How Good Does Mac Have To Be To Get His Option Picked Up?

[Kasmir]

It doesn’t change a damn thing about the sport.

Points scored/allowed isn’t an antiquated stat, neither is turnovers… the very notion is absurd.

Do you even realize that something as apparently simple as the relationship between points scored/allowed and wins is not at all simple? Bill James came up with the Pythagorean Winning Percentage Expectation model for baseball win predictions and Daryl Morey first mentioned the Football version of the theorem. I get that you could give a crap but a lot smart folks have spent a lot of time on this stuff. It is certainly controversial but it is hardly absurd.

Pythagorean expectation - Wikipedia

en.wikipedia.org

 

[Wozzy]

Do you even realize that something as apparently simple as the relationship between points scored/allowed and wins is not at all simple? Bill James came up with the Pythagorean Winning Percentage Expectation model for baseball win predictions and Daryl Morey first mentioned the Football version of the theorem. I get that you could give a crap but a lot smart folks have spent a lot of time on this stuff. It is certainly controversial but it is hardly absurd.

Make a bet with your local bookie, watch your team lose... tell him you don't plan on paying because analytics matter more than points... see if he doesn't break your legs.

Football is pretty simple. You don't need a protractor to figure out who won or lost... or why.

 

[rochrist]

I think you believe objectivity is good and subjectivity is bad. I don't believe anything that simple. PFF is a subjective system and DVOA is an objective model. That's a truthful characterization, but says *nothing* about whether they are useful, let alone whether there results are true or false.

There really isn't a whole lot of distance between PFF 'analysts' grading plays, and Football Outsiders 'analysts' making up a weighting factor so that their algorithm can grade the plays for them. DVOA -could- potentially be better, but it would take metric-buttload of statistical analysis to get there and I don't think they're remotely close to that.

 

[Ring 6]

Because it is crap.

Game winning drives isn't open to interpretation... that's why you look silly for saying Brees wasn't "clutch" enough.

Of course game winning drives is open to interpretation. How about interpreting that he is only near the top because he played so many games. Clutch isn’t measured by trying more and succeeding less often.

How can it be crap if you admit you don’t even know what it is? You are clowning yourself.
And for the record I never said Brees was or wasnt “clutch”.
I said how a qb plays when the game is on the line is more important than cumulative stats. You chose Brees, so apparently you think his play when the game was on the line wasn’t so good. I didn’t say that, you did.

 

[Wozzy]

Of course game winning drives is open to interpretation. How about interpreting that he is only near the top because he played so many games. Clutch isn’t measured by trying more and succeeding less often.

How can it be crap if you admit you don’t even know what it is? You are clowning yourself.
And for the record I never said Brees was or wasnt “clutch”.
I said how a qb plays when the game is on the line is more important than cumulative stats. You chose Brees, so apparently you think his play when the game was on the line wasn’t so good. I didn’t say that, you did.

Game winning drives are not open to interpretation, you start the final drive with a deficit, lead your team down the field and win the game. Brees was third all-time... in history... but in your mind not clutch enough... absurd.

You also argued Brees 32nd ranked defense shouldn't have been a detriment to him winning... ridiculous. In the real world defense matters.

 

[Ring 6]

Points, turnovers, 3rd down and goal line stats are directly related to winning and losing.

Analytics aren’t related to anything, they’re made up nonsense that place value based upon whomever is creating the fake stat.

So do you think all stars are created equal?
If QBA throws for 280 yards and 2 TDs while building his team a 35-0 lead and QBB throws for 400 and 4 TDs against a prevent defense to lose 38-31 so you compare those stats apples to apples?

Or if both QBs have identical stats for the game overall but one has a 3 and out in the 4th qtr and the other leads his team to a game winning score, was their play identical?

Your arguments so far would indicate you believe QBB is better in the first example and they both played the same in the second one.

 

[Ring 6]

Game winning drives are not open to interpretation, you start the final drive with a deficit, lead your team down the field and win the game. Brees was third all-time... in history... but in your mind not clutch enough... absurd.

You also argued Brees 32nd ranked defense shouldn't have been a detriment to him winning... ridiculous. In the real world defense matters.

If wins aren’t a qb stat why are comebacks? Shouldn’t that be a team stat?

Again, I never said Brees wasn’t clutch (there you go again with the lying)
but if game wining drives are the measure of clutch isn’t Derek carr more clutch because he did it in a much higher % of his games.
See the problem with cumulative stats. Is Brees the 12th worst at throwing Ints? The second biggest loser?

Again, I never said his defense wasn’t a detriment.
You can’t seriously not realize how much you are clowning yourself by changing my argument so you have something to argue.
You literally are admitting you are wrong and lying about my comments and arguments. You must be very embarrassed

 

[Wozzy]

So do you think all stars are created equal?

