What If The MLB Were Just a Big Fantasy League?

Sixteen million people play fantasy sports yearly, according to the Fantasy Sports Industry Trade Association. Thus, it shouldn't surprise most when I say "Rotisserie." No, not the cooking style—I'm talking about the fantasy baseball scoring format (refer to Yahoo!'s summary of Rotisserie scoring if you are unfamiliar).

As the end of the baseball season nears, it got me thinking: What if Major League Baseball were one big Rotisserie League? And is there any difference between a team's Roto ranking and their ranking based on regular season wins?

So I did just that. I took each team's stats (runs, RBI, HR, batting avg., and stolen bases, for batting; and wins, saves, ERA, WHIP, and strikeouts, for pitching; those are the standard 5x5 categories in roto leagues) as of Friday, Sept. 26, ranked them, and figured out their Roto value.

Roto value compared to wins

Team	RV	W
Chicago Cubs	258	96
Boston	252.5	94
Tampa Bay	232	96
Philadelphia	217	90
LA Angels	215.5	99
NY Mets	211.5	88
NY Yankees	207.5	88
Milwaukee	201	89
Chicago Sox	188	86
Minnesota	187.5	87
Toronto	177	85
LA Dodgers	175	83
Houston	166	85
Texas	161.5	78
St. Louis	156	84
Florida	152	83
Arizona	147.5	80
Cleveland	137.5	80
Detroit	134.5	73
Colorado	132.5	74
Cincinnati	120.5	74
Baltimore	117	67
Kansas City	117	74
San Francisco	113.5	71
Oakland	105	75
Atlanta	100.5	71
Seattle	82	59
San Diego	69	62
Pittsburgh	65.5	66
Washington	50	59

The Angels have the league's best record, but are only fifth in roto value, and far behind the Cubs and Red Sox.

Now here's the better question: Does Roto Value have any correlation with wins?

Using the past three years' stats (trust me, I would have gone back to 2000, but it takes forever to calculate the RV, mainly due to ties), I figured out each team's RV compared to their win totals. The regression equation to find predicted wins using RV is Wins =~ 0.18 * RV + 53 .

That yields a 3.92 RMSE in the past three years, compared to a 4.10 RMSE of the Pythagorean Formula in the same time frame.

The sample size is very small, but it does seem like RV can predict wins as good as, or even better than, the Pythagorean Formula. Until I find the RV for teams going back to the 1990's or so, however, I'm not going to make any conclusions based on it.