Saturday, January 16, 2010

Saints Are Apparently Underrated In Snapdraft

I did some analysis for the teams selected by players in 3 or 4 of my 25 person contests from the Sanpdraft playoff contests that started last week. I wanted to see how popular the players from each team were among those that used the 'correct' strategy of picking players only from two teams (one AFC and one NFC). As expected, there were four very popular teams, with players from the other four rarely being selected. The surprise is that the Saints were not one of the four that people liked! The Chargers, Colts, Vikings, and Cowboys were roughly equally popular - each selected 10-15 times in my sample. The Saints were only selected by 3 people, and the remaining teams (Cardinals, Jets, and Ravens) were even less popular. I used that information to guide my picks for this week, loading up on the Saints on the assumption that if they do win, I'm not going to be competing with anyone else for the top spot in my contests. I understand that the Saints players at RB, TE, and possibly WR are not the very best for fantasy, but I think that the extra game would still give them a good chance of outscoring better players...particularly given the 3 point bonus for each player on a winning team.

Monday, January 11, 2010

Snapdraft Pricing and Lineups

As I was looking through the teams for the Snapdraft football playoff contests, I found one that had an identical lineup to my team in the same contest. This problem is more common at Snapdraft than it is at other contest sites, such as Fanduel, for two reasons: Snapdraft has the largest multiplayer contests, and they never update player prices during the season. This problem becomes amplified during the playoffs, when there are fewer players to choose from.

So, should I care if somebody is using the same lineup as me? My initial instinct was that this is a disaster. After all, if I win, I'm going to have to share the prize. But let's think about that a bit. In the contest where this happened, the other player and I both selected a lineup made up only of players from the Saints and Colts. I've discussed in previous posts how the structure of Snapdraft's playoff contest virtually guarantees that if those two teams make it to the Super Bowl, a lineup made up of those two teams is going to be the winner. In this case, nobody other than the two of us limited themselves to Saints and Colts. Let's assume that there's a 20% chance that we'll have a Saints-Colts Super Bowl this year. So with identical lineups, there's a 20% chance that I'll win $50 (half of the first prize). Now imagine, that instead of Hartley (the Saints kicker), my opponent had chosen Stover (the Colts kicker). One of the two of us would still win...but we definitely wouldn't share the prize, and the total chance that one of us wins wouldn't increase. I'd have only a 10% chance of winning a $100. Either way, my 'expected' winnings would be $10. And the variance is lower with identical lineups, which is generally a preferable situation.

In other words, having identical lineups isn't as bad as it sounds. Now imagine that five other players had various combinations of Saints-Colts players. Now THAT would be bad...whether or not any of the lineups matched mine exactly.

Sunday, January 10, 2010

Snapdraft Playoff Football Strategy

As I mentioned in my previous post, I think there was a great profit opportunity in the Snapdraft playoff football contests. My thinking was that the correct strategy in large multiplayer contests is always to create a much variance as possible, since it takes an extreme score to win these winner take all contests. Since you get nothing for coming in 2nd (or 10th), there's no point playing it safe. You can always create some variance in weekly contests, by taking players on the same team, whose scores are likely to correlate with eachother...for example, the QB and WR from the same team. In Snapdraft, that correlation (and thus variance) is amplified by the fact that all players on the winning team score an extra three points. In Snapdraft playoff football, it's further amplified by a great deal by the fact that players on the same team either will or won't advance to the next round together, making a team full of players on a Super Bowl winning team almost unbeatable. Since you can't pick a team of players all from one team, the correct strategy was to pick from two teams...ideally those who will make it all the way. I assumed that for the most part, others would not be following this plan, meaning that I didn't have to take into account which team combinations would be most popular. Otherwise, I might have shied away from the best teams.

So I entered $50 of $1 and $5 contests. Almost all were either 25 person contests or 101 person contests, with just a couple of 15s and 20s mixed in. My opponents managed to surprise me with how smart they were and how stupid they were at the same time! I'm generally unable to check live scoring in the 101s without freezing up my computer, but I looked through rosters for all teams in 3 of the 25 person contests. A relatively large percentage of people used the correct strategy - players from just two teams, or possibly two team plus Adrian Peterson if they obviously didn't think one of their teams had a viable running back. However, they almost universally did not pick a New Orleans - Indianapolis matchup! Considering that the Las Vegas odds have those two teams as the favorites, they should have been the most popular pick. Do people think that they know better than the odds? If so, they should reconsider...while the favorites won't always win, the odds are going to do a better job of making predictions in the long run than just about anybody. The only other possibility is that everybody incorrectly assumed that "everybody else will take NO-IND, so I'll be sneaky and take someone else".

Thursday, January 7, 2010

Snapdraft NFL Playoff Contests: Major Profit Opportunity

I think there's a major profit opportunity in the NFL playoff contests that start this weekend at Snapdraft. These contests last through the Super Bowl. Scoring covers the entire playoffs. I'm sure a lot of casual playoffs will miss that minor detail, and fail to plan accordingly. Even those who are aware of the rules are likely to misjudge the correct strategy to maximize their chances. I'll try to write about these contests in more detail tomorrow morning.

Saturday, January 2, 2010

Snapdraft Scoring: NBA Basketball

Here are the scoring rules posted for NBA basketball contests on the Snapdraft site:

2 points - 2 Pt. Field Goal
3 points - 3 Pt. Field Goal
1 point - Free Throw
1 point - Assist
1 point - Rebound

2 points - Steal
2 points - Blocked Shot
1 point - Rebound

Overall, it looks like there's more of an emphasis on good points scorers than in other formats. Categories like steals and blocked shots, where the difference between a stud and a dud might be an average of 1.5 vs. an average of 0.5 are almost irrelevant.

Friday, January 1, 2010

Quarterback, Receiver Combos

I've mentioned before that in multiplayer Snapdraft contests, you want high variance in your scores. One way of doing that is picking players from the same team, whose scores will tend to be correlated. That's particularly true because of the 3 point bonus that players on winning teams get. Either everyone on that team gets the bonus, or they don't. The positions whose scores correlate best with each other are obviously quarterbacks and receivers on the same team. Three combos who I like this week are Favre-Rice, Schaub-Johnson, and Roethlisberger-Holmes. All three have some incentive to try to win and are favored to win.