Where did I even hint at that… I didn’t. In fact earlier in this conversation I relayed a BB quote that I happen to agree with.

Points is the most important stat in football, turnovers is second because it most directly correlates to points.

Don’t lie about what I said, everything is time stamped in earlier posts.

 

[Ring 6]

Do you even realize that something as apparently simple as the relationship between points scored/allowed and wins is not at all simple? Bill James came up with the Pythagorean Winning Percentage Expectation model for baseball win predictions and Daryl Morey first mentioned the Football version of the theorem. I get that you could give a crap but a lot smart folks have spent a lot of time on this stuff. It is certainly controversial but it is hardly absurd.

Pythagorean expectation - Wikipedia

en.wikipedia.org

This is most likely the stupidest statistic ever invented.
And it doesn’t say what you think it does.
If a team wins 11-0 the loses 5-2 3 straight games it is 1-3 but scored 16 runs and allowed 15. This formula would make it seem they are a better than average team, and they are not. Games aren’t played based upon averages, they are played one time, winner takes all.

For all the reasons that cumulative stats over the course of a season or career are not apples to apples between players, this stat is an abomination of reality.
But it’s instructive for fantasy sports.

 

[Wozzy]

If wins aren’t a qb stat why are comebacks? Shouldn’t that be a team stat?

I didn’t create or track the stat, PFR did… I just used it to show how dumb your argument was. Brees is second on the all-time passing yards and passing TD’s list… the notion that he wasn’t “good enough” in and of itself is dumb.

Tom Brady passed for 500 yards and 3 TD’s in the 2017 Super Bowl and lost. Was he not clutch enough… just dumb.

Again, I never said Brees wasn’t clutch

Again, I never said his defense wasn’t a detriment.

Yes you did, it started on page ten and got worse the longer we engaged. Now you want to walk it back because stats made your argument look silly.

 

[Kasmir]

There really isn't a whole lot of distance between PFF 'analysts' grading plays, and Football Outsiders 'analysts' making up a weighting factor so that their algorithm can grade the plays for them. DVOA -could- potentially be better, but it would take metric-buttload of statistical analysis to get there and I don't think they're remotely close to that.

PFF manually grades players not plays so that's a ton of work. FO grades plays algorithmically from box score data. That's why the FO DVOA etc database is able to go back to 1981 (and counting) which could never happen with PFF's labor intensive system (although maybe someday AI will score film).

That's a whole lot of distance, like night and day difference in the approaches.

Again, I've said *nothing* about better or worse as that's an apples and oranges orthogonality.

I think the PFF approach would be very useful if it worked, as it actually produces a ton of data versus the FO approach which is purely derived from existing box score data. That perhaps explains why PFF has NFL team customers while to the best of my knowledge FO does not.

I'm tired of shilling for FO/DVOA; I know you've read this, but for the benefit of others here is FO's explanation of their system and its unique approach:

What is DVOA? | Football Outsiders

An explanation of Football Outsiders' advanced statistics.

www.footballoutsiders.com

 

[Wozzy]

There really isn't a whole lot of distance between PFF 'analysts' grading plays, and Football Outsiders 'analysts' making up a weighting factor so that their algorithm can grade the plays for them. DVOA -could- potentially be better, but it would take metric-buttload of statistical analysis to get there and I don't think they're remotely close to that.

This ^

It’s creating stats to fit preexisting bias. That’s why it’s perpetually changing, results make them realize every season or two it’s flawed nonsense.

 

[RodThePat]

DVOA is a load of crap... so is offensive EPA... whatever that is.

Wow, great argument. But since you love context free "points scored/allowed" metrics, all of the past 6 Super Bowl winners have ranked higher in points per game than they did in points against. The argument remains the same. We're in an offensive era and QBs drive offense. The most certain way to contend is through great QB play. Great QB play is way more likely to carry a below average defense than a great defense is to carry a below average QB. All of those adages are true in 2023.

 

[Kasmir]

This ^

It’s creating stats to fit preexisting bias. That’s why it’s perpetually changing, results make them realize every season or two it’s flawed nonsense.

Nonsense. What preexisting bias are you alleging? The ongoing work they do is to try to improve the fit of their model to actual win/loss results, i.e. for it to backcast results better so it can be more predictive of future wins/losses. That's an immensely difficult problem. An advantage of quantification over qualitative hand waving is that you can actually test your ideas.

 

[Ring 6]

Where did I even hint at that… I didn’t. In fact earlier in this conversation I relayed a BB quote that I happen to agree with.

