Just how important is "clutch", really?

Cool!

Next step is to translate those SD's (which are called z-scores) into probabilities. If you don't have a program that does that you can use the "cumulative" table here:
https://en.wikipedia.org/wiki/Standard_normal_table

That cumulative table tells you the area under the curve below the z-score. So z = 1.77 would be 0.96164 on that chart, meaning 96.164% probability any stat there belongs to a QB that is NOT league average.

So Brady with 2.0 SD has a 97.7% probability he's not league average. As long as it's 95% or above a statistician would say the value is statistically speaking significant. So Brady, Peyton and Kelly all pass, with this one stat.

 

That trend line that you are looking at is the "expected win%" in such games relative to overall win%?

 

NO. That is a team stat, not a QB stat. It doesn't say anything in particular about his ability as a QB, nor does it say anything about his being clutch. Because the Pats have been so consistently good for so long, everyone who has played on that team has basically that same Z-score. But needless to say, not everyone on that Patriot roster is clearly above average.

 

Yes it's a team stat. I just used the shorthand "belongs to a QB". Point is, the team is playing better than expected in this situation in a statistically significant way. Some percentage (you choose.. maybe it's that 15-20%?) of that is due to the QB.

So yes it's a team stat, but it does say something about the QB (how much depends on that percentage). Note that only if it wasn't statistically significant does it say nothing about the QB (statistically speaking).

 

That % depends on the individual games. As I explained previously, some games it might be a big number (shootout with huge passing stats) and some it might be small (defensive struggle where team wins despite poor game from QB).

 

There is a reason "pressured situations" are referred to as "pressured situations..."

Now I'm off to throw a few cubes of solid water into a nice de-warmed beverage from the heat-sucker.

 

No argument there. You just have to subjectively decide what the average percentage was in this case (remember this is comparing to expected, i.e average, drop-off). Personally, I do think the 15-20% makes intuitive sense.

 

You do that, while you're at it, debunk other things from science like gravity, evolution, global warming and relativity...

 

But I don't think there is a meaningful average.

But if you take some %, do you do that after determining the standard deviation or before. Seems to me it should be before. So if, as suggested before, the Pats' expected close game win% was 60% and they actually won 69%, then if you attribute 15% to the QB as you suggest, then Brady is only responsible for 15% of that 9% differential, or 1.35%. But if that's the Brady number, I suspect it is no longer a statistically significant number at all. And with Brady's roughly 100 close games, he'd only be adding 1-2 games above replacement level.

 

By definition there is an average.. you just have to find the one most meaningful to you.

Doesn't matter if you do it before or after you'll get the same result. If you do it "before", then the drop-off is also 15% of what it would otherwise be and the standard deviation changes correspondingly, etc.. Reason mathematically the result is the same is because you're just multiplying by a scalar.. doesn't matter when you do it.

But your final sentence is correct.. the effect seen due JUST to this stat is fairly small.

 

I understand that one can calculate an average, but when you have huge variations from game to game the average doesn't mean much. For example, a team that scores 35, 40, 40, 50, 0, 3, 7, and 7 points in 8 games averages 22.75 pts per game. But that doesn't describe that team's offense at all and is not predictive of anything. Same for the amount of impact a QB may have from game to game and/or across different QBs.

 

Intuitively I disagree. There's something called the Central Limit Theorem that says the distribution (as sample size increases) of sums of random variables is always normally distributed, no matter what types of distributions those individual random variables came from.

The "percentage" the QB is responsible for in any single game will be made up of the sum of many random variables (random in a statistical sense.. which we know is true because what happens on one play is statistically independent from what happens on another play). So that means the percentages the QB is responsible for across games MUST be distributed normally.. and we know this even without having the ability to measure that percentage even once.

Meaning.. "average" is probably the most meaningful single number you could use to summarize the contribution of the QB across games.

 

No, that is YOUR incredibly rigid definition of clutch that fits your argument, that is not the definition of clutch, the definition of clutch is just how I spelled it out, performing well under pressure, which can include rising to the occasion, which is the pinnacle of clutch, heroic, but clutch is simply performing well under pressure.

