HOF
Posted 25 February 2015 - 09:35 PM
Every time I read the Nats rotation, I just stop and think wow. And it's unlikely any other team in their division wins more than about 75 games. They should win the division running away again. Think they have a very legit shot at winning 100 this season.
I hope they crash and burn...F 'em. 
HOF
Posted 25 February 2015 - 09:38 PM
Sorry Shack...I didn't read back your original posta bout the article...well said., sir.
Fair-weather ex-diehard
Posted 26 February 2015 - 06:29 AM
For example, a model with an R-value of .1 might be significant if you're trying to predict the result of basketball games ...
I've used .5 or .6 before. Those aren't necessarily high but I'd still argue that they have significance.
Yes, I am aware that p-values are typically associated with significance, but that is certainly not the lone benchmark for trustworthiness.
The assertion that an R-value of 0.1 is useful in predicting the outcome of a baseball game leaves me absolutely speechless. (I spent a minute squinting to convince myself that you really did put a decimal point before the 1... yep, you really did.)
The fact that you may have "used .5 or .6 before" doesn't say anything about the utility of that level of correlation. When you say they "aren't necessarily high", that implies that they might be high. I can't imagine what trustworthy utility they might have, especially since you seem to be deploying these things for prediction of future single events.
At this point, I'll just say that I'm really kinda flabbergasted and just let it go at that...
"The only change is that baseball has turned Paige from a second-class citizen to a second-class immortal." - Satchel Paige
Fair-weather ex-diehard
Posted 26 February 2015 - 06:47 AM
All the r-value in a correlational study tells you is the strength of the correlation. An r-value of .37 is NOT all that strong. The r-squared value of .137 tells you that 13.7% of the variance in the y-statistic (response variable) can be explained by the value of the x-variable. Basically, the relationship is complete crap.
What's even worse is that the guy just kinda mentioned in passing that he got that by looking at some group of P's in one year and compared to some group of P's for the next year. From there, he goes on to talk about how he might have found evidence of a "repeatable skill". What he didn't even come close to mentioning is that the weak correlation he found is about those groups from 1 year to the next year, both of which are past events, and doesn't mean anything whatsoever about its utility for predicting future events. Yet some people will read that (including his claim that he might have discovered evidence of a repeatable skill) and they'll innocently think that they can use that as a predictor of future events.
You see evidence of this all the time... like when some people talk about FIP being a predictor of future pitching performance. Of course, it isn't even in the same area code as any decent predictors... but what happened was that people saw some R-value in an article once, and were not given guidance about how to make sense of it (or didn't pay attention if it was provided), and so they end up somehow thinkng that R-values are a points on a 1..100 scale of predictive goodness.
When guys write articles like this for a general sports audience (as opposed to a science or math audience), they have a responsibility to help readers understand how to interpret the numbers. Not only do many (most?) baseball writers not do that, they do way worse than that by pretty much encouraging the reader to come to incorrect conclusions. I think in lots of cases, and maybe most cases, it's because the author doesn't know any better either... so it's the blind leading the blind... but with the blind author not even realizing he's blind, much less admitting it to his readers.
"The only change is that baseball has turned Paige from a second-class citizen to a second-class immortal." - Satchel Paige
HOF
Posted 26 February 2015 - 09:53 AM
All the r-value in a correlational study tells you is the strength of the correlation. An r-value of .37 is NOT all that strong. The r-squared value of .137 tells you that 13.7% of the variance in the y-statistic (response variable) can be explained by the value of the x-variable. Basically, the relationship is complete crap. I think Shack has an idea what was going on there...maybe I missed something, but just using my stats knowledge (and I teach it).
A .37 r-value isn't all that strong although I wouldn't call it complete crap. Meanwhile, the discussion is about whether an r-value of .85 is worthless or not.
The assertion that an R-value of 0.1 is useful in predicting the outcome of a baseball game leaves me absolutely speechless. (I spent a minute squinting to convince myself that you really did put a decimal point before the 1... yep, you really did.)
The best gamblers are only right 55% of the time. A model that can give you a small edge is huge.
Here. Suppose we're playing a game. I pick a person from anywhere around the world and you have to tell me whether it's a male or female. And suppose we play this game 100,000 times for a $1 each.
If you have access to google then you'll presume that every single person is male and therefore will be right roughly 50.4% of the time. That means on average you'll be right 866 times more than the times that you'll be wrong and you'll get a nice sum.
