The Problem of “Win-More”
In Magic: the Gathering and other popular card games, advanced players have developed the notion of a “win-more” card. A “win-more” card is one that works very well, but only if you’re already winning. In other words, it never helps turn a loss into a win, but it is very good at turning a win into a blowout. This type of card seems strong at first, but since these games usually do not use margin of victory scoring in tournaments, they end up being a trap—instead of using cards that convert wins into blowouts, you want to use cards that convert losses into wins.
This concept is useful and important and you should never tell a new player about it, because it tends to make them worse at the game. Without a more experienced player’s understanding of core concepts, it’s easy to make mistakes and label cards that are actually good as being win-more.
This is an especially dangerous mistake to make because it’s relatively uncommon for an outright bad card to seem like a win-more card; win-more cards are almost always cards that look really good at first. That means that if you end up being too wary of win-more cards, you’re going to end up misclassifying good cards as bad, and that’s an extremely dangerous mistake to make. Misclassifying bad cards as good is relatively easy to deal with, because you’ll use them and see that they aren’t good; misclassifying good cards as bad is much more dangerous, because you won’t play them and therefore won’t get the evidence you need to update your position.
I call this the “win-more problem.” Concepts that suffer from the win-more problem are those that—while certainly useful to an advanced user—are misleading or net harmful to a less skillful person. Further, they are wrong or harmful in ways that are difficult to detect, because they screen off feedback loops that would otherwise allow someone to realize the mistake.
Even though you’ve said this, people will think you defined the “win-more problem” to be the problem with something that helps you win more only when you’re already winning.
Yeah, this is correct. The naming is confusing.
This is part of a general issue in binary win/loss games.
As you approach the end of a game with a lead in points, the value of scoring more points diminishes. If you don’t adjust your strategy to compensate, you end up taking risks that increase your expected points, but decrease your win probability. Win probability plotting is popular for American Football obsessives.
It’s often very frustrating to be a sports fan, as it’s obvious what’s going wrong with the decision theory in the heads of coaches and players.
This is also one of the standard heuristics in chess. When you gain an advantage large enough, e.g. a couple of pawns, the best course of action is usually to stop trying to gain more advantages and instead trade as many pieces as you can, to avoid counterplay and reach a simple-to-win endgame.
Does it take a really good player to take advantage of this strategy i.e. is knowing about it a trap from middling players ?
Intermediate players may, at times, enjoy a material advantage and start exchanging pieces. However, they may become too focused on exchanging pieces, as if it were a terminal goal, and neglect their position. Once their pieces are placed sufficiently worse, their material advantage might not count as much anymore, and they might lose their advantage and the game.
I recently witnessed exactly this pattern in a game of two weak amateurs, where the weaker player had won a minor piece. In the end, he lost—not because he blundered back, but because his opponent could gradually improve his pieces to the point where he was actually better despite still being down a piece.
However, simplification (even giving back material) to reach a technically won endgame is a viable plan of action for stronger players, so I think it fits the pattern of a “win-more” problem.
Quibbling: Weak players sure, but club level players rarely mess up simplification that much. They need a larger advantage than stronger players to be sure they’ll end up better after simplification, but they don’t usually misjudge that advantage in my experience. So at club level and beyond it’s very standard and effective stategy.
You are correct—although I have also seen club players misjudge a Queen + pawns vs. Rook + Pawns endgame, thinking it was won when in fact it was only a draw.
I wonder if other sports could use a model similar to the Triwizard Tournament. The outcome of the first two events sets the handicap for the third event, which is the only one that really counts. Be ready for some bad ideas.
Adapting this to American Football, it could be that your score after the first three quarters is converted into an advantage in downs or yards in some fashion for the last quarter.
Baseball—each run in the first 8 innings counts as an extra out in the last inning.
In Everyone Else Football… umm… I’m having a hard time coming up with something not boring but less insane than “each goal you score in the first three quarters is an extra ball that goes on the field in the last quarter and can only be scored into your opponents’ goal”.
This isn’t very much “a sport”, but it came to mind that Team Fortress 2′s Payload Race game mode works exactly like this on multi-stage maps. The handicap (starting farther ahead/behind) is very small and usually overwhelmed by team organization and the outcomes of combat, though.
I would watch this variant of Everyone Else Football.
In a failed attempt to make the game “interesting”, the NFL Pro Bowl would do the opposite, and help the team that was behind. They would get the ball again after they had scored if they were still behind. Naturally, that makes it easier to catch up.
