irish eyes
Footballguy
Well, it's Todd Bowles. No business being a head coach at any level.
Fantasy Football - Footballguys Forums
This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!
Post here when coaches do something you disagree with (2 Viewers)
- Thread starter GreenNGold
- Start date
Deamon
Footballguy
He had fruit.
ignatiusjreilly
Footballguy
Bagel would have increased his WP by 4.3%
ignatiusjreilly
Footballguy
Why is that odd? If you run a calculation, of course it's going to come up with a specific number.
As far as "pseudo science", that assumes that people are using it (non-pseudo) scientifically. But that's not what happening. I think most people, from coaches to casual observers, recognize that calculations like that are an input, which should be balanced against other inputs. When I see that the Raiders had an 84% WP, that tells me they were likely (but not guaranteed) to win the game, which is the same thing any casual fan who was watching the game but had no idea of the probabilities would think.
Runkle
Footballguy
Didn't want to spoil it by naming names, but, one guess which coach called this TO down 3 at the end of the 3rd quarter. But it was actually 2nd and 10 after re-checking the game log.
ignatiusjreilly
Footballguy
Reminds me of the time Aqib Talib took advantage of the half-the-distance rule to prevent a TD at no cost (other than to his wallet)
Then Schlereth is an idiot
Chaka
Footballguy
They call him Stink for a reason.
But he's not entirely wrong about analytics. It provides highly questionable numbers that people who don't truly understand them take as gospel truth.
Leroy Hoard
Footballguy
True but I still remember a time when NFL coaches would punt from the other teams 40 just to gain 20 yards in field position. The truth lies somewhere in the middle.
Chaka
Footballguy
That's the problem with analytics, they provide exact percentages for all scenarios as if they could possibly determine with any degree of confidence that a team has a 76% chance of converting a 3 & 3 with a pass and only 58% if they run. Those are nonsense stats. Saying they have a better chance passing vs running is far more intellectually honest (even if that statement may also be untrue).
BobbyLayne
Footballguy
I’m afraid I’m not following.
Analytics does not determine anything.
It presents data to take it into consideration.
They don’t pull the % out of thin air. It’s the result of compiled data for similar situations
Chaka
Footballguy
I am aware of what analytics actually does and what it is presented as, and those are two entirely different things.
Probably a topic for a dedicated thread though.
BobbyLayne
Footballguy
I’m not confident you are based on several posts upthread
but that’s fine we’ll take it up another day
Runkle
Footballguy
Or when they're trying to run out the clock at the end of the game and take an intentional safety to kill an extra dozen seconds without having to risk a kick return or turnover. Usually tell the linemen to hold like crazy to make sure the QB has lots of time to dance in the end zone to use as much clock as possible. Holding doesn't result in a whistle, the play continues. Penalty is... a safety, which they want anyway. So go ahead an mug those guys!
They also provide data like a 75% win probability if you kick a FG vs. 71% if you go for it on 4th and fail vs. 93% if if go for it and convert. Yet the anti-analytics folks yell and scream like mad when the coach goes for it....as if those numbers didn't ALREADY take into account shifts in momentum etc.
MANY (most?) of those who rail against using data like this are simply anti-intellectuals who don't like math and tend put way too much faith in "gut feeling" and tradition. Again...he's an idiot
ignatiusjreilly
Footballguy
GB, you're one of my favorite posters here, but you're really straw-manning with this stuff. Analytics don't "predict", they tell you what has happened historically in similar situations. Also, I don't recall ever seeing any kind of analytically based analysis arguing that a team should run or pass in a specific situation. How could you even make that argument, considering that there are a million different types of runs vs. passes?
I'm sure there are people out there who look at a chart that says going for it on 4th down increases your win probability by 0.1% and automatically think that means a team should go for it, but I think most people understand these historical averages are suggestive, not definitive; they all have margins of error, and each individual situation has factors that may suggest that situation is materially different from the historical average.
Also, to repeat something I mentioned a few pages back, we know for a fact that NFL teams don't make decisions that way, because they are forbidden from doing so. Coaching staffs can only communicate with each other, not the Internet, so there's no way they can have Nerdlinger up in the booth running the numbers for each specific situation.
Chaka
Footballguy
I appreciate that Ignatius, sincerely. But that's the problem with the data, what a conglomeration of other road teams did on 3rd & 3 from the opponents 43 yard line on the left hash mark, playing on Field Turf in a dome has almost zero predictive value for what a specific road team is doing on 3rd & 3 from the left hash mark on the opponents 43, playing on Field Turf in a dome.
