Neuroscientist and avid blogger Jonah Lehrer recently published a great article in the Wall Street Journal about what he and others call “the superstar effect.” The piece is well timed, seeing as it deals largely with the effect that someone like Tiger Woods has on his competition and Mr. Woods has in fact just returned to harass his competitors for the title of “Most Badass Dude Ever at Golf.” Lehrer describes the work of economist Jennifer Brown, who meticulously studied not just Tiger’s performance in high stakes golf games, but the performance of his peers:
Ms. Brown discovered the superstar effect by analyzing data from every player in every PGA Tour event from 1999 to 2006. She chose golf for several reasons, from the lack of “confounding team dynamics” to the immaculate statistics kept by the PGA. Most important, however, was the presence of Mr. Woods, who has dominated his sport in a way few others have.
…
Such domination appears to be deeply intimidating. Whenever Mr. Woods entered a tournament, every other golfer took, on average, 0.8 more strokes. This effect was even observable in the first round, with the presence of Mr. Woods leading to an additional 0.3 strokes among all golfers over the initial 18 holes. While this might sound like an insignificant difference, the average margin between first and second place in PGA Tour events is frequently just a single stroke. Interestingly, the superstar effect also varied depending on the player’s position on the leaderboard, with players closer to the lead showing a greater drop-off in performance. Based on this data, Ms. Brown calculated that the “superstar effect” boosted Mr. Woods’s PGA earnings by nearly $5 million.
The reason, Lehrer goes on to explain, is that when faced with such an overwhelming favorite in the odds, people tend to short sell themselves and not give their best performance, as if the outcome is predetermined. And what’s worse is that this need not even take place in our conscious thought to have an effect. And what’s worse than that is the fact that the phenomenon seems to be most potent with more experienced players. Veteran golfers play a good chunk of their game on autopilot, not wasting mental energy over analyzing every tiny movement, angle, or twitch. But when Tiger Woods is on the fairway, they may begin to overthink their strokes, their choices, and their plan –to engage in too much of what psychologists call “action identification.” The result is that they change the way they play and play worse as a result because they’re wasting their finite concentration on things that didn’t need it yesterday. Writer Malcom Gladwell of The Tipping Point and Blink fame also has a nifty article about this phenomenon, which you can read here.
When we talk about someone “psyching out” the competition, this is what we mean, and it appears to jive with actual scientific research. The WSJ article goes on to discuss how this same phenomenon happens in other competitive environments outside of golf, such as law firms or the executive boardrooms of General Electric, and how it’s especially potent in “winner take all” situations.
…Like, say a game of StarCraft! In the realm of video games, what this all made me think of is the importance of proper matchmaking based on skills and how some games seem to do it a lot better than others. Whenever I jump in to competitive game of Modern Warfare 2, for example, I can’t seem to take four steps without getting owned because everyone else in that game seems to be SO MUCH BETTER than I am. I think many of us have been in a poorly mached game where we round a corner to face the person dominating the top spot on a scoreboard and we just sort of sigh and wait to get headshot rather than try and fight back, especially if we’re squatting at the bottom of the rankings. Halo 3, on the other hand, always seems to group me with people closer to my skill level, and I have a lot more fun and win a lot more matches as a result.
A list of the people who would crush me in any given game of StarCraft II.
The superstar phenomenon is something that Blizzard seems to be actively trying to avoid in its ranking system for StarCraft II with its bronze, silver, and whatever levels of play and the ability to see the ranking of your opponent. Though not perfect and obviously still being tweaked, the system seems to go to great lengths to match players with opponents of similar skill. So I can be relatively sure that I’m not going to waste time second guessing my build order or metagame because I was matched against Tiger Woods, who in the context of this game would be some Korean dude who has been playing StarCraft for 12 hours a day for the last 10 years.
This effect has happened to me, and also appears to compound with the fact that once you win, you tend to keep on winning (and similarly for losing. There’s a name for this effect but I’ve forgotten it). I’ve found that getting a couple of good frags sets me up for a great game, even though I may be against some stellar players (them dying those couple of times probably also negatively affects them). However, being against stellar players can be a scary thing, and I do tend to think twice before trying that daring headshot-but-I-end-up-in-the-open move. Once this low-risk behaviour punishes you, you tend to take even fewer risks, which puts you into that loser’s spiral.
The good news is that you can actually train yourself to get over this (both the loser’s spiral and the superstar effect). Unfortunately, it’s not achieved through actively not thinking about your game. You have to think even harder and re-analyse what you’re doing and why you’re doing it. You have to actively re-inforce positive messages in your mind — get a lucky kill and be sure to say to yourself “See, that wasn’t so hard!”. Similarly force yourself to take the riskier shots, even though they don’t work out in the beginning. The trick is to be in the situation enough that you learn to be your best when you’re against the best.
