why umpire data?

[mack the knife]

I do not use ump data at all, but I notice many posters here live and die with the umpire data. I think this is a mistake, at least as far as how the data are being used. Here is my argument:

I noticed NDFAN, for example, said the ump in the SF/LA game is a homer ump because 49 of 75 home teams have won. But wouldn't you expect home teams to win close to that percentage anyways? In this case, it does seem to be a little higher win % than one would expect on average, but the larger point is that it is probably premature to be calling some of these umps "over umps" or "homers," especially with small samples. What is needed to make the ump data more valuable is a baseline of what would be expected from the home/away teams, all things equal, and how much the ump matters in deviating from this baseline.

For example, if UMP A has been calling mostly Yanks home games, would it surprise you to know that home teams have won 64% of their games?

Maybe one could match up the ump with the home/away line, and see how many games the home team won compared to the average home line. If home teams were lined at -160 ave, that ump better have a home win % higher than 62.5% to be called a homer, likely around 70%, and around 50% to be called a dog ump.

My argument is that ump data by itself might not be very predictive and I don't understand poster's dependence on it, at least the way it is currently used. If someone could match up the ump data with home lines, however, then I think it might me more valuable.

Anyone....?

mack

 

[Palmetto Pimp]

i think ump info can be valuble, but in the circumstance that you are talking about ( a home ump) that is silly, for reasons you have stated

 

[Aces High]

never really use ump date either, :shrug: but

it becomes important in over and under plays, depending on the ump's "strikezone". small zones lead to walks and potential runs for an over, and bigger zones lead to more outs for the under players.

just my 2 cents

 

[mack the knife]

This is what I'm talking about:

"The Ump is 13-9 Over / Under with an average of 10.13 runs scored"

1) for a sample of 22 games, 13 overs is not statistically different from chance (11-11), and if the games were lined -120 over, this is exactly the record you'd expect.

2) If the lined totals on these 22 games averaged 9.5 or 10, then this average of 10.13 is also what should be expected, and not an indication of an over bias.

3) I don't know if averages are the appropriate summary statistic for small samples, because if one game was 13-9, the average is skewed by this outlier.

mack

 

[Bdav11]

True you are mack about the outlier, but there are statistical test which will take the outlier into effect and correct for that particular case which is odd from the norm. I know many stat test that can be run which can be pretty accurate. I have a masters degree in stat and I think if I apply some models, I might be able to help out a little. Any suggestion about categories and where to find data from the beginning of the season could result in some trial and error to work out a pretty accurate model. Let me know if anyone is interested.

 

[Aces High]

Bdav11,

ouch !!!!

flashback to an 8am stat class my senior year with that post. :grins:

gonna hurt somebody :toast:

 

[mack the knife]

I am simply impressed that bdav11 waited nearly TWO YEARS since registering to make his second post ever in this thread. Talk about patience!

 

[loophole]

i am one that believes umpiring stats are important capping factors in definable situations. bdav11, i would love to see someone with some working expertise go over the numbers. best situational stat breakdown for umps on the net is located @ sportsbettingworld. com, although they are currenly a few days behind current because of preparations for football season. if you can, take a look at overall o/u as a predictive tool as well as any correlation between o/u record and k/bb ratio.

 

[Bdav11]

LOL...I have been out of the scene for a while because I got in such a hole from irresponible bankroll plays...kinda like what is going on w/ B1n. I lost allot of money while in school, but was always wanted to apply some of the models I know to find a good one. A year after I graduated and got a job, I now have paid all my debt and have some money to play with. So now I'm back and have made smaller and more paticient plays...used to try and bet 6-7 games a day, and now I might not bet a single one if not to sure on it. I still lose my share, but doesn't everyone??? Anyway, I am not a pro by any means, but would like to help out in any fashion I could. W/ some suggestions about what catagories and where to find the data, I would be glad to post the results of the models in order to find one that works at least 65% of the time...which over the long run, would help increase your bank. Thanks in advance
Bdav11

 

[Bdav11]

loop
I will take a look at those and let you know what I can see. I would also like some other suggestions other than about umps to see about trying to complete a good model on sides as well. Thanks again
Bdav11

 

[kcwolf]

Umpire information, used correctly, is invaluable.

Try looking at the stats in conjunction with your regular handicapping. I use 15-18 keys for every game I play, and each key is adjusted, as far as importance, as the year goes along.

Umpires get more important as the season wears on. gsp is the expert in umpire information, maybe he will reply.

A very good question you bring up, mack.

gl today!

 

[ndnfan]

Absolutely correct in having to use the information correctly.

I still don't consider myself an expert on Umps, but I do like to chart certain Umps tendencies and characteristics. I agree it's foolish to base an entire game on the Ump...if you do that you're in trouble. Seems like this was an angle people got caught up in for a while around here...and NOT using it in the right way at times.