The number of variables that exist between all of those specific scenarios combined and a single specific scenario happening in real time are legion. Lumping all of them together and spitting out a number is a fun exercise but it has little value beyond maybe providing a potential yes/no data point that helps you justify a decision to the media.
And I have no problem with using that data, I dig deep every week trying to make dozens of WDIS decisions, but I fully recognize, and openly acknowledge that I am mostly making W.A.G.s
But too many people who don't truly understand how statistics works, and I'm referring to the announcers and talking heads not posters in here, look at the output stat and think "Oh, they had a 74% chance to accomplish X if they had just done Y, they're stupid for not having done Y!!!" And that is simply not the case.
The "data in" is still low quality relative to any specific situation it is being compared to so the "statistic out" has a very low degree of confidence. And that is before you even consider the algorithms that are used to parse this data.
PhD in Human Genetics here, and even though I am far closer to Peter Venkman as a scientist than, say Milton Friedman, but I like to think I have a decent understanding of higher level statistics.
There is a future for analytics and I am absolutely on board with them in the future and now. But ATM the general football viewing audience gives them far too much weight.
Last edited:
rockaction
Footballguy
I've been saying this for years, but I think you've said it better than I ever have.
That said, moving the needle on some very basic stuff (like punting while you're fourth-and-short on the opposition's 43) was a huge boon to the NFL, and the NFL still hasn't listened to very basic stuff like that.
But I fully agree with your claim that you have to look at the inputs and how they're manipulated into a correlative value before you judge an analytic statistic.
Chaka
Footballguy
To be clear I don't thing people like @ignatiusjreilly & @renesauz and I are ideologically far apart.
I do not discount analytics at all, my beef is primarily with how they are used from a studio pundit and broadcast perspective. I think they use the data irresponsibly. And I don't think you guys take them as gospel truth.
I do not discount analytics at all, my beef is primarily with how they are used from a studio pundit and broadcast perspective. I think they use the data irresponsibly. And I don't think you guys take them as gospel truth.
Deamon
Footballguy
Exactly. And people ONLY bash the analytics the times it doesn't work. It works more than it doesn't (I mean, by definition it does), but it seems every time it works there's never a "wow I wouldn't have done that if I was the coach, but he made a very smart decision". Probably my least fave thing about sports fandom is people blasting a decision after it doesn't work.
The biggest problem I have with this (constantly repeated) argument is that it's deeply flawed. Historical data that doesn't contain these "extra variable" by definition includes ALL the potential extra variables. Thus you over-emphasize the variable you feel argue against a decision while ignoring the fact that many of the events leading to the data set and "analytics solution" have already INCLUDED many scenarios that were similar variable wise.
I think most reasonable poeple realize that the data sets are imperfect...but too many are inherently and obtusely reluctant to account for them at all.
rockaction
Footballguy
I think that's what people like me are asking coaches and the elder statesmen in the game to do. To look or account for them at all. To the analytics guys, I'd ask the same thing I asked the SABR guys in baseball, which is maintain a bit of humility about what statistics can do predictively and what it can't account for.
Adam Harstad, who writes for FBG on theoretical dynasty matters, just tweeted about momentum the other day and said that just because you can't quantify it doesn't mean that it doesn't exist. You can't just wish it away. I forget his argument and the technical terms he was using, but I've always looked at it that this is a human game with human people participating in it. It is subject to limitations and conceits.
That's all I'd say about that. That sometimes analytics pretends to account for everything present in that iterative moment that you have to decide, and as Chaka points out, you might have historical data that accounts for certain things, but not for that specific time at that specific place in those specific circumstances. A human element is also something to take into account, not just a statistical model.
The momentum thing is a GREAT example though. If looking at a specific decision and using analytics, you don't think momentum is already baked into it? It absolutely IS ALREADY BAKED IN!
Where we can (and should) adjust is the TEAM SPECIFIC parameters. You have a monster QB that converts 4th and 2 or less on sneaks over 90% of the time? GREAT....maybe the analytics are under-estimating your odds. You have a midget backup in and you're bigger RB's are gimpy or both guards are backups who've been pushed around all day....maybe it's a little less certain and you can knock 10-15% off your odds.
I wrote one of my favorite scenarios earlier in this thread. 4th and 1 in very short FG range. Per the math...almost always worth it to go for it. Some specific scenarios where you wouldn't have little to do with momentum and a lot to do with score and clock time left (for example up 6-8 already with less then 3 minutes OTC.)