Now suppose that I also tell you the country the person is from. In that case, you can use country demographics to figure out when you should pick males (China, India) and when you should pick females (United States, Europe).
Such a method won't have much predictive power. But it's enough that you'll be right considerably more often and mean you'll pocket considerably more money. You don't necessarily need to explain everything for a model to have value.
Fair-weather ex-diehard
Posted 26 February 2015 - 05:44 PM
A .37 r-value isn't all that strong although I wouldn't call it complete crap. Meanwhile, the discussion is about whether an r-value of .85 is worthless or not.
For the purposes people are inclined to use it (which aren't necessarily the appropriate ones), it can fairly be classified as crap.
Nobody said .85 was worthless... I said it's not "very high"... because it's not "very high" in any meaningful sense... it's only high on the range of possible values between 0 and 1.0, but that's not what matters...if you're gonna use these things as predictors of future events (which is iffy right there, depending various things), then it's more like a C-grade predictor, it's certainly not an A-grade predictor... and it could be a lot worse than C-grade, depending on a couple things....
"The only change is that baseball has turned Paige from a second-class citizen to a second-class immortal." - Satchel Paige
HOF
Posted 27 February 2015 - 11:22 AM
For the purposes people are inclined to use it (which aren't necessarily the appropriate ones), it can fairly be classified as crap.
We're not going to get anywhere at this point. I'll just end this discussion with this article.
http://www.fangraphs...comment-4962609
Can we both agree that this is a great way to butcher statistics?
Fair-weather ex-diehard
Posted 27 February 2015 - 03:10 PM
We're not going to get anywhere at this point. I'll just end this discussion with this article.
http://www.fangraphs...comment-4962609
Can we both agree that this is a great way to butcher statistics?
I thought the purpose of that last chart was the icing on the cake... the cherry on the sundae... etc, etc
Can we agree that, given how people (including authors of stat-based articles) routinely misuse correlation (probably because they simply don't understand what it is and isn't), things would be immensely less-bad if they would use covariance (R-squared) instead of correlation (R )?
"The only change is that baseball has turned Paige from a second-class citizen to a second-class immortal." - Satchel Paige
HOF
Posted 27 February 2015 - 04:02 PM
http://ourgame.mlblo...ching-distance/
So many entirely minuscule things about baseball still to learn every day.
Fair-weather ex-diehard
Posted 27 February 2015 - 04:10 PM
http://ourgame.mlblo...ching-distance/
So many entirely minuscule things about baseball still to learn every day.
I knew the effective pitching distance was pretty much the same between 50' (front foot) and 60'6" (back foot)... but I didn't know what came before that... including that they banned the raised-leg windup in 1885...
"The only change is that baseball has turned Paige from a second-class citizen to a second-class immortal." - Satchel Paige
HOF
Posted 28 February 2015 - 02:24 PM
Fair-weather ex-diehard
Posted 28 February 2015 - 02:38 PM
I didn't know that Carlos Pena had a new job... but he seems to be very good at it...
http://hardballtalk....ctive-velocity/
"The only change is that baseball has turned Paige from a second-class citizen to a second-class immortal." - Satchel Paige
Owner
Posted 01 March 2015 - 12:02 AM
CBS Sports: Spring training positional battles to watch: AL West
http://www.cbssports...o-watch-al-west
HOF
Posted 02 March 2015 - 12:57 PM
http://grantland.com...dgers-mariners/
Some teams in the top 10 that haven't been top 10 teams in quite some time and not very favorable rankings for the teams that played in the World Series last year.
Fair-weather ex-diehard
Posted 02 March 2015 - 04:07 PM
http://grantland.com...dgers-mariners/
Some teams in the top 10 that haven't been top 10 teams in quite some time and not very favorable rankings for the teams that played in the World Series last year.
Spend a lot and get graded highly... don't spend and drop like a rock...
"The only change is that baseball has turned Paige from a second-class citizen to a second-class immortal." - Satchel Paige
HOF
Posted 02 March 2015 - 04:12 PM
Mariners & Nats will be my World Series..
Not going out on some crazy limb, I realize, but I do think Seattle can win the World Series.
HOF
Posted 02 March 2015 - 04:45 PM
Spend a lot and get graded highly... don't spend and drop like a rock...
It's not that simple. He likes some teams to improve that didn't spend much and doesn't like all the big moves. However, while we both agree that spending a lot is far from always being a good idea, would you not also agree that spending a lot generally helps the team in the short-term?
0 members, 0 guests, 0 anonymous users
