CFZ Analytics in Football: What it is and isn’t

[Bobhaze]

I was part of the analytics team for a company called Luxottica, owners of Lenscrafters, Target optical, Pearl Vision, Eyemed vision plan, etc.

Here is what I can tell you about analytics.:

We are influenced by what we see. However, what we see is a matter of perception and it differs from person to person. Analytics gives equal value to what we see, what we think we saw, and what we didn't notice.

For example, how many specific plays do you remember from the last 49er playoff game? Most likely you only remember the plays that make the highlight reel. You remember scoring and turnovers. You might remember a few sacks but that is just about all you remember, However, what about the plays that put you in a position to score? What about the plays leading up to the desperation play that resulted in a turnover?

This is why it is so humorous when a poster says, "I know what I saw"........no, you don't know what you saw. You might know 10% of what you saw and you are basing your opinion on that limited memory.

Analytics identify opponent tendencies, what a team is likely to do under certain circumstances.

Very well said Plastic man!

 

[Bobhaze]

Partly the reason Riley jumped ship from Oklahoma; he knows the Sooners will be further exposed as an inadequate defensive team next year in the SEC.

Next year? They haven’t had a great defense in years! I think both OU and Texas will be similar to A&M in the SEC- mediocre. OU hasn’t won the Big 12 the last 2 years and Texas hasn’t even been that great in the big 12- KState, TCU and OK State have all been better than UT the last decade.

 

[Motorola]

Next year? They haven’t had a great defense in years! I think both OU and Texas will be similar to A&M in the SEC- mediocre.

'Bobhaze' - I noted 2024 because that will be the inaugural season both OU & UT will become SEC member schools competing in football (and other team \ individual sports).

 

[conner01]

My concern is will a coach, or do they, go with the analytics in a certain situation over the gut feeling.

Say 4th and 1, at the 38, and the analytics say go for it against a particular defense, what are the odds of making it. As opposed to ... we have a 260 RB, give it to him ... or kick the FG, we have a good K. I know the score and situation comes into account as well.

Or does it come down to that or not. Is someone in the coaches ear feeding him the analytics. Or do they have all of that before hand, and possibly in the game plan.

Still a lot to take into account and to learn as well.

I think the HC will still go with gut. Getting the data quick enough to make a decision is gonna be an issue. But analytics don’t factor in everything about that decision. Say you are playing a weak offense, 3 points becomes more valuable than if you are playing a good offense and the 7 points means so much more.
Those are the decisions a HC get judged on. Sometimes judged wrong because fans judge the result, not the decision more often than not.

 

[FuzzyLumpkins]

My concern is will a coach, or do they, go with the analytics in a certain situation over the gut feeling.

Say 4th and 1, at the 38, and the analytics say go for it against a particular defense, what are the odds of making it. As opposed to ... we have a 260 RB, give it to him ... or kick the FG, we have a good K. I know the score and situation comes into account as well.

Or does it come down to that or not. Is someone in the coaches ear feeding him the analytics. Or do they have all of that before hand, and possibly in the game plan.

Still a lot to take into account and to learn as well.

You would simply need to adjust the risk/reward assessment to include those specific circumstances. the success of a specific RB or K in those situations can be quantified and weighed as well. Then you can have an informed choice.

 

[JD_KaPow]

They post the numbers without any disclaimers that "this superstar wasn't playing" or "the teams they have played thus far seem to be crap so the numbers may not hold up this week against the Chiefs", etc..

I always hear the numbers thrown about as factual. They may be based on the facts or at least mostly fact of what happened but there is no context. I mean the Ravens may have given up a run on 4th & 3 last week but if everyone watching saw that the OL got away with blatant hold it doesn't change it's in the numbers that the Ravens didn't do well in that situation. I'm not going for a run on 4th & 3 knowing I saw the hold and that Ray is back regardless if the numbers say I have a huge chance of converting.

You can only build so much context in mathematically ..... even that is based on someone's interpretation.

Who's "they?" I don't really care what the game announcers or ESPN personalities say. Those people say all sorts of nonsensical things.

 

[JD_KaPow]

This is generally a good description.
The keynote is ALWAYS missing though.

** IT IS BASED ON PREVIOUS DATA and does not constitute real-time actualization through opponents ALSO having access to that previously compiled data.


Real world example:

4th and 3 from the 30-yard line. --48 yard FG or go for it or punt.
Analytics. Go for it.

Defenses: Spend a tremendous amount of time preparing for 4th down in that area of the field with yardages between 1 and 5 to gain.

analytics 2 years later: DO NOT GO FOR IT YOU GOOFS.

I'd pay a lot more attention to the "context matters" folks if they EVER came down on the more aggressive side. But no, they ALWAYS use these arguments to say why teams should be more conservative than the analytics suggest.