The real danger of the “win-more” concept is that it’s only barely different than making choices that turn an advantage into a win. You’re often put in a place where you’re somehow ahead, but your opponent has ways to get back in the game. They don’t have them yet—you wouldn’t be winning if they did—but the longer you give them the more time they have.
For a personal example from a couple years ago, playing Magic in the Legacy format, I once went up against a Reanimator deck with my mono-blue control deck. The start was fairly typical—Reanimator trying to resolve a gigantic threat to win, while I played many counterspells and hit him with some Vendilion Clique beats. My opponent ended up getting an Iona out (naming blue, obviously), but went down to exactly one life to do so. This was very, very awkward for him, since he couldn’t attack with the Iona, activate fetchlands, or use the alternate cost of Force of Will. But, I had outs—Powder Keg (7 copies of keg/ratchet bomb) and waiting 9 turns, or Vedalken Shackles (3 copies). So I stayed in, and got as many draw phases as I could, and lucked out with a Shackles topdeck, followed by being able to play blue spells and winning the game.
Anyhow, my point is that cards that help you only when you’re winning can turn wins into losses. Your opponents can have outs, and it’s a good idea to take those outs away. If you don’t, then sometimes your opponent will pull exactly what they need to do something ridiculous—say, dealing with a card that keeps them from playing 28 of their 38 spells, and seven of the ten spells they can play take 9 turns to do anything about it.
“Win-more” is definitely the wrong word to describe this concept. I think a better choice is calling it a “close-out” or “finishing” card. The point of these is to make sure that you win when you have an advantage. It also tells you that you don’t want too many of these—many decks run just one or two copies. Dredge, for instance, runs a single Flayer to turn having their deck in their graveyard into a win. My mono-blue control deck ran two Sphinx of Jwar Isle (there were essentially zero answers for him in the meta, and I’ve stolen games with him. That said, one copy would be an Aetherling if it was printed at the time).
Replacing a card with a finisher means that you’ll take fewer leads, but win more games while ahead. Sometimes the right number of finishers is one - when Dredge has a lead, it’s got access to all or most of the cards in their deck. Sometimes it’s more—my mono-blue deck would run between 2 and 6, depending on how I felt about Jace at the time. Often it’s zero, and your game plan is to win with the cards that got you ahead in the first place.
I think it’s important to make the difference between finishers (such as Sphinx of Jwar Isle) and win-more cards (such as Nomad’s Assembly).
A finisher is something you use to end a game quickly. A win-more card is a card that only helps you if you are already ahead.
The Sphinx ends the game in 4 turns no matter what. It’s a good win-condition. Nomad’s Assembly is something that only helps you win if you already have a lot of creatures.
See also The Valley of Bad Rationality.
Can you give additional example of this? I’m looking for parallels elsewhere in life where something is apparently useful, but is only useful in cases where you don’t really need it, and nothing leaps to mind.
Skepticism. In theory it allows you to weed out bad ideas and find good ones, in practice it allows you to dismiss any idea you don’t like.
Trading individual stocks? Often touted as a great way to become wealthy, but in reality only the most skilled are successful at it. Much better to buy index funds.
Picking individual stocks, being wary of “concern trolls” on the internet, making big elaborate complicated plans.
The last one is the only one I understand from your list, but it seems like a rather good concrete example. Making an elaborate plan is a great idea if it’s definitely going to be useful, but otherwise it’s a waste of time—like in HPMOR, when Harry makes a whole scheme of experiments to do that’ll last for many months, but then it turns out his most fundamental premise is totally wrong. The big plan give him a big win when his assumptions are correct, but don’t help him one iota when they’re wrong, and he wastes a lot of time.
Moral: only make large schemes when you have strong evidence that they’ll be very useful, and even then, first carry out any basic preliminary research, to check the assumptions going into the plan.
On #2 - Oh yes. Actual concern trolling is less common or harmful than dismissing actual ally concerns.
The idea of building a Minimum Viable Product comes to mind, especially if there’s an externally imposed deadline. Some of the timed tests I’ve done for IT classes seem like purer examples of this than I’d expect to see in the real world—with some types of grading, you can’t get a passing grade unless your thing /works/, and extra features don’t count for much of anything unless your basic part is working.
The Arctic Expedition thought experiment also sounds similar.
-- My “Murphy’s Law” wall poster
(Note that this particular example is no longer true.)