IE: Analytics may not give you your perfect odds, but most of the variables by which you wanna adjust are already included in the data (in both directions) ...momentum being a perfect example of a variable that's already baked in
Yep, I've been saying the same things for a long time as well. It's ironic to me how certain posters try to characterize anyone who disagrees with them as "anti-intellectual" or "math haters" when it couldn't be further from the truth. Many (most?) of us would love for the data to be higher quality and thus more valuable/predictive but are "intellectual enough" to understand that it simply isn't. The condescension is off-putting.
heckmanm
Footballguy
Any given play when compared to the analytics data is the very definition of "Small Sample Size". Whatever numbers come out of analytics are based on hundreds or thousands of similar situations, but even those didn't come out the same way 100% of the time.
As has been alluded to earlier, a good decision doesn't guarantee a good outcome in a sample size of ONE. If it did, Vegas wouldn't have all those blinky lights.
As has been alluded to earlier, a good decision doesn't guarantee a good outcome in a sample size of ONE. If it did, Vegas wouldn't have all those blinky lights.
ChiefD
Footballguy
I've long argued with friends about what I call "The Psychology of Points" - the part that analytics doesn't factor in. Taking the points can sometimes be the factor that gets a team going. Or supplies a psychological boost to the players.
The coach for the Chargers is terrible at this: he goes for it no matter what without thinking that "hey, I've failed twice in a row - maybe we should get some points on the board so our team can experience a successful drive."
The coach for the Chargers is terrible at this: he goes for it no matter what without thinking that "hey, I've failed twice in a row - maybe we should get some points on the board so our team can experience a successful drive."
Chaka
Footballguy
FTR I am not supporting old-school thinking in my criticism. Just considering the flaws in analytic presentation to the public.
I believe the future of Analytics is bright. There will likely come a day when an AI is calling the plays for an NFL franchise. Or maybe running the whole show. It will be interesting to see competing AIs emerge and consider the possibility of why two different AIs draw different conclusions based upon the same data.
I am not sure who is the official keeper of the analytics thrown around during NFL broadcasts (AWS?) does anyone know if there is public access to their methodology?
How far back does their research go? Are we using Bart Starr as a comparative data point to Josh Allen? Bronko Nagurski to Christian McCaffrey? How does the algorithm handicap a play with Mitch Trubisky under center as opposed to Kyler Murray? Can we simply assume they've got it "figured out"?
To the point of "momentum". Any attempt to distill a concept like "momentum" into a quantifiable variable is inherently magical thinking.
Are all "momentums" equal? Do they have a way to measure "momentum"? Is my "momentum" bigger than yours?
I wasn't even considering "momentum" as one of the limitless variables involved in comparing two seemingly, but not actually, identical situations.
What about fatigue? Or the underlying injuries the 22 guys on the field are uniquely dealing with on any given play? How do those injury considerations change when it's 40 degrees on the field or 75 degrees? Is crowd noise factored in? Does that impact change in the Meadowlands vs Levi's stadium? What about when the Center's wife is dealing with a difficult pregnancy or when Giselle gets bored with Tom's shtick and their preparation is negatively impacted? Does the algorithm have data points to account for that?
The number of variables is staggering, even for super computers running advanced algorithms written by humans.
daveR
Footballguy
In order to go 1-0 THIS WEEK, you have to be ahead after 60 minutes. You don't get to take an average result over 10 seasons. Emotion is important. Momentum is real. (Additionally, the loss of 3 is huge in close games because it invalidates future FGs and 2-pt conversions.)
Deamon
Footballguy
Ya and everyone is acting like 'going for it' based on analytics fails more than it succeeds. They talk about the 'missed 3 points' instead of the 'gained 7 points' or 'still 3 points but easier kick and more time off the clock". And the later happens more than the missing 3 points. People just like to claim that analytics 'don't work' when it fails, but NEVER glorify it near as much when it does.
ignatiusjreilly
Footballguy
It's loss aversion. We fixate on the downside because it causes more "pain" than the upside provides "pleasure".