Yes, it's absolutely true that context matters. Field conditions, who's healthy and playing, relative strength of offenses and defenses, home field, etc. But those contextual things cut both ways. Since the analytics are based on some average state (although that state can have lots of contextual detail built into it), it must be true that sometimes you should be more conservative than the models say and sometimes you should be more aggressive, based on factors that aren't included in the models. And yet, I have never once seen the people who make these kinds of arguments say that a team should be more aggressive than the analytics suggest. It always, ALWAYS goes the other way, toward the traditional, incredibly risk-averse "book."

You know, offenses can "spend a tremendous amount of time preparing for 4th down" as well.

The 4th down stuff isn't in the margins. Back in the day, teams were risk-averse to an unbelievably high degree and their conservatism was massively harming their teams chances to win. Of course, it was mitigated somewhat by the fact that the other team was also making incredibly stupid decisions. It's finally been changing, and it's going to change more.

 

[jterrell]

I'd pay a lot more attention to the "context matters" folks if they EVER came down on the more aggressive side. But no, they ALWAYS use these arguments to say why teams should be more conservative than the analytics suggest.

Yes, it's absolutely true that context matters. Field conditions, who's healthy and playing, relative strength of offenses and defenses, home field, etc. But those contextual things cut both ways. Since the analytics are based on some average state (although that state can have lots of contextual detail built into it), it must be true that sometimes you should be more conservative than the models say and sometimes you should be more aggressive, based on factors that aren't included in the models. And yet, I have never once seen the people who make these kinds of arguments say that a team should be more aggressive than the analytics suggest. It always, ALWAYS goes the other way, toward the traditional, incredibly risk-averse "book."

You know, offenses can "spend a tremendous amount of time preparing for 4th down" as well.

The 4th down stuff isn't in the margins. Back in the day, teams were risk-averse to an unbelievably high degree and their conservatism was massively harming their teams chances to win. Of course, it was mitigated somewhat by the fact that the other team was also making incredibly stupid decisions. It's finally been changing, and it's going to change more.

Missing forest for the trees. Typical of the board though.

If the analytics suggest you be more conservative, then the other side would adjust to that as well making being aggressive more appealing.

Its not that hard to simply reverse the above.
If the defenses spend a lot of time practicing short yardage 30-yard defense and you are told to kick... they know this and will eventually stop practicing it... then it would make sense ot once again go for it.

Nothing wrong with analytics but you should know they are based on prior data and can be countered by the other team also having the data.

If you just want to be aggressive all the time or conservative all the time you don't need to bother with data at all.

 

[CCBoy]

Missing forest for the trees. Typical of the board though.

If the analytics suggest you be more conservative, then the other side would adjust to that as well making being aggressive more appealing.

Its not that hard to simply reverse the above.
If the defenses spend a lot of time practicing short yardage 30-yard defense and you are told to kick... they know this and will eventually stop practicing it... then it would make sense ot once again go for it.

Nothing wrong with analytics but you should know they are based on prior data and can be countered by the other team also having the data.

If you just want to be aggressive all the time or conservative all the time you don't need to bother with data at all.

No need to have to explain facts of training...repetition/rehearsal is always needed, even deployed in a zone of conflict.

 

[AmericasTeam81]

Just wait till AI takes over the NFL.

 

[CCBoy]

I was part of the analytics team for a company called Luxottica, owners of Lenscrafters, Target optical, Pearl Vision, Eyemed vision plan, etc.

Here is what I can tell you about analytics.:

We are influenced by what we see. However, what we see is a matter of perception and it differs from person to person. Analytics gives equal value to what we see, what we think we saw, and what we didn't notice.

For example, how many specific plays do you remember from the last 49er playoff game? Most likely you only remember the plays that make the highlight reel. You remember scoring and turnovers. You might remember a few sacks but that is just about all you remember, However, what about the plays that put you in a position to score? What about the plays leading up to the desperation play that resulted in a turnover?

This is why it is so humorous when a poster says, "I know what I saw"........no, you don't know what you saw. You might know 10% of what you saw and you are basing your opinion on that limited memory.

Analytics identify opponent tendencies, what a team is likely to do under certain circumstances.

How many times did you file for over $100,000 on a tax return?

 

[rnr_honeybadger]

Get ChatGPT to calls plays

 

[plasticman]

How many times did you file for over $100,000 on a tax return?

I wasn't a lead analyst or even close. I was a former store manager with a B.S. degree in math and a fair amount of experience with data bases.

It was a combination of purchase tendencies and loss prevention. My job was to program their searches in either Excel or Matlab. I was a team member but I didn't run Jack.

I came close to a hundred Grande .... once.

 

[CCBoy]

I wasn't a lead analyst or even close. I was a former store manager with a B.S. degree in math and a fair amount of experience with data bases.

It was a combination of purchase tendencies and loss prevention. My job was to program their searches in either Excel or Matlab. I was a team member but I didn't run Jack.