Sadly this only really counts for FPSes. In Golf who knows what tricks there are (if any). If you watch enough tennis you’ll know that the really good players become even better under pressure, whereas the inexperienced or mediocre players get worse. The number of times I’ve seen Serena Williams getting trounced, then on match point winning enough to take back the set, then the match, is astounding. This is what really sets the great players apart — when they meet the all-time greats (like Federer) they’ll bring their A-game and get even better over the sets, as opposed to crumbling under it all.
In addition to the overthinking, I know I also will intentionally lower my own level of effort and just give up and “take the head shot” as you said. When the other team is far better than my own, I’m much, much more likely to create piles of my own body doing dumb rushes to a point instead of thinking strategically.
In regards to Tiger, perhaps the other players feel they need to play better to beat Tiger so they go for longer drives then normal, and take riskier shots but end up with more strokes because Tigers innate Physical abilities are better. On the other hand they could decide they need to play it safer and hope Tiger messes up. The problem is that slow and steady does not usually win the race and they end up with more strokes.
“The reason, Lehrer goes on to explain, is that when faced with such an overwhelming favorite in the odds, people tend to short sell themselves and not give their best performance, as if the outcome is predetermined.”
While I agree with Lehrer’s statement on some level, I do believe it is way too generalized to really be applicable.
Some players thrive on competition (like Tiger), even if it is way above their head. Everyone starts out as a noob and its the best players that are constantly pushing themselves to get to that next level and make that next headshot. If you are intimidated by the best than you can never be the best.
Miles: Also, in the context of golf, Tiger Woods is so good because HE takes longer drives and riskier shots than the “average” person on tour. Aside from his mental tenacity, focus, concentration etc, is the fact that he just plain GOOD and does things with alarming regularity that other golfers only try in practice- like driving the ball far as hell and skipping approach shots and just getting on the putting surface.
I feel like a tool for posting this but I think the blogger and author of the above mentioned article is Jonah Lehrer, not “Daniel” as cited in this blog.
I remember reading the chapter in Gladwell’s book and wondering whether the brain functions involved in the “choking” phenomenon are those that control our fight or flight processes. I know that for me personally, and after the fact especially, I always feel like I’ve had crossed wires whenever I’m dominated online. My e-penor becomes slightly deflated and if it happens often enough it can make me hypersensitive about every little action/decision I make. The more player deaths I experience, the more anal I become. Even in a single player game when I have a friend watching. It diminishes the fun of the game and creates a perpetual cycle of frustration. Inevitably leading to a rage quit.
This discussion questions other subjects like game balance and hardcore vs. softcore too.
@Charles
Regarding the point about a friend watching, there are some theories in Psych that can help explain. There is a theory in Psychology called the audience affect that states tasks you are really good at will be facilitated by an audience (social facilitation), but tasks that don’t come as easily you will perform worse on while an audience is present.
There is also the Yerkes-Dodson law which says that every task has an optimal arousal level. Some tasks require more mental resources (e.g. they are harder, or less practiced) and therefore additional arousal will take away from your available resources resulting in less performance.
It happens to me as well; I play best when I am alone.
As far as performance decreases after consecutive loses, it’s true what Sunny Kalsi said about winning and losing. It’s called the “winner effect” (creative I know), but essentially winners will keep winning and losers will keep losing. It’s found in aggressive interactions in a multitude of species. For humans it’s likely your confidence and self-esteem being shaken, so you start to doubt yourself and feel the opponent is innately better.
I really enjoy your articles but I feel you’ve made a sloppy comparison from the superstar effect described to the video games examples. In the golf scenario you are talking about the very best professional golfer (Tiger) versus other professionals who are just slightly less good. And being in the same tournament with him slightly lowers their performance, due to overthinking or whatever. Then you compare this very poorly matched online games where you just give up and take the head shot because you know the other players are so much better than you. I don’t think this is really the same thing. You also say Blizzard is trying to avoid the superstar effect by _showing_ the player rankings – how could you even have the effect if you didn’t know you were playing a superstar?!
“some Korean dude who has been playing StarCraft for 12 hours a day for the last 10 years.”
Speaking of which… Flash has been tearing it up.
Heh there has been quite some raging in the teamliquid forums over the ranking system. In particular, the more competitive gamers want an absolute ranking system and think this soft relative system won’t help anyone know “really” how good they are. The other side of the argument is the self-esteem, superstar effect that you just explained.
Question for clarification: Does the superstar effect have any benefits of self-handicapping? Is there some effect on self-esteem because.. cmon that guy is just so good or no effect because… dang I just got raped.
I wonder if audience effect cold be taken advantage of in a game design. What if you could see the avatars of other players as spectators. Is there a difference if the spectators are online players or NPC’s? What if the designers put in extra NPC’s as game spectators, but led the player to believe that they were actually players? Would this affect the players’ performance? More importantly, might something of this nature this make the game more fun?
