by Josh Fyman
A cliché that players, especially those on wild-card teams like to dredge up this time of year is that records are no longer relevant, that everybody is now 0-0. The whole idea sounds like bland motivational speak. After all, wouldn’t it stand to reason that teams with superior records would be more likely victors in playoff games? Well, I’m here to tell you that this is only slightly true. In fact, the cliché that all records are dialed back to zero has more credence when one looks at the numbers.
The first, most rudimentary analysis would be to correlate a team’s winning percentage in the playoffs with their records during the regular season. Since the abbreviated playoffs provide a small sample size, I went back five years and correlated the regular season winning percentages of all playoff teams with their winning percentages in the playoffs. The results are illuminating. There is only a 0.14 correlation between regular-season record and playoff record, which is barely significant. Granted that the data are skewed somewhat by the fact that teams that earn a bye do not get to accumulate as many wins, and therefore, suffer a worse playoff winning percentage in the event that they lose. This problem is erased, however, by comparing wild-card teams to each other and bye teams with each other. When examining only bye teams, the correlation between regular season and postseason winning percentage actually drops to 0.10, which is not significant.
Another school of thought is that the teams that finish the hottest are best suited for playoff success. To test this theory, I correlated playoff teams’ December records over the past five years with their playoff success. Turns out that there is only a 0.18 correlation between teams’ performances at the end of the season and their performances in the postseason. Bye teams’ performances are a little more predictive of their playoff performances, with a correlation of 0.23, but that still is not too much.
To drive home the point, over the past five years, in playoff games in which two teams with different records played each other, do you know the likelihood of the team with the better record winning? Ready? A whopping 51.8% probability. So, if you’re picking playoff games based on which team has the better record, you’re just about as well of flipping a coin.
Not only do teams with better records fare only slightly better in playoff games, but it doesn’t even much matter how much better one team’s record is than the other. In a correlational analysis, there was no significant relationship between how many games separated the two opposing teams’ records and the probability of an upset (upset being defined here as the worse record beating the better record). In short, a game with records separated by one game was no more likely to result in an upset than a game with records separated by three, or even four, games.
Is there no way to predict playoff games with some accuracy? Of course there are models that will provide results better than 50/50. Looking at records, however, offers no help whatsoever.
Tuesday, December 30, 2008
| [+/-] |
Trying to Predict the Playoffs? Forget About Records |
Saturday, December 27, 2008
| [+/-] |
Straight Up Winners |
by Bob Burns
It is my contention that the box score data and stats from NFL games are the result of the down, distance, field position, score and time remaining in the game. As they are the result, they can not be the cause of winning or losing. If they are the result, they cannot be the independent variables in a regression that forecasts either scores or wins and losses. I am convinced that the data and stats are not useful for forecasting based on my trying to use them in a regression without success. But Brian Burke has a model that forecasts the probability of wins and losses and the model is based on box score data and stats. This presents a problem, how to explain Brian’s success.
In my opinion, Brian has had success because the data he is using are proxies for wins and losses. If I am correct, I should be able to construct a model based on won/lost records that should outperform Brian’s model because the actual data should out perform the proxy data.I used data from week 5 to week 17 from 1996 to 2006 build the model. The model was set up for home teams only, and used the difference in the win percentage between the home team and the visiting team. A formula was determined that the probability of a win is .60 plus .43 times the win percentage difference.
For example… in week 16 2008 the Giants were home against the Panthers. Both team had the same win percentage, so the Giants had a estimated win probability of .60. Another example is the Redskins at home is week 16 vs the Eagles. The Skins record was .50 and the Eagles was .57. The Skins probability would be calculated as .60 plus (.43 time –0.07) which works out to .57. Both the Giants and the Skins won. [ its my write up so I can pick winning examples LOL ]
That method was then applied to week 5 to 17 in 2007 and 2008 and compared to the same weeks as Brian’s model. Now the question is how to compare results.
First lets look at who had more correct. Brian’s model had 236 correct (his pick with the higher probability won) and 121 losses for a 66% win rate. Mine showed 233 winners and 124 losers for a 65% win rate. But that doesn’t reflect the differences in win probability. I took all the games rated .90 and above, computed an average probability for those games and an average won/lost percentage for that group. I did the same with the .80’s, .70’s, etc. I then ran a correlation between the average forecasted probability and the actual win rates. Brian’s model had a 0.922 correlation and mine had a 0.979 correlation.
There where 358 games and 357 decisions (1 tie). I used the same procedure as above except but group in groups of about 50, starting with the highest forecasted probability. Brian’s model had a 0.948 correlation, and my model had a 0.989 correlation.
So the models are very close in actual results. Are the results close enough to be essentially the same model?
Thursday, December 25, 2008
| [+/-] |
NFL Luck? |
by Ira Haron
The Luck thing.
Here's my 16th week 16¢ worth:
Here's "Luck" according to me (so what?), plus a few extras.