I came close to a hundred Grande .... once.

I was just curious...not negative delving at all. I filed twice in my life for that amount. My returns were about $5,000-$6,000. When I departed the Army and when I departed the Post Office. 1st was in lieu of regular retirement, my wife had cancer and was afraid she wouldn't make if I didn't stay with her when I came on levy.
I went into the Reserves for 3 years. Then, when I retired from Postal employment, I cashed in my personal investment funds, not my postal retirement funds.

 

[CCBoy]

In the spring of 1991, in Dallas Cowboys headquarters, a petroleum engineer named Mike McCoy sat down at his desk to devise one of the most influential innovations in the past three decades of the NFL. He plotted on logarithmic paper every trade involving a draft pick over the previous four years. He wanted to quantify the value of each pick.

Over the previous two years, the franchise’s brash leaders, owner Jerry Jones and Coach Jimmy Johnson, had built a war chest for the draft. They had shipped star running back Herschel Walker to Minnesota for a massive haul, including eight picks, and quarterback Steve Walsh to New Orleans for three more. Now, Jones’s oil-and-gas business partner was trying to figure out how much the picks were worth.

In his office, McCoy created a graph, assigned the No. 1 pick a random value (3,000) and used a regression formula to calculate the rest. The result — the famous but misunderstood and misattributed “Jimmy Johnson draft-pick value chart” — revolutionized the NFL draft.

https://www.washingtonpost.com/sports/2022/04/25/nfl-trade-value-chart/



Truth be known now! Not just another credit to only Jimmy Johnson...statistical analysis.

 

[CCBoy]

But some team executives and analysts believe teams are slowly beginning to shift away from McCoy’s chart. They see it as part of the NFL’s larger analytics movement, which has manifested on the field in passing rates and fourth-down aggression and in front offices with the building out of data departments.

https://www.washingtonpost.com/sports/2022/04/25/nfl-trade-value-chart/

On draft night, as other GMs, coaches, executives and scouts conspire in war rooms across the country, this GM evaluates real-time offers by having an analytics staffer plug them into a model on his computer. He said he considers the charts as gauges, not gospel, because he must weigh so many factors, including schemes, medical histories, interviews and dozens of opinions. The GM said his most valuable resource is the team’s draft board, effectively its internal draft curve.



Even with Jerry Jones...

 

[noshame]

Discipline and execution

 

[CowboyoWales]

Statistics give an indication of liklihood, its never an absolute.

 

[Jumbo075]

Jazzy, the bottom line is, coaches will still make ALL decisions. This just gives them better information to make those decisions. A “gut feeling” can still be made of course. But I would hope it would be an “informed” decision.

Take Barry Switzer’s “gut feeling” about 4th and 1 on the Cowboys own 29 yard line at Philly in ’94. We all remember that that play got stuffed and the eagles turned that into points and a huge momentum swing. IF the analytics showed going for it on 4th and 1 from your own 29 had a high chance of success, that gut feeling was ok even if it didn’t work. On the other hand, if the data shows it has about a 20% chance of success, is that a good “gut feeling” or not?

Also- the analytics info is all given to coaches ahead of time. The analytics people aren’t screaming in the coach’s ear from the press box during the game.

I doesn't nede no ejumicashun. coaches jus nede to yuz there xperence, and not analytics. Futbal aint math!

 

[StarOfGlory]

Jazzy, the bottom line is, coaches will still make ALL decisions. This just gives them better information to make those decisions. A “gut feeling” can still be made of course. But I would hope it would be an “informed” decision.

Take Barry Switzer’s “gut feeling” about 4th and 1 on the Cowboys own 29 yard line at Philly in ’94. We all remember that that play got stuffed and the eagles turned that into points and a huge momentum swing. IF the analytics showed going for it on 4th and 1 from your own 29 had a high chance of success, that gut feeling was ok even if it didn’t work. On the other hand, if the data shows it has about a 20% chance of success, is that a good “gut feeling” or not?

Also- the analytics info is all given to coaches ahead of time. The analytics people aren’t screaming in the coach’s ear from the press box during the game.

That game was 1995, week 15. When I looked at the O-lines and D-lines in that game, you'd say that 80% of the time we would have picked up the first down. The Eagles had two players on the D-line that I don't even remember, and I lived up there then. Then again, if we punted, we put the ball in Rodney Peete's hands to win against the likes of Deion and Woodson. I like those odds too.

There is a 15 minute video of this game on YouTube. Bobby Taylor, drafted specifically because we had Irvin, has two nice pass defenses against him. At the end of the game, when Troy is leading the team back, William Fuller gets to Troy. Troy gets up and is screaming at Erik Williams, pissed off as all hell.

But we kicked the snot out of the Eagles in the playoffs and won the Super Bowl that year!