I’m under the impression that today it is more like:
I don’t think that was ever true.
Superstimulus food.
Most superstimulus food would be very useful to a starving person, so it doesn’t fit the expressed criteria (specifically ”...only useful in cases where you don’t really need it...”).
added as top level comment: http://lesswrong.com/r/discussion/lw/jsr/the_problem_of_winmore/aqgd
Win-more is a subset of narrowness. Cards that are narrow are good in smaller subsets of the situations that exist in Magic than cards that are less narrow. A creature is broad, because it can interact with creatures and there’s almost always an opponent for it to attack. A spell that destroys enchantments is narrow: It requires your opponent to play an enchantment, and is useless in every other situation. Win-more cards are the ones that are limited to being useful when some gamestate prevails, usually involving you have a lot of something occurring. Cards that trigger when 5 creatures attack, cards that cost 9 mana, etc.
Here are a few examples from a cycle of cards in Zendikar that came to mind when I started thinking about this topic:
Quest for the Gravelord is a straight up good card. It’s easy to cast, and requires only 3 creatures to go to a graveyard before it puts a sizeable threat onto the battlefield. It’s somewhat situational, but in any average game of magic, creatures will regularly be dying on side of the table or the other. It has to be in play before creatures die, but can be played after you play creatures, on the same turn as you play a removal spell, or late in the game before your attack step.
Quest for the Nihil Stone on the other hand, is a pretty good example of win-more. Players discarding cards is not very common: usually you have to make them. This means that in order for this Quest to work, you need to have drawn it before you play whatever spell is making your opponent discard. As in: you already need to have been somewhat lucky for this card to be good AT ALL. Assuming you draw it early, putting two counters on this quest won’t be hard for any deck dedicated to it, but making sure an opponent has no cards in hand is a lot harder: This will usually take 2-3 spells on your part. This is part one of the win-more narrowness: Cards that are good when you have several cards that support them. In a dedicated deck, this is usually not a problem, though it’s still a big deal in Limited. The second part of the narrowness: If your opponent has no cards in hand, they are doing nothing. This means that unless you are also doing nothing, you are winning the game. Quest for the Nihil Stone is only good when your opponent is already in a position of failure. And they draw a card every turn. So even if you get them into the position of vulnerability, the mechanics of the game can help them escape.
The key to understanding win-more is that you’re not comparing cards in a vacuum: You’re comparing them to all the other cards they could’ve been. In the time a Quest for the Nihil stone takes to become active and deal 5 damage, a Vampire Lacerator could’ve already dealt 6, 8, or even 10 damage. If your opponent has no cards in their hand and is half dead to any creature you might’ve played, you would be winning anyway. Nihil Stone only makes you win in a situation that’s already great for you, and in which almost any card that has a positive effect would also be good.
Cards can be win-more in some decks and good in other decks and just plain bad in other decks. This doesn’t mean win-more is a bad concept, simply that it depends on context. There are times when Quest for the Nihil Stone is good, there are times when it’s the best card for a situation, but those times are way outnumbered by the times where it doesn’t really do any more than a random creature would’ve done, and that’s not counting the times it does LESS. Don’t get me started on Strictly Better and Mindslaver.
This narrow/broad thig sounds familiar. I haven’t played Magic but a comparative game (a long time ago) and there was a time where I calculated for each card its effectiveness: Basically the chance to deploy it multiplied by its damage and then use the set maximizing the overall total. This means that a very powerfull card that requires lots of other cards and/or preconditions is less valuable than say two average cards. This caluclation of course only works if the effects of the cards are (basically) additive. In that game it was mostly the case.
In MtG there only winning, losing and a draw. Real life isn’t that way. In real life it matter very you get a job being payed 100k or win-more at the job interview and get payed 110k.
If you read the article carefully, this is not relevant to the “Win-More” concept as katydee defines it.
Reading it you are right, but I think that’s a case of highly unintuitive name that provokes confusions.
As a Magic player, I’ve thought that “win-more cards” have a worse reputation than they deserve. You don’t just need to be winning at some point, you need to maintain a winning position long enough to actually end the game. If you’re barely winning, your opponent is much more likely to make a comeback than if you’re totally dominating things.
That’s true, but I feel like the main thrust of the concept is that usually a card that’s pretty good all the time will close out the game just as well as the thing that’s really good when you’re really ahead but otherwise a brick.
That said, any concept that causes you to dismiss a card entirely is a bad one, and probably leads to a lot of sideboarding mistakes.
