These predictions are based on my own "secret sauce" method, which I
definitely think can be improved. I set the nominal coverage at 95% (meaning the way I calculated it the intervals should get it right
95% of the time), but based on testing the actual coverage might be at around 93%.
Intervals are inclusive.
\begin{array} {c c c}
Team & Lower & Upper \\ \hline
ATL & 66 & 86 \\
ARI & 69 & 90 \\
BAL & 71 & 92 \\
BOS &66 & 86 \\
CHC & 75 & 96 \\
CHW & 64 & 84 \\
CIN & 64 & 84 \\
CLE & 68 & 89 \\
COL & 62 & 82 \\
DET & 71 & 91 \\
HOU & 78 & 97 \\
KCR & 84 & 105 \\
LAA & 76 & 97 \\
LAD & 81 & 101 \\
MIA & 63 & 83 \\
MIL & 61 & 82 \\
MIN & 76 & 96 \\
NYM & 73 & 94 \\
NYY & 76 & 97 \\
OAK & 69 & 88 \\
PHI & 45 & 64 \\
PIT & 84 & 104 \\
SDP & 64 & 83 \\
SEA & 65 & 85 \\
SFG & 74 & 94 \\
STL & 89 & 109 \\
TBR & 73 & 93 \\
TEX & 67 & 87 \\
TOR & 75 & 94 \\
WSN & 77 & 98 \\ \hline\end{array}
Interesting features: my model thinks that the St. Louis Cardinals are the best team in baseball, and that the Phillies are the worse. I will add that the model gives a Phillies a true winning percentage of around 36%, (roughly 58 games out of 162), but they've been both bad and unlucky, and so are likely to finish with less than that.
Also note that even at this point in the season, it still can't predict whether all but four teams (the Cardinals, Pirates, Royals, and Phillies) will finish above 0.500 or not.
No comments:
Post a Comment