Sunday, March 13, 2011

Are those using the 'Generalv' strategy in one day tournaments exploiting a loophole or making a mistake?

Joe Vance (aka 'Generalv') is one of the top players in daily fantasy sports. He was one of the top players at Fanduel during the 2010 MLB season, one of the top NFL players, and put an exclamation point on his status among the best players when he took home $25,000 for winning the live FFFC finals in Las Vegas. In January he started playing in daily NBA contests as well, and by all accounts is a very tough opponent in those two.

He recently got some additional attention by employing a strategy in the $4500 Friday NBA Dunk that some people felt exposed a loophole in the multi-entry large tournament format. He had two identical entries finish in the top 5 of the contest, winning large prizes with each. The thinking of those who considered this a loophole seems to be that for a player with a substantial edge on the field, allowing them to enter two identical teams essentially gives them an opportunity to multiply their edge as many times as they want.

I disagree. In fact, I think these people have things backwards. If there's any loophole or 'trick' to taking advantage of the ability to enter the large field tournaments, it's making sure that your teams aren't too similar to each other.

When I'm face with a complex scenario like this, I often find it helpful to look at the math for an extreme scenario with some of the same elements. The large field tournaments have top heavy payout structures, with around 10% of the field taking home prizes, and the top few finishers (especially the winner) taking home the bulk of the money.

I'll use an even more top heavy structure to make my point. Imagine a 100 person contest with $10 entry fees. Let's say the winner gets $700 and the second place finisher gets $300. Imagine you're an average player, with 1% of winning and 1% chance of finishing second. So your 'expected value' from this tournament is exactly $10: .01X$700 + .01X$300. Let's imagine that you enter twice, but use two completely unrelated (but equally good) lineups. You'd expect to win $20 of prizes for your $20 of entry fees. Now imagine that your two lineups are identical. So there are 99 unique lineups in the contest. So there's a 1 in 99 chance that you have the best lineup. If you do, you'll win the entire $1,000 (your two teams would split the top two prizes). There's also a 1 in 99 chance that you have the 2nd best lineup, in which case you'd win $300 (your two teams would split the 2nd place prize). So in this scenario, you'd only get an average of $13.13 of prizes for your $20 entry fee!

That's a pretty good illustration of just how bad using duplicate lineups in these contests is if you're an average player and the prize pool is top heavy enough. The math obviously changes depending how big an edge your best lineup has on the rest of the field, and how much less top heavy the prize structure is. I'll freely admit that I'm not sure how to do the math to find the exact thresholds. But I'm fairly confident that nobody's edge is large enough to make it a good strategy to use identical lineups in these tournaments. Especially when you consider that a player of Generalv's ability could certainly come up with an alternate lineup that, while not quite as 'good' as his optimal one, would still have a substantial positive expectation.

Not only do I think that this strategy is not exploiting a loophole, but I think that those following it a making a pretty big mistake.

No comments:

Post a Comment