The NFL preseason is already underway, and fantasy football players are prepping for the most critical part of the season–the draft. While some people choose to rank players based on gut feeling and intuition, I like to go with a more mathematical approach, making projections based on past performance and historical trends.
Last year, my projections beat out three of the top fantasy football Web sites, CBSSportsline, ESPN, and FFToday. This year, my projections should only get better. Here are what’s changed in my projections from last year to this:
Read More »
Sunday, I released my MLB projections based on a computer model. The individual projections can then be added up by team to create projected standings.
Before I add up any projections, I first adjust any player’s projected playing time that may be off. As I mentioned before, the projected playing time only looks at a player’s past playing time and adjusts for age. So a player such as Jose Reyes, who had had 700 plate appearances in his four seasons as a starter before missing 126 games last year, may have a lower projection than in reality. I also look for players like Rangers OF Julio Borbon, who will receive a starting role for the first time in his major league career.
Once each player is assigned their proper role and playing time forecast, I add up the projected stats for each team. If a team comes up over 4120 outs (AB minus H plus CS) or 1440 IP, I prorate their stats down to those levels; if a team falls short of those benchmarks, I assign replacement-level production for the remaining outs.
I use a Base Runs equation to calculate projected runs scored for offense, and for the pitching staff I simply look at runs allowed. Win percentage is found using Pythagenpat . Then I adjust each team’s win percentage by their schedule to come up with a final tally.
Here are the results:
Read More »
In every aspect, these projections are better than last year’s. Why?
1. Custom weights. Each stat is weighted based on error tests from 1970 onward; for instance, BABIP for hitters is weighted at 0.88, while strikeout rate, which is more stable year-to-year than BABIP, is weighted at 0.49 (i.e., 2009 has a weight of 1, 2008 has a weight of 0.49, etc.). I use the past four years’ stats for hitters and three years for pitchers.
Read More »
How accurate were my objective FEIN projections ? You may think that a completely automated football forecasting system won’t be as accurate as human projections such as CBS Sportsline’s , FF Today’s , or ESPN’s . Let’s see how each set of projections stacked up in accuracy for the 2009 season.
I looked at quarterbacks with 200 pass attempts, running backs with 100 rushes, wide receivers with 40 catches, and tight ends with 30 catches this season (selecting only those with a projection from each system). There were 33 QBs, 47 RBs, 54 WRs, and 25 TEs who met these criteria.
I first adjusted each system’s projections up or down to make the average projection equal to actual average of the players in the sample. (For instance, if ESPN had projected 4.5 yards per carry and the average was 4.2, I subtracted three-tenths of a yard from each player’s projection.)
Then I compared each player’s adjusted projection to his actual number. I squared the difference between the two and weighted that number based on his actual number of attempts or catches, and took the square root of the average for each system.
Here are the results.
Read More »
Through six weeks of the season, most fantasy football owners believe they have enough of a sample size to get a good gauge on a player’s value—Steve Smith of the Giants is good; Steve Smith of the Panthers is bad.
My preseason fantasy football projections forecasted that Vincent Jackson would be a top-10 wide receiver (he’s currently No. 11) and that Anquan Boldin would be outside the top 15 (he’s No. 35). Of course, there were some picks that haven’t played out as predicted, such as Peyton Manning’s projected No. 6 ranking or Clinton Portis’s No. 7 spot at their respective positions.
With that said, a player isn’t only as good as his 2009 stats. Just because Cedric Benson is currently the No. 6 running back doesn’t necessarily mean he’ll be a top-10 back in the final 10 weeks. Most preseason predictions ranked Benson outside of the top-30 running backs, and that should be taken into account when you’re considering selling high on Benson or trading for him.
Read More »
Vegas Watch published its third annual preseason predictions review today and again found that the computer projections are more accurate than human predictions. CAIRO comes out on top this season, with an average root-mean-square-error of 9.24 games, its closest competitor having an RMSE of 9.60.
The FEIN projections I released in March and April fell right in between the CHONE and THT projections in terms of accuracy, with a 9.85 RMSE. But I also said that the Royals’ forecast was optimistic by one or two games since I had accidentally projected them as an NL team; I had estimated a drop of 1.5 wins for Kansas City based on the error, which would lower FEIN’s RMSE to 9.76 , in third place just ahead of Marcel.
After the season, I’ll compare the individual hitter and pitcher projections to what actually happened on the diamond.
Team-by-team forecasted and actual results after the jump.
Read More »
What if I told you that Adrian Peterson isn’t as good as his stats say?
My reasoning is the Curse of the Leading Rusher. You’ve never heard of it before, but it’s an obvious trend. Since 1980, the NFL’s leading rusher has seen his rushing yards fall by 489 yards and his YPC by almost half a yard just one season later. Only six of the 31 leading rushers even increased their rushing yards the following season, and nine had less than 1,000 yards.
Convinced? You shouldn’t be. Their decline is nothing more than regression to the mean and a lack of sample size. Let me explain.
Read More »
It’s hard to make accurate pre-season predictions. Last year, Sports Illustrated’s record predictions had a root-mean-square-error (RMSE) of 3.50 wins. Heck, if you projected each team to win eight games, you’d be off by 3.27 wins each. And if you use the method below, you’d be 3.14 wins away on average.
(DISCLAIMER: I don’t think that these predictions are superior to or any more correct than anybody else’s.)
Looking at all teams since 1994, I ran a regression on their Year X stats to predict their Year X+1 wins. The formula is shown below. The r-squared was .107, with an RMSE of 2.84.
Read More »
John Benson over at THT calculates Marcel projections as if a human controlled the weighting—a yearly weighting of 80%/15%/5%. He says that Jason Bartlett’s projected OPS would be 80 points higher with the more nearsighted weighting.
Another guy with a career year is Ben Zobrist, who is third in the AL with a .961 OPS. Benson calculates a 49-point oversight with the human weighting as opposed to Marcel’s weighting.
Three weeks after saying he won’t join the Minnesota Vikings, Brett Favre has joined the Minnesota Vikings. The contract is between $10 and $12 million, the same as Favre’s $12 M contract with the New York Jets last year.
Here’s Favre’s 2009 FEIN projection (assuming 500 pass attempts), in comparison to Sage Rosenfels’.
Read More »
