Post by eric on Feb 17, 2015 21:29:02 GMT -6
No hay embrague. There is no clutch.
We can improve our predictions of playoffs matchups by a small but statistically significant degree by ignoring the outcomes of regular season close games (decided by five points or less) entirely. It stands to reason that we can improve our predictions of final season standings by doing the same for close games played thus far. Arithmetically, we will predict final wins by:
W + (82 - GP) * (Close% * 0.5 + (1 - Close%) * NonCloseW%)
We use each team's Close% because a team with a very good or very poor record is less likely to play close games. Because we're already 5/8ths of the way through the season and because those games still count, even teams with massively inflated luck (Warriors, Rockets, Heat) or deflated (Suns, Thunder, Hornets) are pretty set in their seeding. The only changes are the Mavericks dropping from 5th to 7th in the West and the Hornets taking the 7th seed outright in the East.
Looking at non-close record further illustrates the disparity between East and West. There are three teams with 70%+ overall: Warriors, Hawks, Grizzlies in that order. If we look at non-close records above 70%, we find five teams: Hawks, Warriors, Grizzlies, Trail Blazers, and Spurs in that order. Another reason to keep an eye on the Spurs... after shattering the record for fewest close games of any champion in NBA history last year (11 to the 2000 Lakers' 14) and shattering the record for highest winning percentage in those games (91% to the 1987 Lakers' 80%), the Spurs have already played 18 close games this year and gone a pedestrian 9-9 in them. Obviously the Spurs did not forget how to be clutch.
There is no clutch.
All we're seeing is the Spurs being historically lucky last year and modestly unlucky so far this year. Give them their close performance last year and they'd be at 39 wins, tied with the Grizzlies for the 2 seed. That's what luck is. Significant, and transient. Keep an eye on the Spurs.
We can improve our predictions of playoffs matchups by a small but statistically significant degree by ignoring the outcomes of regular season close games (decided by five points or less) entirely. It stands to reason that we can improve our predictions of final season standings by doing the same for close games played thus far. Arithmetically, we will predict final wins by:
W + (82 - GP) * (Close% * 0.5 + (1 - Close%) * NonCloseW%)
We use each team's Close% because a team with a very good or very poor record is less likely to play close games. Because we're already 5/8ths of the way through the season and because those games still count, even teams with massively inflated luck (Warriors, Rockets, Heat) or deflated (Suns, Thunder, Hornets) are pretty set in their seeding. The only changes are the Mavericks dropping from 5th to 7th in the West and the Hornets taking the 7th seed outright in the East.
Looking at non-close record further illustrates the disparity between East and West. There are three teams with 70%+ overall: Warriors, Hawks, Grizzlies in that order. If we look at non-close records above 70%, we find five teams: Hawks, Warriors, Grizzlies, Trail Blazers, and Spurs in that order. Another reason to keep an eye on the Spurs... after shattering the record for fewest close games of any champion in NBA history last year (11 to the 2000 Lakers' 14) and shattering the record for highest winning percentage in those games (91% to the 1987 Lakers' 80%), the Spurs have already played 18 close games this year and gone a pedestrian 9-9 in them. Obviously the Spurs did not forget how to be clutch.
There is no clutch.
All we're seeing is the Spurs being historically lucky last year and modestly unlucky so far this year. Give them their close performance last year and they'd be at 39 wins, tied with the Grizzlies for the 2 seed. That's what luck is. Significant, and transient. Keep an eye on the Spurs.