For team shooters, I’d say it’s more about team inertia. A team that does well is likely to stay together and just get better together as they learn each other’s habits. The people on a losing team are more likely to quit to find a more palatable match, so the winning team just becomes a freight train of glory and the people that join are going to be fragmented and continue losing
For me, there is a range.
Lets take street fighter as an example, because that’s where I really learned about competitive video game playing.
If I was playing someone and getting countered on everything I do or they were employing some strategy that was so good I just couldn’t do anything, I would definitely get frustrated and either quit playing or just watch the guy play for awhile.
But, if I felt like I had any chance, if I could take even one in ten rounds or the like, I’d push even harder. I like to be challenged and if I feel I can adapt and learn to the current competition, I’m twice as likely to push harder.
So in the golf example, were I me (a non-golfer) against Tiger, I’d probably quit worrying about the score and try to take in the scenery, bullshit with Tiger, etc. Something to alleviate losing. But if I were say Phil Mickleson, I’d push that much harder when Tiger played.
Maybe it’s that effect that’s getting to these pro golfers. Maybe because he’s so good, they aren’t just relaxing and playing, they are working their asses off to do everything as perfect as possible. And sometimes that leads to not as good as play if you can’t relax into your comfort zone…?
Whatever the case is, for me, I’d MUCH rather play someone who is better than me. Or at least equal. So I can go all out.
First off, sorry folks for not replying to the great comments sooner. I’ve been out traveling and attending a convention all week. It’s all convention stuff during the day, then schmooze, then party, then fall asleep, then repeat. It’s rough, I know.
@ Sunny Kalsi
The stuff with experienced vs. inexperienced tennis players is addressed in that Malcom Gladwell essay I linked to in the OP. Oddly, it seems to be at odds with Leher’s claim that the superstar effect hits more veteran athletes harder.
@ Soft Nonsense
Personally, if I see a superstar in a FPS or similar match, I tend to run. Especially if he/she hasn’t seen me yet.
@Miles
That’s an interesting hypothesis, but not one that the researcher I reference in the OP seemed to be testing. And playing against “the best” isn’t what’s on the table here. It’s playing against someone who is so far above everyone else (or who is perceived to be) that it puts you off your game in a way that nobody else can.
@Charles
Oops, thanks for catching my mistake. Fixed. Also, I think the “crossed wires” thing has a lot to do with devoting limited attention and cognitive horsepower to things that you shouldn’t be. It limits your ability to concentrate on things you should. Concentration is a FINITE resource.
@ Chris
Hrm, cool. You’ve given me a few new things to research. Sounds like there could be a PoG article in what you describe.
@ Gordon
It’s not exactly the same thing, but I think the same phenomenon may be at play. Saying that you “give up and take the headshot” is kind of an exaggeration, sure. But I do think that we may overthink and reduce our estimates of how likely we are to succeed (i.e., self efficacy) which is a prime component in motivated behavior like competing in a video game.
Also, in regards to the SC2 matchmaking, I should have noted that seeing your opponent’s record only folds into this if you are indeed appropriately matched. Looking at their record and finding that you’re outclassed would indeed probably lead to the superstar effect. I could have probably written that more clearly.
@Sabu113
The whole question of how much information to share in matchmaking is really ripe for research. I wish I knew if someone at Blizzard, Valve, or any middleware vendors were really looking at it through the lens of psychology. How cool would it be to do your Master’s thesis on something like that?
Regarding the self-handicapping, I dunno. That’s an interesting question.
@ Anton
That’s an interesting qustion, too. Some games do use custom avatars. 1 vs 100 on XBL, for example, literally has an audience made up of avatars. Other games like WoW allow highly customized avatars that basically amount to the same thing.
@JJZ
Huh, that’s true. In fact, it makes me think that you could directly test a hypothesis about the superstar effect by looking at how often people drop out of a match in the face of a clear mismatch of skill levels. You can’t just drop out of the Master’s Open Golf Tournament, but you can easily drop out of a MW2 game or a fighting game. Those data have got to be out there.
@ Matthew
Right. Same as above.
I wonder if the superstar effect can lead to head hunting. I know if I’m playing a FPS and someone is way ahead of everyone else in kills I will try to kill them, even at the detriment to my own score. I don’t know if that is just a spite thing, or my alternative to rage quitting, but there it is.
I’m not certain if Halo 3 uses it, but Xbox LIVE has an advanced ranked matching system based on chess skill rankings. Microsoft put a lot of time into their matching algorithm. I suspect that many other matching systems have a very ad hoc algorithm not based on research but on whatever the dev felt might work well.
http://research.microsoft.com/en-us/projects/trueskill/Details.aspx