Projection comparison done by BP

[Steve-o]

It is a one year window. What I think both of us are saying is that we would like to see who was consistently the "best" at projecting stats over the last five years.

Although as GTWMA pointed out, there is no statistical significance between PECOTA and Marcel.

[GotowarMissAgnes]

I've mentioned this before, but I'd like to see more comparisons that go into details. I think we know by now that PECOTA is better than most overall, but can we say that one system is better than the other at predicting young players? Old players? Sluggers? Speedsters? Comparisons like that would help determine whether one system is best overall, or whether different systems do well in different areas.

[Field]

I've mentioned this before, but I'd like to see more comparisons that go into details. I think we know by now that PECOTA is better than most overall, but can we say that one system is better than the other at predicting young players? Old players? Sluggers? Speedsters? Comparisons like that would help determine whether one system is best overall, or whether different systems do well in different areas.

Very interesting. Something like this could be really valuable, although there's no guarantee that the projections will be just as accurate in the future due to the human factor unless the projections is purely based on stats/mathematics. But it certainly would be a nice tool to help with rankings and especially sleepers.

Who's ready to start running some regressions?