It's also why I don't really buy the momentum arguments. It's not that I think momentum doesn't exist, it's just that its impact is a lot messier than people think, and it often gets deployed as a crutch argument. Take the example of "taking the points"; if scoring three points gives your team a psychological boost, it follows that scoring seven gives it an even bigger one. And if you have a 60% chance of converting the TD, you're more likely than not to get that boost. Note that I'm not citing momentum as a reason to go for it -- although I could imagine some scenarios where it might be, such as when a team caps off a furious 4th quarter comeback with a last second TD to draw within one -- just pointing out that it's never as neat as some people think, and that if you're using it in support of one action, it could just as easily be repurposed to argue for the opposite action
Deamon
Footballguy
Great stuff here. Loss aversion is definitely a real thing and blurs people's vision quite a bit. Like you say, going for it on 4th and getting the first down might give people a "nice that was great" vs missing it and this thread getting filled with posts. If there was a "post here when coaches do something you agree with" thread, it would clearly be much much smaller. People love to complain and they feel the emotion of the missed play or loss, much much more. So I can see how anti-analytics people are affected by that and inaccurately state that it was a bad decision, almost always AFTER the fact.
ignatiusjreilly
Footballguy
OK look bub, I'm going to take back every nice thing I've ever said about you if you put me in the position of defending NFL announcers. How dare you! HOW DARE YOU!!!!!
Seriously, I agree that we're not that far apart in terms of how we view analytics in a vacuum. The main dispute seems to be how it's described by laymen and in the media. My impression is that it's evolved something like this:
2012
NANTZ: And the Falcons facing 4th and goal from the 1, down 4 in the fourth quarter, will bring out their FG unit
SIMS: This is absolutely the right call, Jim. You gotta take the points here.
2021
MICHAELS: Cris, I suppose you'll tell me that the Chiefs should go for two here?
COLLINSWORTH: Absolutely. My minions at PFF say that it's clearly the right move
MICHAELS: Oh, you weird little nerds, you!
2022
[Amazon flashes a graphic on the screen of a TNF broadcast that shows, without a lot of context, that going for it in this situation is a 60/40 proposition]
I guess that's mostly progress, but I do see the point that the media is not doing enough to explain this stuff to people. In many ways I'd say we're in a similar place as we are with analytics in politics. FiveThirtyEight's data allows us to have smarter discussions about the state of specific races, but the numbers can give people a false sense of certainty, leading them to believe that when a candidate with 30% odds wins the election, it means Nate Silver was "wrong".
Last edited:
I think this is the case many times that teams take a TO instead of just taking the 5 yd penalty. Time outs are very valuable. How valuable you won't know until you need them. Losing 5 yds on a 3rd and 7 from your 35 yd line in the 1st quarter really doesn't change things in the overall scheme of things. Just take the delay of game and keep the time out.
Don't get me wrong, there are times when a delay of game penalty could be big and worth using the time out. But I think there are many more cases throughout a season where the 5yd penalty really doesn't change the game at all.
ignatiusjreilly
Footballguy
Generally agree, though first-half TOs are far less valuable. I might call one in that situation in the first quarter but take the penalty in the third
BobbyLayne
Footballguy
Good discussion on analytics fellas - it’s a tool, a data point, which influences but (obviously) should not dictate.
I like the approach Dan Campbell and his OC Ben Johnson take. They have a lot of reports getting generated by lower level staffers, quality control, self-scouting reports, et al. Analytics is a huge part of developing game plans and red zone schemes. Johnson played QB as a walk on; his double major was mathematics and computer science. He’s obviously the kind of coach who embraces analytics.
In game, though, it’s more about context. They coach by feel. Analytics is a tool but it’s more important to have situational awareness.
MCDC has had some brain cramp moments (first Vikings game), but philosophically I like their concept.
I like the approach Dan Campbell and his OC Ben Johnson take. They have a lot of reports getting generated by lower level staffers, quality control, self-scouting reports, et al. Analytics is a huge part of developing game plans and red zone schemes. Johnson played QB as a walk on; his double major was mathematics and computer science. He’s obviously the kind of coach who embraces analytics.
In game, though, it’s more about context. They coach by feel. Analytics is a tool but it’s more important to have situational awareness.
MCDC has had some brain cramp moments (first Vikings game), but philosophically I like their concept.
ignatiusjreilly
Footballguy
I would actually love to read accounts of how different NFL teams use analytics to inform in-game decisions. Ever since I learned that teams aren't allowed to calculate WP% on the fly, I've assumed they have some general rules of thumb: "If it's 4th and X or less and the ball is on the Y yard line or closer ..." with the HC always having the discretion to do what he thinks is the best call. But I can't recall ever seeing it reported. Maybe it's considered a trade secret or something.
ignatiusjreilly
Footballguy
Not watching the game but can someone explain how the hell this was allowed to happen?
