Doug Drinen
Moderator
Simply comparing playoff records doesn't give a complete picture because it doesn't take into account the strengths of the teams that the coaches were working with. For example, George Seifert was 10-5 in the playoffs during his career while Tony Dungy is 5-7. But Seifert's teams were obviously much stronger, on the whole, than Dungy's. By my count, Seifert's teams were the clear favorites in 11 of those 15 games.
Of course, it is to Seifert's credit --- at least to some extent --- that his teams were such heavy favorites. That means they were great teams, which is evidence that Seifert is a good coach. But that's not what I'm trying to measure here. What I want to know is: what coaches have done the best job of winning games they weren't supposed to win in the playoffs?
So here's the methodology:
1. look at every playoff game since the merger (1970--2004)
2. make an estimate of each team's probability of winning that game (details below for the
crowd).
3. for each coach, add his team's estimated win probability for each game. This gives the number of games he was "supposed to" win.
4. count up the number of games his teams actually won.
5. compare his actual record to his expected record.
For instance, George Seifert's actual record was 10-5. His expected record is also 10-5. So he did no better and no worse than expected. Likewise, Dungy's actual is 5-7 and his expected record is also 5-7. So he too, has been neither a good or a bad playoff coach (at least according to this methodology).
Here is the full list (more fine print follows):
extraCoach Expected Actual wins----------------------------------------Chuck Noll 12-12 16- 8 +3.9Bill Belichick 6- 5 10- 1 +3.9Joe Gibbs 13- 8 16- 5 +3.2Tom Landry 16-15 19-12 +3.1Jimmy Johnson 6- 7 9- 4 +2.8Brian Billick 3- 4 5- 2 +2.2Bill Walsh 8- 6 10- 4 +2.1Bill Parcells 9- 9 11- 7 +2.0Tom Flores 7- 4 8- 3 +1.5Bum Phillips 3- 4 4- 3 +1.4Jerry Burns 2- 4 3- 3 +1.3Raymond Berry 2- 3 3- 2 +1.3Mike Shanahan 6- 5 7- 4 +1.2Jon Gruden 4- 3 5- 2 +1.1Barry Switzer 4- 3 5- 2 +0.9John Madden 7- 7 8- 6 +0.8Dan Reeves 10-10 11- 9 +0.8John Robinson 3- 7 4- 6 +0.7Don McCafferty 3- 2 4- 1 +0.7Herman Edwards 1- 4 2- 3 +0.5Ray Malavasi 3- 3 3- 3 +0.4Jerry Glanville 3- 4 3- 4 +0.4Jeff Fisher 5- 4 5- 4 +0.4Mike Holmgren 9- 8 9- 8 +0.3Marv Levy 11- 8 11- 8 +0.3Tom Coughlin 4- 4 4- 4 +0.3Sam Wyche 3- 2 3- 2 +0.2Andy Reid 7- 5 7- 5 +0.0Dick Nolan 2- 3 2- 3 -0.0Dave Wannstedt 2- 3 2- 3 -0.1Ted Marchibroda 2- 4 2- 4 -0.2George Seifert 10- 5 10- 5 -0.2Tony Dungy 5- 7 5- 7 -0.2Art Shell 2- 3 2- 3 -0.4Red Miller 2- 3 2- 3 -0.5Bobby Ross 4- 4 3- 5 -0.6Steve Mariucci 4- 3 3- 4 -0.6Bud Grant 9- 9 8-10 -0.7Don Coryell 4- 5 3- 6 -0.8Jim Fassel 3- 2 2- 3 -1.0Mike Martz 4- 3 3- 4 -1.0George Allen 3- 4 2- 5 -1.0Dick Vermeil 7- 4 6- 5 -1.0Chuck Knox 8-10 7-11 -1.0Mike Sherman 3- 3 2- 4 -1.2Wayne Fontes 2- 3 1- 4 -1.5Mike Ditka 7- 5 6- 6 -1.5Jack Pardee 3- 3 1- 5 -1.8Dennis Green 6- 6 4- 8 -2.0Don Shula 19-12 17-14 -2.1Bill Cowher 10- 7 8- 9 -2.3Jim Mora 3- 3 0- 6 -3.3Marty Schottenheimer 9- 8 5-12 -3.7Fine print:1. only games played since the merger are counted. So e.g. Don Shula's record in the 1960s is ignored.
2. only coaches with at least 5 playoff games are counted.
3. the "expected records" listed above are rounded to the nearest integer.
4. the formula for estimated win probability is:
WinProb = 1 / (1 + e^(-.423*wdiff - .238*hfa))
wdiff = difference in wins between the two teams (scaled to 16 games)
hfa = 1 if at home, -1 if on road
e = the base of the natural log =~ 2.7181...
