McDermott ranked 5th best coach in terms of decisions this year

[Big Turk]

Interesting analytics on coaching decisions and a site that takes many factors into account and then ranks the coaches. On average, a coach is responsible for losing 3/4 of a game each season based on their decisions. 

 

Here is the Yahoo article that talks about it...Reid is #1 followed by Reich and LaFleur.

https://www.yahoo.com/sports/best-clutch-coach-nfl-look-kansas-city-170746706--nfl.html

 

And here is the coach rankings site:

https://edjsports.com/nfl/coach-rankings/

 

It puts a huge smile on my face that Marrone is dead last.

 

Also the other AFCE teams are not good...Flores is 20th, Gase 22nd and Belicheat 25th

 

 

[SageAgainstTheMachine]

I feel like the critique I hear most often about Reid is about his in-game decision making.  Either that's untrue or Mahomes' dominance corrects for even that.

[billsfan89]

2 minutes ago, SageAgainstTheMachine said:

I feel like the critique I hear most often about Reid is about his in-game decision making.  Either that's untrue or Mahomes' dominance corrects for even that.

 

That was a big issue in Philly for Reid but I heard he hired a clock specialist to help with late game clock management when he got to KC.

[ctk232]

5 minutes ago, matter2003 said:

Also the other AFCE teams are not good...Flores is 20th, Gase 22nd and Belicheat 25th

 

 

Appreciate the analytics and all, but based on their ranking of Gase alone I have to wonder about the methodology here...

Edited December 19, 2020 by ctk232

[Warcodered]

2 minutes ago, matter2003 said:

Also the other AFCE teams are not good...Flores is 20th, Gase 22nd and Belicheat 25th

😭 Belichick in the bottom 10. I'd thought Flores had been doing well except the weird pulling of Tua. Also how can Gase not be etched in stone at 32 with Anthony Lynn near by?

[SageAgainstTheMachine]

I don't really understand how they can say their data yields that the average head coach loses .75 games each year.  So if there are 32 teams that means the HCs account for a net -24 wins per season as opposed to...what?  Wouldn't it by definition have to be net zero or higher or else the data suggests that the HC position in and of itself is a losing proposition?

[The Wiz]

8 minutes ago, SageAgainstTheMachine said:

I don't really understand how they can say their data yields that the average head coach loses .75 games each year.  So if there are 32 teams that means the HCs account for a net -24 wins per season as opposed to...what?  Wouldn't it by definition have to be net zero or higher or else the data suggests that the HC position in and of itself is a losing proposition?

Only if you go by those stats.  PFF has McD at -103242343424324233469693485 for the season.  They don't have the stats but that's the numbers they gave him.

Edited December 19, 2020 by The Wiz

[Malazan]

What is he ranked in clapping though?

Edited December 19, 2020 by jeremy2020

[Nelius]

Saw a coach of the year ranking that had him at 9th prior to the Steelers game I believe. 9th even before that win seems low. Bills win the division and he's got to be hovering around top 5 with room to grow. Embrace the claps. Love the claps.

 

[GunnerBill]

4 hours ago, SageAgainstTheMachine said:

I don't really understand how they can say their data yields that the average head coach loses .75 games each year.  So if there are 32 teams that means the HCs account for a net -24 wins per season as opposed to...what?  Wouldn't it by definition have to be net zero or higher or else the data suggests that the HC position in and of itself is a losing proposition?

 

Was exactly my thinking.

[Billy Claude]

11 minutes ago, GunnerBill said:

 

Was exactly my thinking.

 

It is at least partially due to the coaches calls on critical plays.  Based on their models of what the better decisions are, coaches tend to make the "wrong" decision more often than the "right" decision on critical plays.  So the average coaches decisions on those plays lead to negative expected points.

 

[GunnerBill]

2 minutes ago, Billy Claude said:

 

It is at least partially due to the coaches calls on critical plays.  Based on their models of what the better decisions are, coaches tend to make the "wrong" decision more often than the "right" decision on critical plays.  So the average coaches decisions on those plays lead to negative expected points.

 

 

So every team would be better without a head coach? I mean what is their statistical baseline? A Head Coach who makes every decision the same as their statistical model? That is a completely worthless baseline. 

