Concepts some posters could benefit from learning

[RayClay]

Ray, I don't understand this post and your critique of mine.

WRT Yaz, my response is: sample size. go look at the # of regular season games he played, then go look at the # of playoff and WS games you are comparing it to.

Ray, I don't understand this post and your critique of mine.

That goes without saying.

Yaz had enough clutch situations to compare them to his mean averages for 1967 and career.

I'm sure you'll agree that regression to the mean is expected, that's fundamental.

Certainly his .875 average in the crucial 1967 was a small sample, but as a statistician I'm sure you'll agree there are plenty of evaluations that can be made. Do you want me to list the best tests to use with small samples?

For a .326 hitter 2.61 hits per 8 at bats is the mean. He had 7. In a normal distribution that is way off the extreme end of the curve. If anyone can do standard deviations and substitute the mean batting average for the arithmetic mean of 8 numbers, I'd appreciate it, I'm a social studies guy.

One would expect a pretty fair sample of playoff would regress to his career average (mean) of .285, I'm sure you agree. At 352 and 369 for WS and playoff averages, every clutch indicator is well above the mean, none below it, none close.

 

[makewayhomer]

I've seen that on FO. But remember that baseball Sabremeticians still consider a single season a relatively small sample size. And your first point talked about clutch situations, which reduces the sample size even further. And about kickers, which makes the sample miniscule by comparison. Yet that was the starting point for your statements?

yes...what is the confusion?

 

[pats-blue]

very high. you have very little data. the postseason sample is WAY smaller than his regular season sample.

by that theory you could pick out his games on Tuesdays, find a better BA on those days, and include that Yaz was a much better hitter on Tuesdays

This is absolutely hilarious! So if it fits your opinion it is ok to use but when someone SHOWS you stats of a player that seems to be "clutch" you discount it becaus of a too small of a sample size? You are funny!

So how big is the sample size of AV kicking a 29 yd FG in a driving rain where his plant foot slips? How big is the sample of kicking a FG in a Blizzard?

EXACTLY WHAT is your point at all? Do you have one? "Variance" and there is no "clutch"....wow facinating! You are trying to improve the board? Exactly how? Talking about statistics that you are pulling out of your ass? You wrote a whole paragraph on AV using stats you completely MADE UP!

You did apologize to another poster for being condesending in your initial post but you continue your crusade of trying to quantify football in baseball terms yet when posters pointout stats that point out "clutch" you dismiss them out of hand with various excuses. How is this discussion you claim to be trying to encourage.

You don't really seem to have a point except to try and banish the use of the word "clutch" yet other posters have PROVEN at least using your own arguments that there is such a thing even if you dismiss them even though they are using your "logic". You can use stats to make any argument you want so pretty much your "raising the board above rah, rah" is nothing but hot air of a fan of some other team that you don't even have enough pride to claim.

 

[RayClay]

Ray, I don't understand this post and your critique of mine.

WRT Yaz, my response is: sample size. go look at the # of regular season games he played, then go look at the # of playoff and WS games you are comparing it to.

I am comparing it to the mean. Seriously, get a basic text book.

 

[pats-blue]

yes...what is the confusion?

The confusion you are claiming "too small of a sample" when your lack of "clutch" theory wa blown out of the water by RayClay using Yaz...yet your diatribe about AV uses and even more miniscule sample yet you claim it as PROOF of your theory. Basically the confusion is that you talk out of both sides of your mouth and he is just pointing that out for all to see.

 

[makewayhomer]

I am comparing it to the mean. Seriously, get a basic text book.

would this text book say anything about sample sizes?

 

[makewayhomer]

The confusion you are claiming "too small of a sample" when your lack of "clutch" theory wa blown out of the water by RayClay using Yaz...yet your diatribe about AV uses and even more miniscule sample yet you claim it as PROOF of your theory. Basically the confusion is that you talk out of both sides of your mouth and he is just pointing that out for all to see.

I don't get what your saying. I'm saying that the samples are too small in both cases. We can't call Yaz clutch, we can't call AV clutch.

 

[makewayhomer]

but hot air of a fan of some other team that you don't even have enough pride to claim.

it's seriously sad that me disagreeing with the CW wisdom on this board makes me a fan of another team.

 

[Patriotic Fervor]

It's a mistake to let statistics/numbers define life. Mathematics is not life, in spite of howls of protests you'll hear from statisticians who often have no life except their numbers.

Statistics are, and always should be, used as a lamp post is for a late-nite drunk - i.e., more for support than illumination.

Useful in trend analysis (which I gather is what this phenomenon Sabremetrics is about), statistics have a power of their own, but suffer severe limitations when applied to everyday life.

By definition, the cannot possibly quantify the human element, which, for example, Ray Clay offers up in his analysis of Yaz.

There is a time to step up, which the great ones invariably do. The regular season, in any sport, can be defined as the hum-drum of existence, as the working man understands it. It's what the athlete does for a living, and what he gets paid for. That's his 9 - 5 job, so to speak.