Also, people who tend to choke, can at times be clutch, just as people who tend to be clutch can choke, this "automaton" you're seeking that has to be clutch at all times, is fantasy, and in no way correlates to real life.

If you can sit there and say there is no such thing as choking or clutch, you aren't a very observant sports enthusiast, or you're doing it on purpose to elicit a reaction, i.e., contrarian, but the contrarian does help us to think so...

 

At the end of the day, we're back to the same thing: good to great players doing in "big" situations the same thing they do all the time: play great. Bad players do in "big" situations what they do all the time: play badly.

If it's the norm, how is it clutch? Just because you want to say it's somehow different, or bigger, or whatever, doesn't make it so.

 

So you're saying that Dalton plays close to his avg in the playoffs?

regular season 88.4

playoffs 57.8 with a high of 67.0, 1 TD and 6 INTs for his playoff career.

No excuse for bad team, poor D, no weapons or what not, he has simply choked in every playoff game he's been in.

 

Do you even read what people write? Playing below your norm may be "choking", if you want to call it that. But playing at your normal level isn't clutch. It's normal. Further, if a QB plays at his normal level for 56 minutes, and plays at his normal level in the last 4, if you call the final 4 minute clutch, you'd also have to call his first 56 minutes clutch. At which point, there is no clutch. There's simply his normal level of play.

Good/ great QBs play good/ great in the first quarter, second quarter, third quarter, and fourth quarter. That's why they're good/ great. In the fourth, good/ great QBs do what they do...they make plays...and people sensationalize those plays, acting as if the plays from the previous 56 minutes were less important. But the plays from the previous 56 were important. Without those plays, the ones in the final 4 would be meaningless.

 

I haven't been reading the other posts in this thread, but I personally think the definition of clutch is to not be mentally affected in high pressure situations. Conversely, choking is the opposite, letting your nerves get the best of you when things matter the most. So yes, clutch, to me, can be performing normally (or better than normal) under stressful, high stakes conditions. To me, someone who is clutch does not feel the mental effects that lead to choking:

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2708085/

 

Playing at a normal level during high pressure situations is not normal, people are more inclined to perform worse in high pressure situations, that is precisely what clutch is, being able to perform your task at or near your normal level during high pressure situations.

QB's that choke, far out weigh those that are clutch.

 

But it is normal for good/ great QBS.

Fineas has explained a multitude of times why what you're calling "high pressure situations" really aren't, for these athletes who are the best in the world.

 

This is now like the third different definition of choke in the thread.

LOL

 

Just to bring you (partly) up to speed, that exact link and description of what definition science tends to use was talked about in post #391 and #392 on page 20.

I proposed what is arguably a better way of defining it for team sports (but reduces to the same definition in the link for an individual). Since it's possible in team sports for another unit/person on the team to be responsible for an observed difference between an individual's stat in non-pressure vs. pressure situations, I think you want to subtract off the average drop-off (difference) league-wide from non-pressure to pressure situations first, THEN look at the differential. We're calling that "Differential Performance Under Pressure" = DPUP.

Keep in mind it's just a stat.. how it's interpreted is up to the observer, but it does measure along a "choking" to "clutch" continuum.

 

He hypothesized that for professional athletes the points on their stress vs. performance curve in non-pressure vs. pressure situations are similar. No one knows whether that's true or not, and you can't directly test it because we can't directly measure the amount of pressure in different situations.

However.. IF you assume to know which conditions have greater pressure than other conditions, then DPUP allows you to go after that problem, with the key uncertainties being: 1) how much difference in pressure there was, and 2) the proportion of the statistically significant difference being due to the individual.

 

Well that is pure BS, Peyton is a choker, as has been pointed out, his performances vary way more than they do in the regular season, look at how many times his team has exited the playoffs on a bad performance by him, and conversely Flacco upped his play in the playoffs, these debunk that theory.

You make it sound as if that being an athlete makes you immune to being human, they are just gifted people, nothing more, and are subject to all things human.