Pip's Invitation
Footballguy
The wind is bad and Santos had missed an XP from that direction. They were at 4th and 27 after a couple of sacks (4th and 32 after they took a delay of game penalty to give the punter more room.) Still a bad decision, though.
ignatiusjreilly
Footballguy
I have no idea what just happened in the Pats-Raiders game but multiple people deserve to be fired for that.
ETA: Seriously, everything about that made no sense. Why are you running a draw? Why are you lateraling like you’re behind? Why is Meyers running backwards and flinging the ball wildly?
ETA: Seriously, everything about that made no sense. Why are you running a draw? Why are you lateraling like you’re behind? Why is Meyers running backwards and flinging the ball wildly?
Last edited:
ignatiusjreilly
Footballguy
College edition (but NFL, too): I had no problem with Michigan’s decision to go for it on 4th and goal on their opening drive, but the Philly Special needs to be mothballed for the next decade, or else someone needs to come up with a wrinkle that gives the offense more flexibility. As soon as the defense covers the QB, the play has no chance
Hot Sauce Guy
Footballguy
I can’t believe McCarthy’s abysmal end-game management on TNF hasn’t been mentioned. With 5 mins on the clock, the Titans got not one, but 2 more possessions.
What the actual F was he doing throwing deep on 2nd down?
Just atrocious.
What the actual F was he doing throwing deep on 2nd down?
Just atrocious.
Hot Sauce Guy
Footballguy
So, the chargers. In a meaningless game. Had starters in.
Mike Williams carted off with a back injury & now being reported that Joey Bosa is hurt, too.
That all sounds pretty not smart to me. Maybe rest players if the week 18 game doesn’t count?
Mike Williams carted off with a back injury & now being reported that Joey Bosa is hurt, too.
That all sounds pretty not smart to me. Maybe rest players if the week 18 game doesn’t count?
Tau837
Footballguy
Williams' injury has been reported as back spasms. Assuming that is true, he should be fine and will play in the wild card game. Bosa did not suffer a new injury or reaggravate a previous injury. As far as I know, no others were injured.
I think sometimes people overlook the reality of fielding a team for a game. The fewest players the Chargers played on offense this season in any of the first 16 games is 17. The fewest on defense is 16. They have 3 special teams specialists who don't overlap on offense defense (LS, PK, P) and 3+ players who play special teams and do not play on offense or defense, even though they are listed with an offensive/defensive position (they play on coverage and return teams).
That is 39 players at a minimum. The team only has 46-48 players active on gameday, depending on whether they activate 0-2 players from the practice squad. Today, they activated the max number (2) and had 48 active players. The Chargers absolutely should have held out or limited the snaps for any player dealing with a current injury, and they did that -- SCB Callahan was inactive because he is dealing with a current injury. Otherwise, they should have shifted more snaps to the depth players in the various rotations, and they did that in the second half. If there is any valid criticism, it is that they should have done that sooner.
Meanwhile, it was appropriate for the offense to try to build on its good game last week, its first good game in weeks, and build some momentum entering the playoffs. And on defense, both Bosa and James needed game reps with the starting defense to re-integrate and be ready for the playoffs; they got them and exited the game in the 2nd and 3rd quarters, respectively.
Last edited:
BobbyLayne
Footballguy
Coaches that go for it on 4th down to win the game?
Hell yeah
Hell yeah
Hot Sauce Guy
Footballguy
Agree to disagree.
Chargers had nothing to play for. Bosa re-injured his groin for nothing. He absolutely could have been rested, as could Mike Williams.
Williams has been notoriously fragile. And FYI, it was reported as more significant than back spasms. They took him for X-rays.
I don’t see why you start an important pass rusher coming off a groin injury and an oft-injured WR in a game that can’t help your team.
If either player misses the playoff game, it will prove to be a horrible decision.
ignatiusjreilly
Footballguy
I did think it was a slightly closer call given that the Lions were up 4. If the lead is 3, it's a no-brainer.
But overall, this is a point I've hit repeatedly in this thread: Don't give the ball back to HOF QBs with a chance for them to win the game! Get one yard and put them away.
BobbyLayne
Footballguy
the discussion we were having is why not
kick FG they can still tie/win the game
make it & it’s over
didn’t come all this way to play it safe
no hesitation from the sidelines
Hot Sauce Guy
Footballguy
Plus it’s not an automatic that if they don’t get it they lose.
GB still needs a TD to win it, and a FG can’t tie. And they have no TOS.
So at worst, you make the Packers drive 70+ yards for a touchdown with 0 TO. And to that point they’d been containing the packers pretty well - extremely well the last series.
I loved the decision there.
Similar threads
