The Wisdom of Crowds
Supermufacho has submitted a great article titled “On Picking your Dream Team and the Wisdom of Crowds”. The second year DT coach looks at how group think affects AFL Dream Team.
On Picking your Dream Team and the Wisdom of Crowds
I have already changed, re-changed, deleted players, made parallel but slightly different teams, made teams with different bye structures, different guns and rooks ratios, and all manner of other things already this pre-season.
As I am not privy to any insider information (insert Heath Shaw jibe), these moves and counter-moves have been based entirely on the advice, comments, and forum posts of people I have never met. I have no idea whether or not their information is credible or not, no idea whether they are trolling or not and so on.. This got me thinking about the crowd dynamics of pre-season Dream Team selection, and Dream Team itself as a giant exercise in the “wisdom of the crowd”.
To anyone that didn’t watch “The Code” the other week, the principle of the wisdom of the crowd posits that reliable answers can be produced to even complex questions by averaging the guesses (educated or not, serious or not) of a crowd of people. The larger the crowd, the more the average tends towards the correct answer. The best way to get reliable approximate answers is to ask the group every time – the many are smarter than the few. How many jelly beans in this jar? Ask even a medium-sized group of people and you will get a surprisingly accurate answer. What is the answer to a certain question on “Who wants to be a millionaire?” Turn to an expert, and he’ll tell you the right answer 65% of the time. Turn to the audience, the crowd, and they are correct 91% of the time.
So you could hypothesise that in order to be successful at Dream Team, to generate a “correct” answer, all you really need to do is follow the crowd. Judging by how many teams are the same or very similar come Round 1, this is what a lot of people do, consciously or not. Interestingly, all the good and sh*t advice, helpful and strategic comments, informed and misinformed articles, even out (just like joke guesses in the jelly bean experiment), and most people have pretty much the same team. But, this is where Dream Team and crowds get a bit more interesting…
The key thing to note here is that in picking your Dream Team to win the competition (or even your league), you don’t just need to pick the “right” answers when everyone else picks the right answers (lets just say that it’s 91% of the time for arguments sake), you need to pick the right answers in those cases where the crowd is wrong. In other words, you need to gamble, go against the crowd and win. Not too much obviously, because 91% of the time the crowd will be right, and you will suffer because of this (the number is probably a lot lower in the context of Dream Team, but you get the drift).
Individuals can and do outperform the group choice, but not the same individuals every time. Your outperforming the group is not necessarily essentially related to some special attribute you possess (better information, more rational internal deliberation etc). So again, the best way to get a reliable approximate answer is to ask the group every time. But you won’t win doing this.
The somewhat unique thing about Dream Team is that, from a crowd-choice perspective, unlike the answers in game shows or the number of jelly beans in a jar, there are no real right answers, only answers relative to the crowd – because you are competing against the crowd and your success is gauged relatively in relation to them, you are not in pursuit of an objective answer. There is one out there in the future (a player that will have the highest average etc) but it doesn’t matter if you don’t have him, as long as no one else, or the majority of other people, has him – because it is your points relative to other people’s points that matters, because you are competing against them, not against the maximum possible (hypothetical) score.
Having been reasonably successful in my rookie year of DT last year (top 2% – OK, not that successful, before you all you pros start cawing) having absolutely no DT knowledge whatsoever, past the “cheat sheet” produced by DT Talk, this is reinforced by my personal experience. I picked Suckling on a whim after the Grimes injury and a few other things I vaguely remember (Heppell and Hibberd from the start which turned out to be genius, Lower because I grew up in Norwood and he was mature-age, Puopolo for the same reason), so I felt like I was making my own decisions. But these decisions were already very much filtered through the knowledge of the crowd. In the above cases, I benefitted from the wisdom of the crowd (although not relatively, because most other people did too). But, I also suffered from the crowd wisdom – Foley was average, Morton was sh*t, Everitt was sh*t, Sylvia burnt me from memory.
So, if you want to finish high in the ranking, go with the crowd every time. Literally, go to your Assistant Coach and build an entire team based on the percentage ownership statistics (under the cap of course). Byes might make this tricky, but that’s probably another argument.
The moral, however, of this pseudo-scientific story is that you will find safety in the crowd, but never GLORY! So grow some balls and go unique. That’s what I’ll plan on doing – until 5 minutes before lockout, when I will pick Goddard even though I think he will have a sh*t year – because the majority of people have him anyway and so act as relative insurance for my potential loss.
The “unique” approach is of course tempered by the fact that you statistically will not outperform the crowd consistently, so of course there is a hell of a lot of luck involved – which is why, at the end of the year when they interview the winner of Dream Team, the crowd will collectively sigh “That guy??… Really?? Are you serious??”
It will be interesting to see how many teams are the same this year – rookies and a few mid-pricers seem more open in 2012, and there are certain bandwagons that you just won’t be able to afford not to jump on (like Curnow last year for instance). And you can jump on safe in the knowledge that everyone else is jumping on to. If it fails, everyone fails – no relative personal loss incurred.
It would be worth exploring how people knowing the percentage ownership (through the Assistant Coach) affects the crowd-choice dynamic, especially in pre-season, and maybe one of you stats geniuses out there can tell me how that might work. Further to that it would also be interesting to see what happens if everyone takes my advice and picks a “most selected by percentage owned” 30. Likely, everyone’s choices will converge, giving everyone pretty much the same team, and the one fool that chooses their favourite player, against all DT reason, scoops the pot! So, Kurt Tippet fans, your time is now!
Submitted by Supermufacho.