 

To be fair Peyton is the most overrated QB in history. Always had great (if not HOF) talent around him, playing in a dome. Playing in a dome, at home, in the playoffs, his teams have lost 8 playoff games at home.

8 one and dones. Most of them at home.

Let's not even talk about his latest two super bowl performances (good thing that defense was stellar this last time around).

 

Joe Namath is in the Hall of Fame. I wanna see how you argue Peyton is more overrated than Namath. Here are his stats so you can see how bad they are:
http://www.pro-football-reference.com/players/N/NamaJo00.htm

For me, Peyton in his prime was one of the best ever. I agree he underperformed in the playoffs. Still.. no way he's that overrated IMO.

 

It's not BS, Finster. People become accustomed to pressure, the more they're exposed to it. The good/ great NFL QBs have been exposed to pretty drastic pressure for many, many years.

 

Also, I don't think that just because choking, or playing less than normal under pressure, exists, that that means that clutch exists.

Clutch perhaps exists in the way that cbrad has defined it, but, that doesn't really mean what Average Joe Sports Watcher means when he calls a player "clutch."

 

To be fair Brad, you can't compare stats from different eras, Namath's numbers were good for his time.

6 seasons he was top 3 in yards, leading the league 3 times, 7 times in the top 5 in TDs, 4 times in the top 2, leading the league once, and to be clear, I think Namath is overrated as well.

However, in his SB year, he was very good in the playoffs, and very efficient in the SB, posting an 83.3 rating and completing 60% of his passes, which are very good marks for the time, Manning has never done the equivalent.

Manning's playoff record is very deceiving, you could say look at 2003, he had a 106 playoff rating, which would lead you to believe he had a good playoffs, but he had a 35.5 rating in the AFC championship game.

2004 he had a 107 rating, but a 69 rating in their loss, in 2006 when he won the SB, he had one of his worst post seasons, a 70 rating, throwing 3 TDs and 7 INTs in 4 games, the defense won for him that year.

I agree with Jdang as far as being a very overrated QB, and I think he has a strong argument for most overrated, simply because of how high he is rated, as one of the GOATS, where Joe Willy is not even close to that conversation, but when it matters most, in the playoffs, Manning has had way too many failures, and that is where the metal meets the meat so to speak, which is why I think he is a very overrated QB.

 

There is no guarantee that they get better at dealing with it though, Manning is a great example, 2013 when he set all those records, and had a 115 rating, what happened in the SB, he choked, as he has done many many times in the playoffs, and he has had a vast amount of experience.

 

Manning's numbers are FAR more impressive relative to his era. Compare Namath to Staubach or Jurgenson to see how far away he is from those that had good passing stats in his time.

And how the hell do you put a guy who threw 173 TD passes vs. 220 INT's in the regular season over his career in the Hall of Fame? My point is.. Namath should NOT be in the Hall of Fame, but Peyton certainly should be.

And you can talk about how good Peyton's teams were, but you put a decent QB on those teams they won't get anywhere close to Peyton's passing stats. That man was one of the greatest regular season QB's and maybe the best on-field coordinator ever.

Playoffs, and also in many other pressure situations, yes Peyton is statistically speaking almost as much of a "choker" as Brady is "clutch" using DPUP (ignoring for the moment other ways people might define "clutch"). But no way IMO is Peyton overrated like Namath is. Peyton the regular season QB was as elite as we've seen, and I think most people aren't overrating him much in the playoffs anyway.

 

Except that the average is different for each QB and attempting to answer that question with a % impact effectively answers the bigger question. So it just layers guessing on top of guessing in a pseudo-scientific way to make it seem objective and scientific when it really isn't at all.

 

Of course we can test it and the data supports what I am saying. The most pressure-filled single player events, like FGs or basketball FTs in closing seconds/minutes of close games, do not show any notable drop from the norms. On average, NFL kickers and NBA players perform as well in those late pressure situations as they do in regular, non-pressure situations. Those events happen after a stoppage of play, the game on the line, the cameras and fans focused on one player, etc., and we see no real drop in performance overall. Those are the events these guys live for and have dealt with their whole athletic lives.

 

So you admit that pressure can affect their play?