Then the playoffs (or a very important regular season game) comes up, and while there is undoubtedly money involved, that's not what motivates the guy at this particular point in time. He'll call on his reserves, which are always there, as he needs them.

You can't play a Super Bowl game the way you an early September game, because the stakes have grown exponentially. It's a different beast altogether.

If this type of thing were the be-all and the end-all of existence, it's adoption would have been universal a long time ago, and if it was the be-all and the end-all of existence, and it worked as some hope or want, there'd be no reason to play the games. You could just do computer simulations, and while Madden is somewhat entertaining, it's a far cry from the real thing.

As a case in point, I offer up the game of Futures (commodities) trading. Random walk theory is big here, but if it were reducible to statistics, numbers, and associated forecasting, the major grain houses would have cornered their respective markets long, long ago.

All these numbers signify is past performance, and where human beings are involved, the past is indeed prologue, and past performance, as the disclaimer goes, is not necessarily indicative of future returns.

In short, use, but don't abuse, the numbers the game generates. Be appreciative of their usefulness, but understand from the get-go that that self-same usefulness has limitations.

That having been said, enjoy the game!

 

[RayClay]

very high. you have very little data. the postseason sample is WAY smaller than his regular season sample.

by that theory you could pick out his games on Tuesdays, find a better BA on those days, and include that Yaz was a much better hitter on Tuesdays

Yaz played 23 years. If you compiled all his Tuesday statistics and compared them against his career batting average and it showed a differences of two or three standard deviations you should infer a relationship. Absolutely.

 

[makewayhomer]

Yaz played 23 years. If you compiled all his Tuesday statistics and compared them against his career batting average and it showed a differences of two or three standard deviations you should infer a relationship. Absolutely.

I should have used a better example.

the idea is that comparing his 65 post-season at bats with his 11,988 regular season at-bats is an exercise in futility.

I'm quite certain I could find other random 65 at bat samples in which he performed well above his 23 year average, and others in which he performed horrendously.

 

[makewayhomer]

If this type of thing were the be-all and the end-all of existence, it's adoption would have been universal a long time ago, and if it was the be-all and the end-all of existence, and it worked as some hope or want, there'd be no reason to play the games. You could just do computer simulations, and while Madden is somewhat entertaining, it's a far cry from the real thing.

PF, there are various successfull baseball teams who are now SABR run organizations, and almost every team has at least 1 SABR type guys on their staff. the Red Sox, Indians, and A's chief among the SABR run team. this stuff IS very prevelant and is being used to make decisions today, whether you know it or not.

In short, use, but don't abuse, the numbers the game generates. Be appreciative of their usefulness, but understand from the get-go that that self-same usefulness has limitations.

I have never said that #'s should be 100% of the decision. they are extremely useful, but the human element must always play a part in the decision making process. this is why the pioneering SABR baseball site Baseball Prospectus has Kevin Goldstein on their staff, who is an old school crusty scout type, not a statistician.

and here is a real world example from just this month from an online chat with the new Pirates GM.

The Pirates upper management has widely ignored OBP (on base percentage) in the past. How important will OBP be in player evaluation under your leadership?
-- Eric S., Pennsboro, W.Va

We are going to utilize several objective measures of player performance to evaluate and develop players. We'll rely on the more traditional objective evaluations: OPS (on base percentage plus slugging percentage) , WHIP (walks and hits per inning pitched), Runs Created, ERC (Component ERA), GB/FB (ground ball to fly ball ratio), K/9 (strikeouts per nine innings), K/BB (strikeouts to walks ratio), BB%, etc., but we'll also look to rely on some of the more recent variations: VORP (value over replacement player), Relative Performance, EqAve (equivalent average), EqOBP (equivalent on base percentage), EqSLG (equivalent slugging percentage), BIP% (balls put into play percentage), wOBA (weighted on base average), Range Factor, PMR (probabilistic model of range) and Zone Rating.

That said, we will continue to stress the importance of our subjective evaluations. Succinctly stated, we believe that a combination of quality objective and subjective analysis will allow us to maximize our probability of success and to make the best possible decisions

 

[texpat]

The best team always wins. How do I know? Circular logic.

There is only one method of determining the best team, that is the team left after all games have been played.

Statistics are only a by-product. The only purpose for every game, playoff and Super bowl is to determine the best team, therefore the team remaining is, by definition, the best.

The outcome of any single game determines only the best team between those 2 teams on that day.

For example, let's say that the AFC has the 5 best teams in the league. Let's say that by some fluke, Dallas wins the NFC, and then beats Pitt in the Super Bowl, does that mean Dallas is better than NE. And then is Dallas the best team in the NFL. I say no, they are only the best team between themselves and Pitt. (Now with all that said, Dallas may be better than 28 or so teams in the league, but the result of the SB doesn't determine that.)

 

[pats-blue]

I don't get what your saying. I'm saying that the samples are too small in both cases. We can't call Yaz clutch, we can't call AV clutch.

