PrOPS and the '08 Cardinals
by ViperLjs on Jan 14, 2009 2:52 AM EST
Email
PrintHow lucky were the Cardinals' hitters last year? There is a way to tell.
There is a stat that exists that shows a player's Predicted OPS, or "PrOPS." PrOPS is a stat referenced on Hardball Times that attempts to accurately gauge a player's performance without the "luck factor" of having balls fall in for hits or not. I don't see this stat referenced very often, and I'm not quite sure why. It seems to be on solid ground statistically, and it could be quite useful. You can read about it from the source here: http://www.hardballtimes.com/main/article/introducing-props/
To take the words of the stat's creator, J.C. Bradbury, PrOPS
"uses batted-ball types (line drive rate, groundball-to-flyball ratio) and a few other things to generate the typical outcome for a player who hits the ball in this manner."
These variables include:
- Line drives per batted ball
- Groundball-to-flyball ratio
- Walk rate
- Hit-by-pitch rate
- Strikeout rate
- Home run rate
- Home park of the player
It's basically the same kind of measure as Fielding Independent Pitching (FIP), only for hitters. He goes on to say about it's significance,
"There is a highly statistically significant relationship…between a player’s over/under performance and his decline/improvement. And the greater the the deviation between PrOPS and OPS, the larger the reversion is the following season. For every 0.01 increase/decrease in a player’s over/under performance, his OPS is likely to fall/rise by 0.008 the following season. For example, a player with an OPS 10 “points” above his PrOPS, can expect his OPS to fall by eight points in the following season. That is quite a reversion."
So any player that has a positive difference for a season is probably going to be worse the next year, while a negative difference would indicate they will probably be better.
Now lets take a look at how the Cardinals performed last season:
Predicted Actual Difference
Yadier Molina C .761 .740 -0.021
Albert Pujols 1B 1.076 1.114 .038
Adam Kennedy 2B .714 .692 -0.022
Khalil Greene SS .706 .599 -0.106
Troy Glaus 3B .890 .856 -0.035
Skip Schumaker LF .780 .765 -0.015
Rick Ankiel CF .878 .843 -0.035
Ryan Ludwick RF .969 .966 -0.003
Joe Mather OF .881 .780 -0.101
Chris Duncan LF .809 .711 -0.098
Brendan Ryan INF .665 .596 -0.069
As you can see, only one player who is likely to get a significant number of AB's next year overperformed their predicted OPS, and that is none other than Albert Pujols. Judging from Bradbury's methodology, we can expect Pujols to have a 1.083 OPS next season.
Every other player under-performed their predicted OPS, meaning they are likely to have better seasons next year than last, or at least not regress significantly. Given that the Cardinals were 4th in the NL in runs scored in '08, this data is cause for fans to be optimistic about the offense, at least, for next year.
Some interesting conclusions from the data:
Ludwick's season was not a fluke.
Ludwick's PrOPS matched very closely with his actual OPS, with only a -0.003 difference. Luddy's OPS basically matched perfectly with his performance, so it's more likely than not his success will continue.
Khalil Greene should be better.
Greene's difference of -0.106 says that he was quite unlucky last season. Combined with his career numbers, this looks to be more evidence that we can expect a bounceback from Greene next year.
Rick Ankiel has a ton of potential.
His excellent debut season was actually supposed to be better. Given that it was only his second full season of hitting period, let alone in the majors, and he was injured during a part of it, I'd say we can expect great things from Rick as he develops.
Joe Mather deserves playing time.
While his rookie campaign went well, he was unlucky. Judging from the formula, Mather should have around an .860 OPS next season.
One should not simply look at this statistic and take it as Gospel. Career performance, injuries, age, ballpark, and other variables are very important in determining a player's likely future performance; however, PrOPS is a useful tool in seeing how luck has affected past seasons, and what to expect in the future.
31 comments
|
2 recs |
Do you like this story?
Comments
Wow, we seem to be in good shape next year
I would also love to see some PrOPS for minor-leaguers last year, specifically Freese, Rasmus and Jones.
I would love that, too.
Sadly they are not provided. Maybe we can lobby the Hardball Times to start keeping track of the minors……
I think that
you are drawing some incorrect conclusions from the data. prOPS probably has its uses, but it’s not better at predicting what a player will do than something like a simple Marcel projection.
For instance, Ludwick will have to sustain a LD rate that nobody sustains or make some strides in other areas for his prOPS to stay where it is.
Exactly.
Lduwick’s season wasn’t a fluke in that he was hitting the ball well and the results were in line with that. That should be distinguished from a sustainable performance, which it also likely wasn’t. Ludwick had like the 3rd highest LD% in the majors last year — that’s not likely to happen again just due to random variation.
I should have made it clearer
that those conclusions were based on the assumption that the PrOPS data was valid, and a good indicator of future success. That’s why I put the last paragraph in. Also, Ludwick’s season wasn’t a fluke, and I never said that he will continue to have a .960 OPS; I simply said he is more likely than not to continue to have success. You can’t simply rule out an excellent season based upon the fact that he was hitting the ball too well over it’s course; that it clearly skill and something influenced by the batter. While he is not likely to have another season that good, it does indicate that he possesses a solid skill set and is able to hit the ball hard consistently; that should be taken as an indicator of future success in my view.
yeah
sure he might not have quite as high of a line drive rate, but even if he’s close he is very dangerous.
by Cards Fan in Chitown on Jan 14, 2009 3:09 PM EST up reply actions
Yes.
The degree of his continued success is what I’d question. You’d expect his LD% to regress toward MLB average. He’ll still have an above average LD% but will it be the 3rd highest again next year? Probably not. Ludwick should be a solid contributor moving forward.
