dankok8
01-02-2022, 04:04 PM
The Titles Over Expected (TOE) metric has has been updated and there are now preseason and postseason versions colloquially known as pre-TOE and post-TOE.
The pre-TOE scores are based on the preseason odds and are the same as in the previous thread (http://www.insidehoops.com/forum/showthread.php?500379-Expected-Championships-Won-and-Titles-Over-Expected).
The post-TOE scores are new and based on championships odds prior to the beginning of the postseason. The betting odds are converted to expected titles (ET) by the same methodology as before. Note that post-TOE completely ignores seasons in which the player’s team misses the playoffs; obviously their title odds in that case are zero. Here are the post-TOE scores since the merger so the starting season is 1976-77 instead of 1984-85 in the other thread. As a result we have Magic and Bird's full careers as well as Kareem (Lakers career only), Moses, and Dr J added on here.
Postseason Odds -- Titles Over Expected of Some Legends: 1976-77 to 2020-21
https://i.postimg.cc/HxWRwTWY/Postseason-TOE.png
The advantage of using preseason odds is it allows for an assessment of team building. The disadvantage is it does a poor job of accounting for injuries and trades. Postseason odds are not affected by trades and regular season injuries but also don't account for injuries that occur in the playoffs. Postseason data also gives a "pass" to teams that miss the playoffs. However this is hardly a major issue because the vast majority of teams that miss the playoffs have very low title odds anyways. No player on the list is really underrated or overrated due to this factor at all.
Overall, preseason and postseason odds give surprisingly similar results.
My current project is focused on going even further back in time. I really want to include Bill Russell (probably #1 all time), Wilt Chamberlain (probably a massive negative score) and guys like Oscar, West, Baylor, Pettit, all of Kareem's career etc. I'd also like to integrate series odds and injuries into the model too. Based on how odds are reduced following major injuries perhaps that could be included in the model. It probably makes little difference on a career level because luck tends to even out for most but on a year by year basis, it can affect perceptions of who should be favored to win. This sort of thing has already been done so I don't have to start from scratch to make the model but it will take research to get the odds prior to the mid 70's.
The pre-TOE scores are based on the preseason odds and are the same as in the previous thread (http://www.insidehoops.com/forum/showthread.php?500379-Expected-Championships-Won-and-Titles-Over-Expected).
The post-TOE scores are new and based on championships odds prior to the beginning of the postseason. The betting odds are converted to expected titles (ET) by the same methodology as before. Note that post-TOE completely ignores seasons in which the player’s team misses the playoffs; obviously their title odds in that case are zero. Here are the post-TOE scores since the merger so the starting season is 1976-77 instead of 1984-85 in the other thread. As a result we have Magic and Bird's full careers as well as Kareem (Lakers career only), Moses, and Dr J added on here.
Postseason Odds -- Titles Over Expected of Some Legends: 1976-77 to 2020-21
https://i.postimg.cc/HxWRwTWY/Postseason-TOE.png
The advantage of using preseason odds is it allows for an assessment of team building. The disadvantage is it does a poor job of accounting for injuries and trades. Postseason odds are not affected by trades and regular season injuries but also don't account for injuries that occur in the playoffs. Postseason data also gives a "pass" to teams that miss the playoffs. However this is hardly a major issue because the vast majority of teams that miss the playoffs have very low title odds anyways. No player on the list is really underrated or overrated due to this factor at all.
Overall, preseason and postseason odds give surprisingly similar results.
My current project is focused on going even further back in time. I really want to include Bill Russell (probably #1 all time), Wilt Chamberlain (probably a massive negative score) and guys like Oscar, West, Baylor, Pettit, all of Kareem's career etc. I'd also like to integrate series odds and injuries into the model too. Based on how odds are reduced following major injuries perhaps that could be included in the model. It probably makes little difference on a career level because luck tends to even out for most but on a year by year basis, it can affect perceptions of who should be favored to win. This sort of thing has already been done so I don't have to start from scratch to make the model but it will take research to get the odds prior to the mid 70's.