Yes you said AV wasn't clutch...but you used stats to point that out using "stats" you pulled out of your butt. The fact is that he made many high pressure kicks and you assigned arbitrary numbers to "making" kicks in a blizarred. You also assign the same "stats" to kicks made in different weather as to those made in a dome and then finally you assign the SAME importance to a kick made or missed in the 1st quarter to one with a few seconds left on the clock. Fine to use stats when talking about flipping a coin but humans feel "pressure" during certain acts and times and that effects heartbeat, breathing, ect and can effect motor function to a degree which isn't shown how you interpret stats.

I suggest while you read your "stat" books maybe, one play a sport and you would understand more. Second read Lt. Col. D. Grossman's book, The Psychological Effects of Combat. It may to some appear to not be relevant to your "stats" but I assure you it is very applicable. The effects of a high pressure situation effect the SNS system of the body just like those in combat and police situations Lt. Col Grossman discusses in his book alibet on a smaller scale. The effect on a professional athlete would be in the lower ranges of the Police Officer and Soldier scale and the effects on the SNS system accordingly would also be on the lower side but STILL prevelant.

Some police officers react during high stress situations better than others. Some simply shut down and others respond "heroicly" or do what needs to be done to solve the situation. The same thing with soldiers...some never fire thier weapons in combat and others win the Medal of Honor. These are your "clutch" performers on the battlefield and on our streets.

Some athletes are able to deal with high stress situations better and thier bodies and minds provide them with more focus rather than inhibiting them like others due to SNS system overload.

Stats be damned. Anyone that thinks a kick or an at bat in the beginning of the game or in the regular season isthe same as in the end or win or go home situation has never played a sport higher than gym class or sat on the bench. It is simple as that. You are trying to use math class to quantify a situation it appears you know nothing about and you don't even realize it.

 

[RayClay]

would this text book say anything about sample sizes?

Yes and it would reveal many ways of testing differently to draw inferences.

For instance you keep mentioning variance. I bet you don't know what it means in statistics. The only thing it means is related to standard deviation, which I've mentioned about three times and you don't seem to be familiar with.

This is Variance

In a normal distribution with 2.61 hits per 8 trials being the mean (for a .326 hitter), perhaps a statistician can work out how many standard deviations 7 hits would be.

 

[RayClay]

The outcome of any single game determines only the best team between those 2 teams on that day.

For example, let's say that the AFC has the 5 best teams in the league. Let's say that by some fluke, Dallas wins the NFC, and then beats Pitt in the Super Bowl, does that mean Dallas is better than NE. And then is Dallas the best team in the NFL. I say no, they are only the best team between themselves and Pitt. (Now with all that said, Dallas may be better than 28 or so teams in the league, but the result of the SB doesn't determine that.)

All it means is Dallas is the best in the NFL, nothing else is proven.

Not true. They keep score. There are not judges scores. Everyone can have an opinion of the best team, that's totally subjective.

There are extraneous variables, such as schedule. Every year, teams start again with zero wins.

I go back to my example. If writers and other experts were to judge, the 1960 Celtics would never have been considered the best.

Trust me, I grew up with them. They were considered less talented than the Sixers and Lakers and St. Louis Hawks.

They were underrated as much as the 2001 Patriots.

Yet they won 11 times. That's why they keep score.

 

[RayClay]

I should have used a better example.

the idea is that comparing his 65 post-season at bats with his 11,988 regular season at-bats is an exercise in futility.

I'm quite certain I could find other random 65 at bat samples in which he performed well above his 23 year average, and others in which he performed horrendously.

You simply compare means. There are probably 20 means tests in your book.

 

[makewayhomer]

You simply compare means. There are probably 20 means tests in your book.

ok, this is rude, but you need to stop talking about statistics. it's clear you know nothing about the topic.

the fact that Yaz experienced a short hot streak is basically meaningless.

in 1995, Mike Benjamin had a career batting average of .186. he then went on a 14 for 18 tear (!). what does this mean? it means absolutely nothing.

it doesn't mean the next time June 14th rolls around, he's going to on a 14-18 streak. it happened once. but that is not a predictor of future events.

 

[RayClay]

The confusion you are claiming "too small of a sample" when your lack of "clutch" theory wa blown out of the water by RayClay using Yaz...yet your diatribe about AV uses and even more miniscule sample yet you claim it as PROOF of your theory. Basically the confusion is that you talk out of both sides of your mouth and he is just pointing that out for all to see.

Saving money by getting a kicker who's statistically better except in the clutch is like saving money by getting a bullet proof vest that almost stops a bullet.

 

[RayClay]

ok, this is rude, but you need to stop talking about statistics. it's clear you know nothing about the topic.

the fact that Yaz experienced a short hot streak is basically meaningless.

in 1995, Mike Benjamin had a career batting average of .186. he then went on a 14 for 18 tear (!). what does this mean? it means absolutely nothing.

it doesn't mean the next time June 14th rolls around, he's going to on a 14-18 streak. it happened once. but that is not a predictor of future events.

this is rude, but you need to stop talking about statistics. it's clear you know nothing about the topic.

You're killing me.

Just for fun, could you at least look up some statistical jargon on google since you've pwned me so?

Variance is a start, do you know what it means?