Also known as “a little learning is a dangerous thing”, as Alexander Pope noted in early XVIII century...
The real problem with “win-more” cards is that they’re conditional. Being conditionally good is a common criticism of many magic cards. The advice to new players is simple: think about how likely these conditions are to be met. If they can’t make this estimate, copy a more experienced player’s opinion.
It’s also possible that a “win-more” card is also just a decent card even when you’re not winning. For example, if you play MTG now, you’ll know that 2 desecration demons on the play against G/R monsters is both very good, and that the second demon is “win more”.
So I don’t really see a problem with playing “win more” cards. The problem is playing too many conditional cards.
I’ve never known this to happen, and I’ve been playing magic for years.
I think that ‘win-more’ is the probably least helpful concept in card game strategy. Not just for new players, but for experienced players as well. (Its annoying enough to read people say something is win more on MtG sites, now I have to read it on LessWrong as well!)
I’m of the opinion that almost everyone would become better players and would be better at card evaluation if they eliminated the concept of Win More from their vocabulary. It almost never is used to actually successfully avoid using an actual bad card, and almost always results in misclassifying good card that would actually help you convert games into a win as a bad card.
What term should we use for concepts that are net harmful for ALL users?
I rarely follow MtG sites so I didn’t know about this term, but if I did it would have really simplified a recent conversation where it was difficult to explain why such good cards had no place in the deck we were building. Any vocab term that can bundle a group of arguments together in a way that is easy to convey and understand will simplify constructive conversation.
I agree. That’s why I used it as an example here. The concept of win-more is useful—but only to high level players. Since very few players are high enough level to benefit from the concept, nearly everyone who uses it does so incorrectly and in fact becomes worse.
“Wrong?” “Bad?”
I don’t think so. It’s interesting that the LW article makes top 10 in google for google win-more card. There one article http://magic.tcgplayer.com/db/article.asp?ID=8867 about the concept.
It names Iona, Shield of Emeria as 9 mana card as example. That leaves to me who played magic 10 years ago the question of why you need a concept to tell you that you shouldn’t play cards that cost 9 mana.
This post reminds me that “using Google to judge the popularity of things” is a good example of the problem described in the OP. Many times on the internet I’ve seen people claim that something is more or less popular/known than it really is based on a poorly formulated Google search.
Also, compared to when you last played, high-cost cards are more likely to be viable.
I’ve seen it too. Even Nate Silver did it in this New York Times blog post, where he estimates the number of fans for each team in the National Hockey League “by evaluating the number of people who searched for the term “N.H.L.”” Using his method, Montreal is the only Canadian market with a team for which it is estimated that fewer than half of the people are avid hockey fans (as he defined it).
In Montreal, French is the official language and the language spoken at home by most people.In French, the NHL is called the “Ligue nationale de hockey,” abbreviated “L.N.H.”
Web search engines aren’t really designed to deliver comparisons of popularity, anyhow; those numbers are pretty much a way of saying “look, we index a lot of stuff!” rather than an accurate count.
Systems like Google Books’ Ngram Viewer are designed to compare popularity of terms — though that one indexes over a corpus of works in print, which is not the same as the Web.
This is better, but it’s also common to get Ngram viewer wrong—eg not realizing that a word has multiple meanings which may have changed over time, or not realizing that there are two different ways to phrase the same thing, etc.
Hmm, I don’t see the LW article on the first page at all. Perhaps this is different search customization?
In any case I see several other articles on this topic, as well as many forum discussions about it, people asking whether specific cards are or aren’t win-more, etc.
Are you suggesting that this concept is a country-example to the argument put forward in this essay that was recently listed in the rationality blogs? At the moment you and the author have given your expertise as evidence, it would be nice to have something a little more concrete.
I consider fashion a real-life win-more ‘card’.
A concrete example in a situation where you have an all-or-nothing choice like in a card-game is job application: Either you get the job or you don’t get it. Investing enery and money in an expensive business suit and optimum outward impression is a win-more strategy because it only helps once you already have the face-to-face interview where your appearance will not much difference (except in a few professions) but at best win-more. Your enery would have been better invested into writing more or better applications and a thorough preparation.
I’m not sure if this generalizes to fashion being in general a win-more strategy.