This formula is based on a logit regression of all playoff games since 1970. Try it out, it gives answers that are fairly reasonable in the majority of cases.
Of course, it is to Seifert's credit --- at least to some extent --- that his teams were such heavy favorites. That means they were great teams, which is evidence that Seifert is a good coach. But that's not what I'm trying to measure here. What I want to know is: what coaches have done the best job of winning games they weren't supposed to win in the playoffs?
So here's the methodology:
1. look at every playoff game since the merger (1970--2004)
2. make an estimate of each team's probability of winning that game (details below for the
crowd).3. for each coach, add his team's estimated win probability for each game. This gives the number of games he was "supposed to" win.
4. count up the number of games his teams actually won.
5. compare his actual record to his expected record.
For instance, George Seifert's actual record was 10-5. His expected record is also 10-5. So he did no better and no worse than expected. Likewise, Dungy's actual is 5-7 and his expected record is also 5-7. So he too, has been neither a good or a bad playoff coach (at least according to this methodology).
Here is the full list (more fine print follows):
extraCoach Expected Actual wins----------------------------------------Chuck Noll 12-12 16- 8 +3.9Bill Belichick 6- 5 10- 1 +3.9Joe Gibbs 13- 8 16- 5 +3.2Tom Landry 16-15 19-12 +3.1Jimmy Johnson 6- 7 9- 4 +2.8Brian Billick 3- 4 5- 2 +2.2Bill Walsh 8- 6 10- 4 +2.1Bill Parcells 9- 9 11- 7 +2.0Tom Flores 7- 4 8- 3 +1.5Bum Phillips 3- 4 4- 3 +1.4Jerry Burns 2- 4 3- 3 +1.3Raymond Berry 2- 3 3- 2 +1.3Mike Shanahan 6- 5 7- 4 +1.2Jon Gruden 4- 3 5- 2 +1.1Barry Switzer 4- 3 5- 2 +0.9John Madden 7- 7 8- 6 +0.8Dan Reeves 10-10 11- 9 +0.8John Robinson 3- 7 4- 6 +0.7Don McCafferty 3- 2 4- 1 +0.7Herman Edwards 1- 4 2- 3 +0.5Ray Malavasi 3- 3 3- 3 +0.4Jerry Glanville 3- 4 3- 4 +0.4Jeff Fisher 5- 4 5- 4 +0.4Mike Holmgren 9- 8 9- 8 +0.3Marv Levy 11- 8 11- 8 +0.3Tom Coughlin 4- 4 4- 4 +0.3Sam Wyche 3- 2 3- 2 +0.2Andy Reid 7- 5 7- 5 +0.0Dick Nolan 2- 3 2- 3 -0.0Dave Wannstedt 2- 3 2- 3 -0.1Ted Marchibroda 2- 4 2- 4 -0.2George Seifert 10- 5 10- 5 -0.2Tony Dungy 5- 7 5- 7 -0.2Art Shell 2- 3 2- 3 -0.4Red Miller 2- 3 2- 3 -0.5Bobby Ross 4- 4 3- 5 -0.6Steve Mariucci 4- 3 3- 4 -0.6Bud Grant 9- 9 8-10 -0.7Don Coryell 4- 5 3- 6 -0.8Jim Fassel 3- 2 2- 3 -1.0Mike Martz 4- 3 3- 4 -1.0George Allen 3- 4 2- 5 -1.0Dick Vermeil 7- 4 6- 5 -1.0Chuck Knox 8-10 7-11 -1.0Mike Sherman 3- 3 2- 4 -1.2Wayne Fontes 2- 3 1- 4 -1.5Mike Ditka 7- 5 6- 6 -1.5Jack Pardee 3- 3 1- 5 -1.8Dennis Green 6- 6 4- 8 -2.0Don Shula 19-12 17-14 -2.1Bill Cowher 10- 7 8- 9 -2.3Jim Mora 3- 3 0- 6 -3.3Marty Schottenheimer 9- 8 5-12 -3.7Fine print:1. only games played since the merger are counted. So e.g. Don Shula's record in the 1960s is ignored.
2. only coaches with at least 5 playoff games are counted.
3. the "expected records" listed above are rounded to the nearest integer.
4. the formula for estimated win probability is:
WinProb = 1 / (1 + e^(-.423*wdiff - .238*hfa))
wdiff = difference in wins between the two teams (scaled to 16 games)
hfa = 1 if at home, -1 if on road
e = the base of the natural log =~ 2.7181...
This formula is based on a logit regression of all playoff games since 1970. Try it out, it gives answers that are fairly reasonable in the majority of cases.
Just going off W-L record, BB needs to go 6-4 in his next 10 playoff games to match Gibbs' 16-5 record. 6-4 in the postseason isn't easy. Shanahan is 0-3 since his last SB win.
