is no way you knew that your opponent is going to top deck that AoE card. You might have suspected that, considered that as one of the options, but you did NOT know that. Why is it important? Because it will make it easier to objectively analyze your plays. Treat random situations as exactly what they are – random situations, not something you could have predicted or – even worse – had some control over (this one is called “illusion of control” and it’s another cognitive bias).

Outcome Bias in Hearthstone

When talking about analyzing your play, outcome bias plays a much bigger role. One of the main ways to improve your gameplay is looking at your past decisions and judging whether they were correct or not. If it was correct – you want to repeat it, or similar decisions, in the future. If it wasn’t – try to not do the same thing in your upcoming matches. So, you should clearly see how outcome bias might affect that.

Let’s start with an example. You play against a Hunter and you’re relatively low (let’s say 14 health), but you’re still out of range of most of their burn cards. The only way this Hunter deck can kill you is Unleash the Hounds + 2x Kill Command + Candleshot, after he played one Candleshot already. The chances of him having this combination of cards is low, as he drew only half of his deck. You can either play around it by not playing any more minions (you have two on the board), or drop another big minion, which would play into that combo, but make your board significantly better and put Hunter on a faster clock if he can’t kill it. You decide to drop that big minion and not play around that combo, but it turns out that Hunter has it and wins the game. Outcome bias might make it very difficult to analyze this play, especially right after you got punished. Instead of thinking about the choice and the chance of success, you think about what happened – you lost the game. A LOT of players would be quick to call it a bad decision, or a misplay, while in reality, I’m confident that it was the correct play unless you had some psychic-level reads. You can go ahead and analyze the chances of Hunter having those four cards in his hand, but I’m sure that playing around it would be a massive mistake. Even if you wouldn’t die immediately, you would give Hunter more time (he would put you in lethal range next turn with his Hero Power alone) and make it easier to come back on the board.

But what’s the problem? It doesn’t really matter if you think that this decision was correct, while it wasn’t, right? Well, wrong. If you think that a certain decision is correct, you’re more likely to repeat it in the future. Maybe when a similar situation pops out next time – you will face a small risk of lethal – you might go for a sub-optimal play, because there is a small chance that your opponent will kill you otherwise, repeating the same mistake.

What you should do instead is simply ignore the outcome. Outcome is irrelevant and there is a very easy way to illustrate it. If you could play a 0 mana card that always starts in your hand, that has 40% chance to win you the game on the spot and 60% to lose it, would you run it? Assuming that you don’t get any gold for those wins (you only gain/lose stars), it seems like a terrible deal, right? Anyone will tell you that. Even if you end up winning 5 times in a row, it doesn’t make your decisions to run this card correct, because mathematically it will do you more harm than good. You can look at every decision in a similar fashion. While it’s usually impossible to assess a single decision’s impact on the entire game precisely, you should focus on (usually estimated) chance of success and not on what happened after.

Another thing I should mention is that making the correct decision is not as easy as picking the more likely outcome. Risk assessment is another very important skill. You need to understand how many risks you can take in a certain matchup/situation and how taking a risk now might increase your win chance in the context of the entire game. I mention it because I know some players, who always assume that a play that has a higher chance of short-terms success is the correct one, while in reality, a risky play (e.g. taking a 40% chance) might be necessary to get out of certain situations or win a bad matchup. Those plays are much harder to analyze and are also affected by outcome bias.

When it comes to outcome bias and risky plays, I’d say that it’s something a lot of Hearthstone casters are guilty of. While I enjoy watching tournaments, it sometimes hurts when casters put too much attention into the outcome instead of analyzing the choice itself. Casters are also working with more knowledge than the players, so it’s easier for them to be biased. Another quick example to illustrate this issue – you play a proactive, Midrange deck against a Big Spell Mage and you build a board of 3-4 health minions, playing right into Flamestrike. But you need to do it, because taking a slow approach is not an option in this matchup. When listening to casters’ commentary on such plays (not exactly THIS one, but taking a risk of not playing against X card that your opponent is quite likely to be holding), I often feel like they concentrate on the outcome way too much. I’ve heard the same casters praising a calculated risk when it paid off in one match, and then questioning it in another, even though both situations were very similar. And that’s exactly the opposite of how you should analyze the line of play, which is one of the main jobs of casters.

And finally, I have some tips on how you can avoid outcome bias when analyzing your plays. Most importantly, you need to realize that such a bias exists – it’s the biggest step towards not being affected by it (or any other bias for that matter). Additionally, try to not analyze your plays right after seeing the outcome. I should probably say it about your matches in general – from my own experience, analyzing a game right after it’s over is not a great idea. This bias is incredibly easy to notice when you’re thinking critically. However, right after a close win or after your opponent has topdecked lethal, you will most likely be more emotional, and emotions aren’t exactly on their way with logic. If you wait some time, even a few minutes, you will become more emotionally detached from that given match. That’s where replays are handy – either recording your own games or simply using the replay tool built into Hearthstone Deck Tracker is a great way to look at the past matches. Go back to the play you aren’t sure about and try to look at every option you had. Don’t think about whether the play worked out or not in the end – just go back, analyze the play and play alone. Look at your potential options, try to figure out whether the choice was correct. Instead of looking at your own replays, you can also try to analyze the plays pros are making (on streams or during tournaments), but still looking only at the plays, no matter if it worked out or not. Pros also usually explain their reasoning in such moments (when they’re streaming, not in the tournament setting, obviously) – e.g. “it was the correct play, because X and Y, but it failed, because opponent had Z, which was unlikely”. That’s the part you should be paying most attention to, not to the Twitch chat spamming “misplay” or “you shouldn’t have done that”.

Closing

That’s all folks. While hindsight & outcome biases might not affect how you play the game directly, they’re incredibly important to understand and not fall for. Learning from your previous plays is one of the most important things if you want to improve. Not properly recognizing them (especially the outcome bias) might not only stop your advance, but make you a worse player than you were before. If you make a bad play and get rewarded with a good outcome, the outcome bias might force you into thinking that it was the correct play all along, and you might repeat it in the future. Bad habits are easy to make and hard to break, so try your best to not fall for those biases.

If you have any questions, or want to add some more examples, be sure to leave a comment below.

Also, please keep in mind that I’m no psychologist or sociologist, so I’m mostly basing my definitions on well-known online sources, such as wikipedia. If I made a mistake somewhere, was wrong about something etc. please let me know.

Good luck on the ladder and until next time!