Mack...I know what you're trying to say as far as stuff such as "HOMER UMPS" and stuff, but when the numbers are so far out of whack, it becomes more of truth rather than coincidence. I'm the last person that will use MEANINGLESS TRENDS, and I don't use any trend unless it has meaning to it. Alot is coincidence, but you have to be able to separate the few that aren't

There's way too much to get into on this subject, and there's people out there that know much more on Umps than me, but if you don't think an Ump can change the whole perplexion of a game in certain matchups....I have to disagree. There's times when control pitchers who rely on hitting the corners really benefit from an Ump with the larger strikezones. It's a lot easier to pitch when you're ahead of the count than having to "groove" a 3-0 or a 3-1 85 MPH fastball to these power hitters and that 85 MPH fastball is what some of these control pitchers have at best.

Also certain Umps will call the high strike or low stike and not the corners or vice-versa and it's important to matchup the pitching styles with the Umps tendencies. For example a pitcher that likes pitching down in the strikezone is not gonna benefit from an Ump that will give the high strike call and not the low one that he needs to get to be effective.

But again, the Ump is just another key or angle you should use when capping.....I never base a play on just the Ump, but i will eliminate a play based on it in certain cases.

I do know what your trying to say though...for example: almost like people that post about a team that struggles vs lefties, yet vs the particular starting lefty going today, over half the players on the team are hitting over say .400 career vs him. It's all about the certain situation and the amount of weight you have to apply to the situation....which will vary depending many factors.


Don't know if this helps any, but the topic can go a lot deeper than I just went.

Good Luck to you!

-ndnfan

 

[cisco]

If this ump information isn't important then why is it that you never know who the ump is at the start of the series?
Because it IS important.

If I see Hirshbeck is 11-1 to the under, you can bet that i'm on the under as well.

 

[ndnfan]

cisco...You're correct. Always heard the Umpires schedules are not released to the public for both gambling and security reasons.
If you don't incorporate Umping into your capping and are successful at baseball...congrats to you....I just know I can't grind out what I consider a successful baseball season without knowing the Ump info in the majority of the games.

Think the same goes for NBA officials which are crucial in totals at times and I think a lot of sportbooks would agree. Sometimes you can get a big advantage on the 2nd half totals in the NBA after finding out the refs.

 

[mack the knife]

Nice collection of thoughts so far.

I want to reiterate my main point that seems to have missed. I never said the umpires don't matter. I clearly said that umpire data is often used incorrectly because (1) sometimes there are so few cases, the data are basically meaningless, and (2) sometimes the data are not sufficiently different from what already be expected otherwise.

Cisco: two questions. If Ump A is 3-0 to the under, do you think this is a valuable indicator? What about 4-0? What about 6-1? At what point (number of cases, skewness within the data) do you think a bias can be detected i these small_N examples? Second, why do you think 11-1 would continue? Literally, this implies 44-4 is just around the corner, which it ain't. The past is not always predictive of the future beyond what the odds/line offers the bettor....

mack

 

[nostromo]

The argument is also valid for other stats averaged over small sample data.

Team's record vs lefties is one of them (as there are far less LHP than RHP), especially if you use a number like team's batting avg vs lefties over last 10 games - typically they face 3 LH starters over that stretch and it will make a world of difference to that number if just one of them is either Glavine or Hampton.

Same goes for stats like pitcher's ERA over L3 (what if one of them was at Coors Field?) and others.

 

[infinii]

I may be ignorant by saying this but if you guys are claiming that the sample size is too small for the trend to have any validity.

Can't I argue that a large sample size is invalid as well because the rosters and coaching is different? different coaching staffs being very important because the style of the head coach may be more inducive to over or under.

 

[jmizeus]

using umpire information is very valuable! are u nuts?
here is a good example!

cincinnati plays pittsburgh in a 3 game series!

game 1 goes under
then game 2 has a ump w/an advantage for an over play -well there is a cappers point of view you have a nice play on the over in game 2!

now if game 1 was an over and game 2 's ump is rated asan over -i might lay off! think thyis is a very big key in judging total plays!

 

[mack the knife]

Infinii,

In almost any situation, a larger sample size is superior. But the data in the sample have to be equivalent. That is, data over 5 years on a college team is probably not much better than the prior year alone because of player turnover, etc.. Rule changes in sorts can also make prior data incompatible, like in hockey with the OT rule change. But if all you are looking for is whether an ump is a homer or not, I don't see what difference 1999 and 2002 are--they should be pooled together. For most teams, I use the single, current year data.

jmizeus, I have no idea what you are saying, but you seem to entirely misunderstand the thread. Please read again. I am getting tired of repeating that I never said umpires don't matter. But I think you have given me an excellent example of why I said the umpire data are often misused....

mack

 

[jmizeus]

sorry,maybe i jumped the gun! maybe ur not saying that! but i just showed you my thoughts on an ump! i basically use them for totals only! to busy to go into it any further! but i think my edge i use is a n ump thats record sways towards either the over or under -depends on the last game in the series! i think totals you can make money year in and year out if you on that just alone! like i said that is my opinion!@ i know alot others might cap the umps differently,but they also might have the time to do so!

gl and like i said sorry for misunderstanding you

MIZE