The problem with economists calling such choices “irrational” is that they’re assuming a lot of things, like the ability to reiterate at no cost, which isn’t realistic. The example he used at the beginning with the volunteer shows this: if he loses the $20 wager, he doesn’t eat lunch. If he wins, he gets $40-so what? So he can buy two lunches? He doesn’t need two lunches to prevent going hungry. So his choice to reject the 50/50 offer is absolutely rational from an evolutionary psychology POV. Ignoring real world needs like this is one silly thing economists often do. Another example: would you wager your car if you need it to get to work and keep your job? Who cares if you win a second car, if you’re risking your job to do so? This is okay in board games of course because there’s nothing at stake related to actual survival.

When he's talking about tracking, the best part was me having removed that emotional evaluation from my brain after years of competitive magic. It's the same reason people hate mill in magic, but as a competitive player, you learn the actual benefit of cards like tracking and the almost always insignificant loss that comes with mill.

Holy crap this is one of the most useful things I've ever heard. Not just in game design.

One of the best GDC talks I've ever watched. It reminded me of how in Potion Explosion, nobody ever wants to use the -2 point tokens that allow them to take an extra marble on their turn because nobody likes losing points. But taking an extra marble can be game changing.

TL;DR loss aversion takeaway 8:05 Framing 20:00 [Player improvement] 29:00 Regret & competence 31:25 Endowment effect 37:00 Q & A 43:30 Personal comment: when I played FF, loss aversion was a thing for me; I would grind through game without using alot of the good stuff unless it was for a big boss fight I thought I would need it for. Thanks for the video.

Geoff's "Achievement Relocked", which goes deeper into loss aversion and it's applications to game design, is one of the most fascinating books I've read.

Another note about level draining in DND; not only was it emotionally devastating, it also was a pain in the ass when it came to calculating character stats, especially in a crunchy game like 3rd edition. I think that's a bigger reason why it's been removed. It never felt quite as bad in the video game adaptations like Neverwinter Nights

18:00 "A deadly disease is rapidly spreading." Little did he know he was predicting the future.

This was a deeply fascinating talk. It's gotten me to think a lot more about how I practice loss aversion in ways that aren't always logical.

12:29 level onDraining - What a lot of people don't realise is that in early DnD combat was to be avoided - a lot of game was about NOT having physical conflict as it was so dangerous. Undead were TERRIFYING as they could Drain Levels! The worst thing a monster could do was to take the precious XP...HOWEVER also in early DnD Clerics were amazing against undead. They were a 'hard counter', and so the reason Turn Dead was sopowerful was because Undead were so powerful. Gygax was a clever guy

One comment on the Level draining ability, especially in 3E: I'm more uncomfortable with recalculating all those modifiers, remembering which feat was gained at which level or for spellcasters redoing the entire spell selection (on a Sorcerer that's even worse) and basically having to redo the entire sheet from the ground up is faster. BUT you can technically exploit this for respecing, which is standard starting from PF so the only useful role is lost. Speaking from experience that wasn't drained

Wow, that was an amazing talk on the topic

Great talk about of a lot of important psychological principles applicable to game design!

On the comment of player vs player compared to player vs game for emotion in the QA, my thoughts... Players will always associate loss to the game unless it is a player choice. Let's say your playing a game and you have to gain victory points. Your opponent draws a card that says "opponent loses 1 victory point." So you lose a point, whatever, it's on the game. Now let's say that card said "Gain 1 victory point or remove 1 opponent victory point." If they remove one of your victory points, it will feel a bit more personal only because it wasn't the game that did it to you but the player's choice. However it is still primarily on the game since either way, it creates a 1 point difference. Granted the player always had a choice. I mean he could just not used the card, but there is an 'expected loss.' I mean players, especially competitive ones, will feel more upset if your opponent didn't use the card at all. Why? Seeing an opponent detriment themself is also upsetting because it starts to invalidate your win. Let's take this more extreme. The card says "gain 2 victory points or an opponent loses 1." And they chose to let you lose a point? Oh now there is a lot more on the opponent for that loss because the opponent made a choice that was bad for them and worse for you. Having players sacrifice their own gain to shut you down hurts far more and takes entirely away from the game.

In the casino part of the mobile game I'm playing, you have to play Double Up 10 times in a row for the most efficient earnings. You always choose Low or High depending on which one is the most probable, but when you stand on a 8, it's 50/50. I got so sick of always choosing the wrong choice, that I litterally just decided to ALWAYS pick High on a 8, regardless of what my guts feelings tell me. Feels much better that way.

Great talk! Enjoyed it a lot.

This was a fantastically interesting watch

Great topic and presentation! Good on you!

Great talk!!

Tracking would be a staple in every deck in Yu-Gi-Oh!

What a fantastic lesson!

