| Original Source | Author | Published |
|---|---|---|
| Correlating player offensive stats against RE24 — TangoTiger on Twitter | TangoTiger | 2025-07-21 |
Tangotiger @tangotiger 2025-07-22
1.7*OBP + SLG is best combination of the two, in terms of matching to runs created AT THE PLAYER LEVEL
Team-level correlation, as fun as it is, is simply too limiting. You can’t extrapolate team-level correlations to Aaron Judge or any player outside of team-levels of OBP, SLG
Tangotiger @tangotiger 2025-07-22
And thank you for running those different versions, that even all absurd versions, including OBP + 4 * SLG still gives you a correlation close to wOBA
It shows that this is just “big good, low bad”, and there’s not enough team-level variations to tease anything out of OBP, SLG
Michael Kasper @KasperStats 2025-07-22
Interesting! So is there still value in the team-level correlations? Or should it all be done on a player-level? I have a whole bunch of player-level stuff I’m preparing tonight for tomorrow…
Tangotiger @tangotiger 2025-07-22
People LOVE doing it at the team-level (and interestingly they only use the batting side and NEVER the pitching side)
But the reality is, it’s too limiting. The issue with players is: what are you correlating against? So that becomes problematic.
In the end, we know how…
Tangotiger @tangotiger 2025-07-22
… runs are created, it’s RE24. And THAT is what you are trying to correlate against. Or RE24/PA to get a rate stat.