 

Never said it was scientific. I said it's something you subjectively decide upon. Point is, there IS a ground truth answer to this question. Whether your estimate is accurate or not who knows. And "taking the average" is NOT where the problem lies. You have to compare apples to apples, so you have to do that. Anytime you take an average you remove information. So? It still contains useful information.

How do you know FT's in the last 2 minutes are "greater pressure" situations than FT's at other times? You have to ASSUME that (read the post you quoted again). I mean I'd assume it too, but you're the one that said pressure in regular season, playoff and championship games is essentially the same, showing that people can disagree on these things. Point is, we can't directly measure pressure here.

And FG's are also team stats, given how technical everyone here wants to be about things being team stats.

Either way, the problem you'll have in trying to "prove" anything here is what "similar performance" means. I mean there's always a distribution, so there are always some players, even in individual sports, that are statistically speaking performing better than expected.

 

Your Peyton as a choker position is funny because according to many of the measures of clutch that people propose, Manning is notably better than Brady. For example, in the very same close (7 pts or less) games set they were looking at for win%, when you actually look at the QB performance in the last 6 minutes Manning (94.2 rating) did far better than Brady (79.5 rating). Flacco is also below average in those scenarios (73.2 rating). So while you may have a sense or impression that Brady and Flacco step it up under pressure and Manning cowers under pressure, some fo the more relevant measures show that to be untrue.

Yes, athletes are human, but humans adapt to their environments and when you do something and practice something enough you do become largely immune to the pressures. Think about something like public speaking. For people who aren't used to doing it it is very pressure-filled and stressful. For people who do it a lot, it is second nature. So for someone like Obama, there is no place or crowd to pressure-filled or too intimidating. He can give a speech to literally the entire world and he's not going to start stuttering, sweating, etc. It is second nature to him. Similarly, the pressure associated with walking a tightrope from extreme heights would be unbearable for most people. But there are some people who do that all the time without breaking a sweat. With practice and expertise, the pressure factor pretty much disappears. And in the scheme of things, the pressure in sports is pretty mild. The ramifications are winning or losing a game and maybe some impact on future salary (although that pretty much never comes down to performance in an isolated pressure instance or two). Nobody's life is at risk. Nobody's family health or safety is at issue. Etc.

 

I've never argued that pressure can't negatively affect play.

I've argued that the good/ great players, who also happen to usually be the "clutch" players, are good/great all the time. I've argued, as has Fineas, that even role players, like Horry and others, who are said to be "clutch," aren't performing, really, above their norm.

Clutch is something that we think is something, because humans want to believe in greatness. It's why we love super heroes.

 

Is there a real debate over whether there is more pressure on a FT with 1 second less in a 1 pt game than with 40 minutes left in a 20 pt game? I think that's got to be pretty close to undisputed. I believe that pressure difference doesn't matter much because the average NBA player makes FTs in both situations at about the same rate, but I don't think there is a real argument that there is no difference in pressure.

FGs are only marginally team stats. Sure, there may be a bad snap on 1-2% of FG attempts and a bad hold on another 2-3% of FG attempts, but as among players the kicker has got to be responsible for 90+%. Yes, that's a rough estimate, but I think it's on the conservative side. I'd say its probably more like 95+%. Weather, on the other hand, is a real factor and needs to be looked at for comparing two or more kickers, but overall it doesn't seem to have any more effect in close games than in non-close games.

 

If you go the DPUP route and compare drop-offs relative to league drop-offs, Peyton is either average or worse than average in most clutch stats, opposite of Brady.

Of course, if you use a different definition like "absolute performance in pressure situations", then you're right.

 

Personally, until you said playoffs and championships are essentially the same pressure-wise, I thought that no one would argue they were. So I'm betting there's someone out there who would argue pressure on a FT with 1 second left is no different than in other situations. But no, IMO there is no serious debate about that, and IMO there is also no serious debate that playoffs and championships are greater pressure situations than regular season games (just my IMO about what most people think though).

And yes, FG's are only marginally team stats. I'd also agree that it's at least 95%+ on the kicker.