1 of the best lectures that l have never attended!

It's extremely interesting, I wish I could access more study about it. I will look for them but do you have any references named in the video?

Great talk. Interesting about Catan as an example for endowed progress. You start with two victory points because villages are worth 1VP and you start with two, so you aren't just endowed with 2VP for the sake of 20% progress as manipulation to feel invested, it's more for narrative effect that your people have just settled on the island and have a base to begin expansion. The game goes up to 10 instead of 8 for easier score counting. Getting 2 free villages and roads at the start definitely creates the effect of endowed progress but I don't believe it was designed with that in mind.

I think this explains a lot about the community reaction to Mojang not adding glow bugs... or showing their concept art for birch forests.. the only difference being fungi on trees and taller trees. the taller trees were added in world gen... but they said it was just concept art and not a promise.... ppl felt they had had something taken away, and focussed on that more than what they had gained

Really good talk, but at the questions segment I perked up because "wait a minute? Was that Pip from SUSD podcasts?"

50:00 Is that Pip from Shut Up & Sit Down?

"so you are telling me there is a chance ?"

Immediately my math-brain begins to consider which are the _rational_ decisions, and then I stop, remind myself of his warning not to over-analyze my options, and concede; I'm shamefully forced to admit that-by default-I'll willingly decide to decide exactly the same way everybody else does. Even though I know better, give me a safe gain and a gamble to avoid loss every time!

Fascinating talk, applicable to areas other than board games. Are casinos and online gambling companies applying these techniques? The odds are already stacked in their favour!

Really great talk. Thanks!

15:00 CORRECTION: Values range from one _PENNY_ ($.01) to 1 million($1,000,000).

"a singed companion cube comes into the last scene of the game" Did Geoff just predict Alyx?

29:15 Not really, not if people can invest money in some way and get profits from it before they have to give money back. It could be pretty interesting idea for the game - where you have basically predict your future capabilities ;) I know that it is pretty obvious but it is worth pointing it out that you cant use this trick for a long term, or if players have free access to a ways of spending money.

OMG. I played HS for a five years and never thought that Tracking removes other two cards from the deck. Silly me. XDDD

If i had to chose someone randomly, i would also pick Rob Daviau 😊

31:35 Does anyone know what this study is called? I can't find it and I find it hard to believe that 10% of people still won't switch if it's 10x the amount Also I don't think that the advice on sticking with your gut feeling if you're not sure to change, is wholly because of this. I think that there is merit in that advice because you could be overthinking things

Is positive interaction(Race for the Galaxy, Mottainai, Keyflower) processed that much differently from negative take that(even in cases that are mathematically similar)?

about tracking discarding the 2 other card is actually much better than keeping them in your deck. Playing a 28 card deck is stronger than playing a 30 card deck.That is the reason why there is no « 0 mana draw a card do nothing » card. It would be auto include in every deck. I didn’t like tracking when i first played the card. But when you stats building becks, you realize that the 29th and 30th card are the worst cards of your deck, and that it would be stronger to have a 28 cards deck.

Talking about Loss Aversion and not even mentioned WoW's "rested" mechanic? Basically they created mechanic where if you farmed a lot you became "unrested" and got 50% exp penalty. Noone liked it. Then they flipped it and made penalty exp standard, and made rested "you get 100% more experience" - everyone loved it.

Regarding the experiment at 30:00 : The participant's unwillingness to switch to the other cup seems kind of rational to me, if you factor in information asymmetry and trustworthiness of the experimenter. After all, the experimenter might only give the opportunity to switch to participants who made the correct choice in the first place, enticing them with more and more money until they switch - and lose.

Meanwhile, Yu-Gi-Oh! players: "Lose 10 cards from your Deck to draw 2? Good deal!" (Pot of Desires)

You're playing deal or no deal, there are two boxes left. One box contains £1 million and the other a penny. The phone rings, it’s the banker, he offers you half a million pounds do you say no because it's half a penny less than the average value of the boxes?

Something I've noticed that no one really comments upon is that opportunity cost isn't a real cost when it involves an infinite resource. How can it be seen as a lost opportunity when by all logic it's still there? With deals involving the infinite resource against a finite resource are struck, loss only works in one direction. How do trade-offs work with one-sided post-scarcity?

This is very interesting.

I feel like I am one of the 10% outliers in the regret section. I'm constantly 2nd guessing myself and would jump at the opportunity to switch because it would feel worse to have that fear validated.

With casino chips being abstracted from money, it takes that a step further with digital chips in online casinos.

Losses are also mathematically more significant than wins. Consider a base of $10. A loss of $5 sets you at $5 which is 1:2 ratio A win of $5 sets you at $15 which is 2:3

To the last question about rewards/lose aversion, I think Monopoly does that great. Rolling doubles lets you go again, but roll doubles 3 times in a row and wind up in jail. Everyone feels like that player (even when it is themselves) deserves to go to jail because they are going too fast.