Points is the most important stat in football, turnovers is second because it most directly correlates to points.

Don’t lie about what I said, everything is time stamped in earlier posts.

You are implying that all stats are created equal by arguing that cumulative career statistics are more important than making the plays that result in winning or losing.

There is no question that belichicks comment is 100% right, but if you really believe it why are you judging okay by stats instead of winning?

 

[Kasmir]

This is most likely the stupidest statistic ever invented.
And it doesn’t say what you think it does.
If a team wins 11-0 the loses 5-2 3 straight games it is 1-3 but scored 16 runs and allowed 15. This formula would make it seem they are a better than average team, and they are not. Games aren’t played based upon averages, they are played one time, winner takes all.

For all the reasons that cumulative stats over the course of a season or career are not apples to apples between players, this stat is an abomination of reality.
But it’s instructive for fantasy sports.

I'm not advocating for a Pythagorean approach because it indeed a simplistic model. But at least it's an attempt to quantify how the handwaving claim that points scored/allowed matter relates to actual wins and losses.

American Football is so popular that even smart analytic nerds follow it and spend time thinking about it. Even statisticians and scientists can be football fans. And with $12 billion bet annually on NFL games there's a lot of money riding on better predicting wins, losses and their derivatives.

 

[Ring 6]

I didn’t create or track the stat, PFR did… I just used it to show how dumb your argument was. Brees is second on the all-time passing yards and passing TD’s list… the notion that he wasn’t “good enough” in and of itself is dumb.

Tom Brady passed for 500 yards and 3 TD’s in the 2017 Super Bowl and lost. Was he not clutch enough… just dumb.

Yes you did, it started on page ten and got worse the longer we engaged. Now you want to walk it back because stats made your argument look silly.

You said wins are not a QB stat. Then you brought in career comeback wins as an argument for Brees’ greatness. Why would comeback wins be a qb stat?
The funniest thing is I never argued Brees wasn’t great.
I said, again,
The value of a qb is in making plays that decide games, not building up a bunch of stats.
Somehow you took that to mean you had to defend drew Brees.
By the way Brees is 4th in all time wins, so I’m not sure why you chose him to defend your argument that accumulating stats is more important for a qb than winning.
Again, my viewpoint is that the reason you play the game is to win, so anything that isn’t contributing to winning doesnt gave much value to me.
You seem to think the purpose of the game is to show off pretty stats.

 

[Ring 6]

I'm not advocating for a Pythagorean approach because it indeed a simplistic model. But at least it's an attempt to quantify how the handwaving claim that points scored/allowed matter relates to actual wins and losses.

American Football is so popular that even smart analytic nerds follow it and spend time thinking about it. Even statisticians and scientists can be football fans. And with $12 billion bet annually on NFL games there's a lot of money riding on predicting wins, losses and their derivatives.

Points scored and allowed really are all that matters because whoever scored more points wins the game.

What these analytics try to do is take all of the hundreds of variables and try to see which ones correlate (not causation) to winning.
Football games are not played on a computer. The ability of one player or team to score points or stop points from being scored is an actual ability. The guy who completes the highest percentage of passes is not the same as the guy who is most likely to complete the game winning pass in a critical situation.
The defense that gives up the fewest yards is not the same as the defense that it is toughest to get into the end zone on when it’s do or die.

If analytics really were meaningful predictors then betting would be easy.
But analytics will never be meaningful predictors when you try to accumulate events from different games, on different days, against different opponents, with different game situations, fluke plays,, momentum, injuries, etc.

The better team is the team that wins not the one that accumulated stats that result in a formula saying they should have.

 

[Kasmir]

Points scored and allowed really are all that matters because whoever scored more points wins the game.

What these analytics try to do is take all of the hundreds of variables and try to see which ones correlate (not causation) to winning.
Football games are not played on a computer. The ability of one player or team to score points or stop points from being scored is an actual ability. The guy who completes the highest percentage of passes is not the same as the guy who is most likely to complete the game winning pass in a critical situation.
The defense that gives up the fewest yards is not the same as the defense that it is toughest to get into the end zone on when it’s do or die.

If analytics really were meaningful predictors then betting would be easy.
But analytics will never be meaningful predictors when you try to accumulate events from different games, on different days, against different opponents, with different game situations, fluke plays,, momentum, injuries, etc.

The better team is the team that wins not the one that accumulated stats that result in a formula saying they should have.

We certainly agree that whomever scores the most points wins the game. But point differential is not a very good predictor. Hence all the analytic efforts.

 

[Ring 6]

We certainly agree that whomever scores the most points wins the game. But point differential is not a very good predictor. Hence all the analytic efforts.

Predictor of what?