Sorted from Best-to-Worst, then by Luck, is my number-crunched relative "Power", "Influence" and (if betting was legal) my "Index" which, after combining with such logical considerations as home/away, injuries, who's hot, who's not, who cares, who don't, who's angry, who's on the SI cover, and other important "tangible" football fudge factors could, possibly, maybe, be used to update and augment the numbers to then compare the opponents and lay a wager (?) when judged vs. the LV spread. My Luck factor works for me because of the relatively few numbers of NFL games. If played to infinity, Luck would asymptote for everyone to 1.00. This way, I see the "skew" which is what I wanted.
My SB column is what the rest of my calcs (not sent) determine who is a real SB contender. I can't figger why the Ravens are not in the elite group but it makes my stats even more interesting and SB-ing more exclusive. No, I don't know what the Raven's' flaw is (yet). I normally do 65-70% W-L and 55-60% ATS with "Locks" a little (but not much) better. And, no, I haven't bet on a game since my wife and I vacationed in
OK. Back to Luck:
Several things that I have observed in all sports is that "Good teams, like people, make their own Luck"; "Luck favors the prepared mind'; "Opportunity knocks and they answer the door"; "Bad teams stand around looking to find ways and excuses to lose"; so, call it Luck or whatever, "Good teams have it, make it, and take advantage of it". Here's another , "Worry is rust upon the blade". Some teams seem lucky for a day but it doesn't last, does it? Over the course of the season, "Good teams find ways to win". I also say, and say again, "Don't ask a bad team to win for you", and "In the NFL, more games are Lost than Won (Luck?)." Maybe
Rank Team Luck Power Inf Index SB? 1 Tennessee Titans 1.4 2.4 1.1 26.9 SB 2 N.Y. Giants 1.3 2.0 0.9 23.1 SB 3 Baltimore Ravens 1.1 1.8 0.9 20.2 - 4 Pittsburgh Steelers 1.3 1.7 0.8 19.4 SB 5 Carolina Panthers 1.3 1.6 0.7 18.4 SB 6 New England Patriots 1.2 1.5 0.7 17.8 - 7 Philadelphia Eagles 1.0 1.5 0.7 16.9 - 8 Indianapolis Colts 1.4 1.4 0.6 16.7 - 9 Atlanta Falcons 1.2 1.4 0.7 16.2 - 10 N.Y. Jets 1.1 1.3 0.6 14.9 - 11 New Orleans Saints 1.0 1.3 0.7 14.8 - 12 Minnesota Vikings 1.1 1.2 0.6 14.2 - 13 Tampa Bay Buccaneers 1.1 1.2 0.6 14.1 - 14 Dallas Cowboys 1.1 1.2 0.6 13.7 - 15 Miami Dolphins 1.3 1.2 0.6 13.6 - 16 San Diego Chargers 0.9 1.1 0.6 13.4 - 17 Chicago Bears 1.1 1.1 0.6 13.3 - 18 Buffalo Bills 0.9 0.9 0.5 10.6 - 19 Arizona Cardinals 1.1 0.9 0.5 10.3 - 20 Green Bay Packers 0.6 0.9 0.6 10.1 - 21 Washington Redskins 1.1 0.8 0.4 9.4 - 22 Denver Broncos 1.1 0.7 0.4 8.5 - 23 Houston Texans 1.0 0.7 0.4 8.5 - 24 San Francisco 49ers 0.9 0.6 0.4 7.4 - 25 Jacksonville Jaguars 0.7 0.6 0.4 6.8 - 26 Seattle Seahawks 0.6 0.4 0.3 4.8 - 27 Cleveland Browns 0.6 0.4 0.3 4.6 - 28 Oakland Raiders 0.7 0.3 0.2 3.1 - 29 Kansas City Chiefs 0.3 0.2 0.3 2.4 - 30 Cincinnati Bengals 0.6 0.1 0.2 1.4 - 31 St. Louis Rams 0.4 0.0 0.2 0.2 - 32 Detroit Lions 0.0 0.0 0.2 0.0 -
Now sorted by Luck:
| Rank | Team | Luck |
| 1 | | 1.4 |
| 8 | | 1.4 |
| 2 | N.Y. Giants | 1.3 |
| 5 | | 1.3 |
| 15 | | 1.3 |
| 4 | | 1.3 |
| 9 | | 1.2 |
| 6 | | 1.2 |
| 14 | | 1.1 |
| 22 | | 1.1 |
| 17 | | 1.1 |
| 21 | | 1.1 |
| 12 | | 1.1 |
| 10 | N.Y. Jets | 1.1 |
| 13 | | 1.1 |
| 3 | | 1.1 |
| 19 | | 1.1 |
| 7 | | 1.0 |
| 23 | | 1.0 |
| 11 | | 1.0 |
| 18 | | 0.9 |
| 16 | | 0.9 |
| 24 | | 0.9 |
| 25 | | 0.7 |
| 28 | | 0.7 |
| 20 | | 0.6 |
| 27 | | 0.6 |
| 30 | | 0.6 |
| 26 | | 0.6 |
| 31 | | 0.4 |
| 29 | | 0.3 |
| 32 | | 0.0 |
Posts