[Billy Claude]

10 minutes ago, GunnerBill said:

 

So every team would be better without a head coach? I mean what is their statistical baseline? A Head Coach who makes every decision the same as their statistical model? That is a completely worthless baseline. 

 

I couldn't figure out what their baseline is.   If the baseline is a coach who makes every decision the same as what they feel is their optimal then all coaches would have an overall negative impact -- just that bad coaches would be more negative than good ones.

 

In any case, to a certain extend the baseline doesn't matter -- only how one coach compares to another matters.  Perhaps that is why they only release the rankings.

 

[GunnerBill]

1 minute ago, Billy Claude said:

 

I couldn't figure out what their baseline is.   If the baseline is a coach who makes every decision the same as what they feel is their optimal then all coaches would have an overall negative impact -- just that bad coaches would be more negative than good ones.

 

In any case, to a certain extend the baseline doesn't matter -- only how one coach compares to another matters.  Perhaps that is why they only release the rankings.

 

 

I think the baseline if that is right is utterly nonsensical. But I suppose you are right that as long as it is consistent it doesn't affect the ranking too much.

[Billy Claude]

3 minutes ago, GunnerBill said:

 

I think the baseline if that is right is utterly nonsensical. But I suppose you are right that as long as it is consistent it doesn't affect the ranking too much.

 

I am guessing on the baseline but it would explain why the average is negative.

[Doc Brown]

Whenever I see something like this my first instinct is to look for the flaws in methodology.  Unless the result is something that fortifies my preconceived notions.

 

I think this is a brilliantly conceived model and should be considered the gold standard when it comes to analytic based head coaching decisions.

[Billy Claude]

16 minutes ago, Doc Brown said:

Whenever I see something like this my first instinct is to look for the flaws in methodology.  Unless the result is something that fortifies my preconceived notions.

 

I think this is a brilliantly conceived model and should be considered the gold standard when it comes to analytic based head coaching decisions.

 

 

Sarcasm I assume?    Definitely one has much more confidence in this kind of analysis in baseball or basketball but it is still fun to think about as long as you don't take it too seriously.

 

 

[BobbyC81]

4 hours ago, Nelius said:

Saw a coach of the year ranking that had him at 9th prior to the Steelers game I believe. 9th even before that win seems low. Bills win the division and he's got to be hovering around top 5 with room to grow. Embrace the claps. Love the claps.

 

 

The writers at Sports Illustrated who vote should give him a first place vote since they projected the Bills to win 7 games.

[Big Turk]

8 hours ago, SageAgainstTheMachine said:

I don't really understand how they can say their data yields that the average head coach loses .75 games each year.  So if there are 32 teams that means the HCs account for a net -24 wins per season as opposed to...what?  Wouldn't it by definition have to be net zero or higher or else the data suggests that the HC position in and of itself is a losing proposition?

 

No not necessarily. Just because you make the wrong decision by analytics doesn't mean it won't work in your favor when you do it. If analytics says there is a 56% you should go for it and you punt, it is the wrong decision but that doesn't mean it will lose you a game. It simply means that based on analytics it was the wrong decision.

 

Also, each coach decision is independent of other coach decisions. It's not like if McD makes a good call based on analytics in this instance that the other coach then has to make a bad call to make it balance out.

 

 

Edited December 19, 2020 by matter2003

[SageAgainstTheMachine]

23 minutes ago, matter2003 said:

 

No not necessarily. Just because you make the wrong decision by analytics doesn't mean it won't work in your favor when you do it. If analytics says there is a 56% you should go for it and you punt, it is the wrong decision but that doesn't mean it will lose you a game. It simply means that based on analytics it was the wrong decision.

 

Also, each coach decision is independent of other coach decisions. It's not like if McD makes a good call based on analytics in this instance that the other coach then has to make a bad call to make it balance out.

 

 

 

I see.  I didn't delve deeply enough into this at first to see that they were exclusively talking about 4th down decisions.  I thought they were saying that the net total of all decisions by all coaches was an average of 3/4 of a loss per season per coach. 

 

Still though, I think it makes more sense to frame the stat in terms of expected points rather than expected wins/losses.  This is one of those rare statistical moments where I think it's inappropriate to consider each event independently of the opposing team.  For example I didn't mind McDermott trotting out Bass 6 times against the Jets when we beat them 18-10 but had he done the same against a high powered offense it would have been way too conservative.