I must be old school, because I'm ok with losing stuff in games . It's part of the game.

it's a cool talk and i believe what he's saying, but i also chose the opposite choice for a lot of scenarios he mentioned

Is this loss?

you could've said "the unhappiness of losing the $20 is/feels more intense than the happiness of gaining the $20

Every time i see "80%" I think to myself ""Those are Hydro Pump odds" and go for it.

At 11:07, I would hate that change to the card as it's strictly worse. In the original situation you have knowledge over which cards you no longer have in your deck, in the alternate version, you lose that information, but still have 2 cards you lost. EDIT: ah, you mention this seconds later.

Jeff!!!!!

His example of the 20 dollar bet demonstrates marginal utility and diminishing returns. Loss aversion is emergent behavior of those principles at best.

Funny how we get the companion cube back at the end of Portal 2

nice

7:40 no it's not. It's better over all for a large group not for a individual person. It is far better to get something slightly lesser them to have nothing

Hello gdc. Great content as always.i have one doubt Can we repaint game and character of game and publish it in other language with same mechanism . Is there any copyright issue for that from designer or developer of that game?

Oh boy, the deadly disease part didˋt age very well, did it? But great talk nontheless.

He told the person he won $5 but then have it sit there for an hour, keeping him on his toes whether or not he will actually get it :p

The issue with the expectation value is that from the perspective of the participant, it is unclear how the probabilities are REALLY implemented. There could be cheating involved. So, guaranteed to get $3000 is better than 80% $4000. Because who knows what that 80% really mean and who throws the die.

Calling cutting off the longest road with a settlement in Catan an edge case, clearly hasn't played enough Catan 🤣

47:05

Am i the only person who didnt care about the companion cube? Tbh i thought it was a joke, i had no qualms about incinerating it, though i did look around if there was some trick just out of having been told i had to do something, being a puzzle game i thought maybe there was a clever little puzzle to it...

50:40 because there was no question asked, that was just brabbeling

So loss aversion like this is not actually irrational in many cases. One of those being the diminishing marginal utility of money as mentioned in this video. Its COMPLETELY rational to take 100% of $1 million vs a 50% chance of $3 million for that reason alone. But there are other reasons. Take this game for example: ukposts.info/have/v-deo/l3amqY-KhG2qsKM.html . Average returns of playing is 15% gain per play. However, for any number of plays, the majority of players will end up losing money. The massive winnings of the few pull the average up. So expected value isn't everything. People rationally realize they're not the chosen one and won't likely do better than average. Even if this is subconscious. Its a modern double-myth that loss aversion is irrational. Too often we hear "oh that's actually a myth" and assume someone saying something's a myth is more likely to be correct than the person propagaing the actual myth. In this case, the idea that its a myth, is the myth.

Level drain is worse than death in some ways. Player death feels like an end, a conclusion to that characters story. Level drain feels like the time the player spent building up the character is lost. With the additional penalty of having to remove all the gains of those levels and recalculate stats. It also doesn’t make thematic sense for most monsters that have it. Early versions were even worse because it all depended on the DMs roll.

About Theory of Choices - did you knew original doom had precisely 7 types of the enemy? :)

Wheres proper captions?

i dont see how taking an 80% over a sure bet on a gain is the wrong financial choice the thing 20% chance to loose might aswell be 100% because there is no garentee you can easily be the 20% who gets nothing, the problem here is none of these questions have context which the human brain uses to make all of its descions so by asking me if i'd rather make a sure bet or a risky one without asking me my financial situation before hand- its meaningless information without the context of the thinking behind the descions.

All those negative emotions are bad for a person's well being and also feed the Archons.

This guy doesn't measure the personal situation into his things. I gain access to way more things with free $3000 than I would with an extra 1k and a chance to get 0. One is a sure 3000, the other one could end you with 0. If you played the game like 10 times, yes, choosing B would be almost always right.

That's why all this time people are still agonizing over the death of Aerith.

27:31 Shots fired at League of Legends (well, and TBH entire of the industry standard).

