Fantasy Points = Real Life Wins /boggle

Hey everybody, the World Series starts today! The San Francisco Giants are coming off the momentum of having completed their postseason's second improbable three-game winning streak in the face of elimination while the Detroit Tigers have enjoyed six full days of inactivity while they waited for their Fall Classic opponent to be decided. I wonder who has the momentum going into this one? But Game 1 starter Barry Zito has his work cut out for him, since there's one very important game-changer that the Giants don't have: Justin Verlander.

But enough about what's happening in the actual world of baseball, let's talk more about the fantasy point system I've been using to evaluate players all season. It is indeed the same system I've been using for the last decade, but has now become infinitely easier to calculate thanks to ESPN's fully-customizable fantasy baseball program. I've mentioned those guys so many times that they should seriously be sponsoring this site by now... maybe my chances would be better if I used their standard scoring instead of my own system. But this is what I'm used to, and furthermore, I'm setting out to prove that the aggregate point totals of all the players on a team pretty accurately reflects that team's wins. First, here's how the point system shakes out:

I know baseball is a game of 3's, but this fantasy point system is essentially based on 5's. Batters get 5 points each for runs scored and driven in. Then we move to the basic building blocks of run production: Total Bases. Each one is worth 5 points: 5 for a single, 10 for a double, 15 for a triple, 20 for a home run. The progression makes sense - the more bases you reach per hit, not only do you put yourself in a better position to score, but the more potential runs you stand to drive in. Scoring position also plays a part in why steals are worth so many points - if you successfully steal a base, you at least put yourself in scoring position, whereas if you are caught, you erase yourself as a baserunner in a nonessential play. You lose a point every time you have a hitless at-bat, with an extra point lost for a strikeout (because in that situation you can't even advance the runner). A walk is worth three-fifths of a base hit, which is somewhat of a compromise, considering it puts you on base, but gives you very little chance of driving in a run in turn.

I'll be the first to admit that win-loss percentage is overvalued in this system, but in fantasy you want to make sure you're drafting the top pitchers on each team, which in terms of starters, are generally the ones in the position to rack up the most wins. The measure of the top relievers on each team is generally saves, hence the big point value there. Hey, I'm not saying it's a perfect system. Keeping with the motif of 5, each pitcher gains 5 points for each out pitched (which works out to 15 per inning) but loses 5 points for each baserunner allowed (via either a hit or walk). In another three-fifths compromise, a strikeout gains a pitcher an extra 3 points, which essentially makes each batter struck out worth 8 total points. Each run allowed costs a pitcher 10, because allowing runs is how a pitcher loses games after all.

Normally this point system is used to evaluate players personally, but for this experiment, I ran the combined stats of each team through the program to see how total points scored (both batting and pitching combined) corresponds with how many games that team won. However, I didn't use the exact same formula when calculating the team statistics - or, rather, I did and then I quickly realized where I went wrong and made some adjustments. The first adjustment was for batters: sure runs are important (scoring them is, after all, how you win games), but that's no reason to count them twice - i.e. for both runs scored and runs driven in. And while I don't think it's fair to exclude unearned runs as if they don't count, I removed RBI from the scoring. For pitchers, it doesn't make sense to include in the scoring system the very statistic you're trying to test for, so I removed wins and losses. There is also no reason a team should be rewarded for winning a close game, so saves had to go as well. This next chart contains the new points for each team in terms of both batting (BPts) and pitching (PPts). The most important column for our purposes are the last two: total points (TPts) and points divided by total team wins (TPts/W). Check out the results:

Lots of numbers, I know, but here's an overview of the important stuff we learned to look for in math class:

• BPts range from 13,787 (NYY) to 9,668 (HOU) with an average of 11,692. This number speaks volumes about the Yankees' offense, as their park ranked just 17th in terms of ESPN's park factors. The median is 11,803 (the average between KCR and TOR) - just slightly higher than the mean.
• PPts range from 12,354 (TBR) to 5,507 (COL) - a much larger spread. I guess the effects of the humidor in Denver are finally wearing off as we're starting to return to the batting/pitching discrepancies of the pre-2002 days. The average team scored 9,424 PPts, with the median at 9,657 (between PIT and CHW).
• TPts range from 24,105 (WAS) to 17,153 (HOU). Notice, much to my delight, that the highest point total corresponds to the team with the most wins and vice versa for the team with the fewest wins. In the old system (including all the stats) the Astros were still on the bottom, but the Yankees had the top spot. These two teams also occupy the top and bottom spots respectively in terms of points per game. The average is 21,116 and the median is 21,583 (ARI and LAD).

Now for the new and exciting category: TPts/W, which shows how many points each team scores per win. The first thing we notice is that the team with the fewest wins (55) has the most TPts/W (HOU with 311.873). However, BAL has the fewest TPts/W with 229.204 and they managed 6 fewer wins than the top squad (WAS with 98). This anomaly kind of makes sense, considering this was a team that was not expected to play well. Plus, they vastly outperformed their run differential - based on their runs scored vs. runs allowed, they should have managed only 82 wins. It might be the case that fantasy points have a better correlation with a team's expected performance than their actual performance, but that's a question for a different day.

Not that there isn't a decisive pattern emerging already as it stands: teams with the six lowest TPts/W all won at least 90 games, while the teams with the six highest TPts/W all won fewer than 70 games. It seems that teams that win more games use their points more efficiently in that it takes fewer of them to get a W. However, there's a lot more work to be done: the average of all 30 teams is 263.117 TPts/W, and the 3 teams closest to that figure won 88 (STL, 261.136), 73 (TOR, 265.110), and 90 (TBR, 265.378) games, respectively. So there's clearly not a precise correlation. However, that analysis will have to wait, as we've just had a sighting of the Justin Verlander we've come to expect in the playoffs in San Francisco. Let's watch some World Series!