Comments on Advanced NFL Stats: Bill Walsh on Randomizing

Anon, that would be interesting to have in the NFL...

2010-06-09T01:46:37.495-04:00

Anon, that would be interesting to have in the NFL. 15 games in conference, 3 inter-conference, 18 total. Best team in each conference goes to the Super Bowl. That would be fun.

I've really never had a problem with the best team not winning, so long as it's a surprise and an uncommon occurence when the underdogs do win. To me, it's just a part of sports. The best team doesn't always win. It could be because of lucky breaks, injuries, good team has a few bad days, etc. What really bugs me is when long regular seasons essentially determine nothing, like in the NBA and the NHL. Yet I still kinda have to follow it because the regular season is quite predictive of who will be the best in the playoffs.

I do find myself gravitating towards table systems. That's probably one of the reasons why I watch F1 racing so much, and why I stopped following NASCAR after they instituted "The Chase."

Thanks to your awesome work Brian, in the past few...

2010-06-07T02:46:35.241-04:00

Thanks to your awesome work Brian, in the past few months I have come to understand the incredible importance of the field of Game Theory in analyzing the NFL. A huge part of football strategy is the "indifference principle". And non-random behavior in the play calling is an arbitrage opportunity for the opposing team. It's really cool to see that Bill Walsh was on to these intricate principles long ago. Also, thanks for posting that awesome resource of playbooks. Terrific content as usual.

Oops. Sorry, I had the wrong link to the Walsh com...

2010-06-04T17:22:34.237-04:00

Oops. Sorry, I had the wrong link to the Walsh compendium. It was pointing to a Walsh 49ers playbook. Fixed.

Team goes 14-2, gets the 1 seed, then loses in ...

2010-06-03T08:32:28.585-04:00
Team goes 14-2, gets the 1 seed, then loses in the divisional round. They obviously choked, right? Great regular season team but not a playoff team, eh? Perhaps, but it's more likely that one of their losses just happened to fall on week 19, out of sheer randomness.

And this is why I dislike playoff systems. They often don't determine the best team. I greatly prefer a table system whereby every game counts an equal amount.

I just had a very quick check on the 2008 play-by-...

2010-06-02T18:47:53.283-04:00
I just had a very quick check on the 2008 play-by-play data (thanks for sharing it) and by my reckoning on 2nd and 10, teams call runs 44% of the time when the previous play was an incomplete pass and 33% of the time when the previous play was not an incomplete pass.

Based on your Sep 2008 article you link to, 33% would be about right to smooth out the aberration at 2nd and 10. Definitely a case of alternating play-calling over true random.

There's not a huge difference in average gain between the two (4.54 yards after an incomplete vs 4.66 otherwise) though, which suggests that perhaps defensive coordinators haven't picked up on this trend either, or they have but haven't found a way to capitalise on it yet.

Wasn't Nolan already in SF by 2005?

2010-06-02T11:26:33.043-04:00
Wasn't Nolan already in SF by 2005?

People are horrible at randomizing because they ar...

2010-06-01T18:44:46.287-04:00
People are horrible at randomizing because they are horrible at figuring out what is truly random, basically.

Team goes 14-2, gets the 1 seed, then loses in the divisional round. They obviously choked, right? Great regular season team but not a playoff team, eh? Perhaps, but it's more likely that one of their losses just happened to fall on week 19, out of sheer randomness.