Difficult to get past 17:41 in 2022

Someone hasn't watched A Beautiful Mind.

um choice A in the first example given is not 'better for you overall'. it might be better on average for 10~ participants or whatever, but the fact that it mathematically abstracts out that way only demonstrates the limitations of that kind of abstraction. the people who are part of that average that walk away with $0 are not going to feel better about it because on 'average' everyone walked away with $3,200; not unless everyone agrees to pool and split the proceeds evenly this shit is just so disconnected from reality like, the only way you can make sense of someone arguing that the choices made are paradoxical is if $3,000 is literally nothing to them, and ergo it is always worth the risk of walking away with nothing because if you do oh well who cares? and it makes all the sense in the world that the same person who sees meaningful value in a guaranteed $3,000 is going to take the odds on losing $4,000 or losing nothing because losing $4,000 and losing $3,000 is exactly the same thing to them: it would mean they are so fucked that nothing else would matter. being $3,000 vs $,4,000 in the red is a laughable distinction, but at least with the second option there is a 1/5 chance you'll be okay the talk gives the impression of someone that has never once experienced money problems

I understand the intended affect of ripping up cards in legacy games, but there is another angle to it- you can never "start from scratch" with another group or hand it off to someone else, making it really a one-time-shot sort of game, devaluing it in that sense. As a player who buys their games very selectively and wants to keep them in mint condition, I will NEVER listen to you if you tell me to destroy something I bought, it's a rather presumptuous ask.

"A bird in the hand is worth two in the bush." ~ useful adage .....Saying that taking the 80% chance to get $4,000 or nothing is in our "financial interest", over getting $3,000 for sure, is a bit of a slippery slope. Let's say you choose to roll the dice and you get the $4,000, what about rolling again for an 80% chance of $5,320 or nothing. And what about rolling for an 80% chance of $7,075 after that? And $9,410 after that? And $12,516 after that? In a vacuum, if it is in our "financial interest" to take the deal at $4,000, then it will always be in our interest to take the same deal for the same percentage of gain for the same probability. If you follow this logic you will eventually bust and be left with $0, even though each individual deal was in our "financial interest" to take. At what point, then, does it become rational to take the 100% chance of less money? I would argue that in context, it is hardly ever in our "financial interest" to take the deal, despite probability being in our favor, because the potential gain is only useful to a very small portion of people's starting wealth. For example, if someone starts with only $100 wealth, $3000 represents quite a substantial increase, a 3,000% guaranteed gain, while $4,000 only offers a 29% potential increase beyond that when you consider that the $3,000 is considered to be already in the bank. In context, that is the only valuable comparison here, 3,000% guaranteed gain ($100 to $3,000) vs 29% possible gain ($3,100 to $4,000). You would be foolish to take the deal in this case. On the other end of the spectrum, if you start with $1 million, $1,000,000 to $1,003,000 moves the needle much more than $1,003,000 to $1,004,000, the latter of which would have hardly any impact on overall wealth. You would have to really dig somewhere in-between to find an edge case where the potential increase of wealth is worth it over the guaranteed gain.

Comparing certain outcomes to statistically superior outcomes in the individual case is nonsensical. The person on the gameshow, or in whatever other situation will only experience a single instance of the statistical event. It cannot be accurately avaraged out. With simple reductio ad absurdum the value if a single 50% all or nothing deal is not one half of all, because while the avrage may be numerically comperable to a certain outcome, the relation of the 50% chance of falioure with any concrete guaruanteed sucess is a division by zero. To put it more accurately, with the 3k dollars, 80% chance of 4k dollars thing, on awrage you might make 6.6% more on taking the chance, but on a single instance you have an avrage 50% chance of 6.6% gain, and a 20% chance of a numeriaclly infinite loss. So the calculation is either a non sequitour, or if you want to use infinites, infinitely biased against whatever margain of superiority the gamble provides. The more instances you look at the further the chance of a total loss in each case diminishes, also teending to infinitesimal, but it will only actually reach it in an infinite nuber of cases, so the purely rational choice is to never take a gamble with a chance of total loss unless you are taking it an infinite number of times.

"Uhm" - Geoffrey Engelstein 2017

So my thing is, with regards to the "receiving $3000 or 80% chance of receiving $4000", there's another factor that I consider that I can't be alone on, and that's the validity of the random chance. I would need to see the guts of the mechanism that is determining the "80% chance" before I could fully believe it. If some guy walks up to me and gives me the above option, I would also say I want the $3000, only because I don't really believe that I have an 80% chance if I go the other way. I could very well have a 0% chance because I'm just taking the guy on his word that I have that 80% chance, AND it would benefit him if my chance were actually lower or non existent. Whenever there is something for someone to gain from lying to you, it's best to assume they will lie. Funny enough, some of the only circumstances in life where I actually believe I have the chance I'm told I have, is in the realm of a video game I've already paid for and that has no extra content I can purchase to "enhance" my game. A setting where no one has something to gain from lying to me.

The way this is framed is bad. If Choice A guarantees me 3000$ and Choice B gives me a 80% chance to get 4000$ and 20% chance to get 0$, it's perfectly reasonable to choose A, especially if you are poor (in general, the more value you give to the money, the more likely you are to choose A). The concept of taking averages as the "rational" thing to do is very abstract and doesn't really relate to game situations or real-life situations at all.

Sigh, and again it's all about the money when we speak about gain and loss. :/