I’m pretty sure it doesn’t. Most situations you encounter in everyday life are far more iterated than job interviews: you’ll only conduct a face-to-face interview once or twice for a given job, but you might go on a dozen dates with a potential partner before you enter into a serious relationship; you likely meet potential friends and contacts a few times a week; and you’ll meet with your coworkers almost every day for years. Even if optimizing your appearance adds little marginal advantage in any given situation, it adds up quickly, and I suspect that advantage can actually be fairly large in situations where there’s not much to go on except a first impression.
Actually, I’m not even sure it’s generally win-more in the context of technical job interviews. There your skills probably make the most difference, but it’s not uncommon for someone without much technical savvy to have veto power, making personal presentation something to satisfice if not to optimize: you can be the best engineer in the world, but that won’t matter if the hiring manager thinks you look like a bum. For non-technical jobs—especially anything managerial or customer- or public-facing, which are not at all uncommon—it can easily be a full-blown optimization objective.
I in general agree with this assessment. It does add up. But there might be an area where it doesn’t.
To give a concrete example: During school I avoided fashion, pop culture and socializing. I felt that it gained me nothing and take time away from my much more interesting subjects like math and (later) computer. Arrogantly I thought myself above this. Intentionally misunderstood jokes. Committed to be different. This made me an outsider during school (luckily without becoming subject to bullying). But it did allow me lots of time to study.
Only much later when my interests expanded to include evobiosociopsychology did I really grasp the effect. I believe that I have been a quick learner since. The only penalty that appears to be difficult to compensate is relations. I have only a small circle of acquaintances and I’m not sure I can (or want to) build it up. But I’m also not completely convinced that the business landscape requires this.
To end the example: Is it worthwile to focus ones skills on one area—an intellectual area that my promise to pay back with compound interest later—and to pick up other areas later (though possibly completely excluding some areas)? Or should one rather advance all areas at once (albeith possibly with different weight)? Or does it depend?
For me it was a most successful strategy so far—though it could have been just luck and chance.
I don’t think that even generalizes to job interviews. Your appearance makes substantial difference, though often not consciously.
Appearance is not fashion. You should come across as optimistic and capable. And if the job requires it you obviously must have proper clothing. But reading latest fashing magazines, buying an expensive stylish suit, haircut and manicure and whatnot (not including cases where this can be overdone) is a win-more card which I still think costs more than it gains.
I hope I exaggerated enough. And remember: This is a possibly contrived example for the OP.
I think you’re trying to draw a hard line where none exists. In general, there is no clear boundary between matching your appearance to the situation and mere fashion games; of course there are situations where costs outweigh potential gains (you probably don’t want to go into debt to buy an Armani suit), but it’s possible to overinvest in almost anything, and I don’t think there’s a case for generalizing that to the entire category of fashion.
Job Interviews can often involve assessing how much pay to offer someone. I’m not sure you can do the math on how much “fashion” gets you how much more money, but I wouldn’t be surprised if it compares favorably to a lot of salary negotiation tactics. Also: Many people go to many job interviews over the course of their lives. Spending money one time to optimize a suit or other outfit can be spread out over many opportunities to wear it.
I also think more generally this is a big fallacy among nerds, the notion of “Anyone who would dismiss me for how I dress is not someone I want to hang out with anyway”. You have to remember that you’re competing with everyone else a person may or may not interact with, and consider how many people you meet who you never have an in-depth conversation with. Everyone uses many criteria to judge whether to talk to people for the first time, or talk to them again or seek them out at parties or whatnot. “fashion” or more generally appearance and a sense of personal style is one of these criteria. If you look interesting or like a hoopy frood or this can make more people want to talk to you. If you dress like a goth, some people will really want to talk to you and some others will want to avoid you. Fashion isn’t just paying attention to what runway models wear, but to what sort of image and identity you yourself have. There’s not even anything actually wrong with going around all the time in cargo shorts and a shirt with a Carbon Dating joke or a My Little Pony on it. Just recognize that even your “default” outfit is signalling something to everyone you meet, and your social interactions will in a not necessarily visible way all be colored by it.
No doubt. There is significant impact of appearance. And I agree that it is kind of a nerd fallacy to assume (or insist) that inner values should qualify only or mostly. It took me a while to learn that.
But it is also a fallacy—and a potentially expensive one—to assume the opposite that you have to follow every fashion or trend (of your social peer group). Moderation and balance is the key here I think.
Yeah that’s one reason I don’t like the word fashion/fashionable: it combines trendiness with style in a way that can make you focus on the wrong side.
related: http://ctr.sagepub.com/content/5/3/1.short