Fair enough
and I didn’t really pay attention to the last paragraph like I should’ve. Mainly this
So any player that has a positive difference for a season is probably going to be worse the next year, while a negative difference would indicate they will probably be better.
and this
Every other player under-performed their predicted OPS, meaning they are likely to have better seasons next year than last, or at least not regress significantly.
jumped out at me.
My point was that prOPS doesn’t really add anything to the discussion about what they’ll do next year beyond any of the decent projection systems. But it is interesting when looking at a past season.
It does add something though.
Say the data on Ludwick went the other way, and he had a very lucky season. He wasn’t hitting the ball all that hard; hits just seemed to be coming down in the right spots. Marcels does not distinguish between a fluky season and a non-fluky one; all are treated with equal weight. His projection would still be for an .879 OPS. If his season had truly been that flukish, then I would say that we shouldn’t expect his success to continue next year despite what the projections may be. With small variations and sample sizes it’s not very useful, but over the course of a season it tell you how much the player’s overall line (which the projection systems use) was influenced by luck.
BTW, I’m not really sure about the methodology of the other projection systems; do the Bill James, CHONE, ZIPS, and PECOTA projections take batted-ball types or some kind of “luck” factor into account?
Presumably
the smaller the sample size the more useful it is, correct? As the sample grows, prOPS and OPS should converge if it’s doing what it’s supposed to be doing.
However the good systems take much more into account that just one year of data, and I know that Dan S at least has talked about hitter babip and on-contact rates. It’s probably better at predicting OPS than the previous years’ OPS, but PECOTA/ZiPS/whatever will be better still.
In your link a few posts down Bradbury says this: PrOPS isn’t a forecasting system. It’s backwards-looking tool. It might have some use in projecting, but I haven’t put much effort in to applying it in this way.
You're right
Its not meant to be a prediction tool by itself. All it tells you is whether or not the player is likely to be better or worse the next year. But used in conjunction with the real projection systems, I think it does have some predictive value.
Bradbury says about it’s predictive usage:
“PrOPS is not a stand-alone projection tool. You should not look only at a player’s PrOPS and assume it’s exactly what the player should be doing. When I look at it, I also consider the player’s recent hitting history, injuries, aging, and all that other stuff we sometimes use to evaluate hitters. But when I see a player have a career year, and his PrOPS don’t show it, I start to get suspicious.”
Basically he sums up my position with this comment:
“It’s not an air-tight projection system, but there seems to be some information there.”
I can agree with
all of that, and there’s information beyond OPS. Interesting information at that.
Until someone figures out a way to combine it with something that’s already better to predict future seasons, there’s really no point in using it to look forward, besides the fact that it’s interesting. All that it’s going to do is take 3 or more years of information, and make Pujols’ batted balls look more like Izturis’. Until someone demonstrates that it improves on the accuracy of the best forecasting systems, it’s a novelty stat, or at least, poorly named.
I agree.
The biggest change Bradbury should make is to differentiate between different types of hitters’ batted balls. A power hitter’s line drives and fly balls should not count the same as a slap hitter’s. That alone would do wonders for its predictive value, IMO.
Ludwick wasn't exactly a fluke
But nobody has LD rates like he did. Either we’ve seen the most spectacular player in the history of baseball have his first good season, or he was just simply in some kind of amazing zone. I’m hoping for the first, banking on the second. I think he can and will still be a good player this year and from here on out, but I don’t think he’ll be quite as good as he was this season.
"Your Holiness, I'm Joseph Medwick. I, too, used to be a Cardinal."-Joe Medwick, to Pope Pius XII.
by redbirdnation8206 on Jan 14, 2009 8:52 AM EST reply actions
I still find the data interesting FWIW...
it looks like we can expect substantial improvement from Mather, Greene, Duncan and Ryan…though the latter three need substantial improvements to justify a roster spot. It’d be interesting to compare the batted ball types to look for outliers (like Ludwick’s LD %) and see which players can be expected to improve/decline in that area, then incorporate that into their prOPS number.
Sometimes I wonder,
"Why is that frisbee getting bigger?"
...and then it hits me!!
was albert lucky....
or is he just that good because he hits ball so hard and places the ball in the right spots that it looks like he is lucky. Are there any trends in his career numbers that would prove it was luck or just another stat to show how good he is?
Is it weird that I would rather the payroll be more like the Marlins than the Yankees?
He's outperformed
his prOPS four out of the last five years. I don’t really know how prOPS is set up, and I do know that Bradbury is a smart guy, but IIRC it assumes that all LD are created equal. Same thing with LD+.120 or whatever people like to say about hitter babip.
Luck of the Irish.
You guys knew Albert was Irish, right?
This is pretty interesting stuff, but I’m glad some commentors are urging caution. Does anybody happen to have an idea of why Mather might have been so unlucky?
It would make sense to me...
That guys like Albert would tend to outperform their props a bit (an Albert fly ball is more likely to turn into a double in the gap than a Miles fly ball, because on average it will be a longer, harder-hit fly ball). Also, you would expect speed-burners to get a few more hits than expected by prOPS on ground balls. Guys with no speed and little power (Molina, cough), should be expected to underperform their prOPS a bit, IMO.
by mikedallas45 on Jan 14, 2009 11:08 AM EST up reply actions
Also,
Greene has a history of underperforming his prOPS by a bit, but nowhere near as much as .106, so he does indeed look like a good bounceback candidate:
2004 Greene Khalil .783 .795 .012
2005 Greene Khalil .772 .727 -0.045
2006 Greene Khalil .791 .747 -0.045
2007 Greene Khalil .775 .759 -0.016
2008 Greene Khalil .706 .599 -0.106
by mikedallas45 on Jan 14, 2009 11:15 AM EST up reply actions
Excellent work, Viper
you’re right — we should be doing more w/ this stat. This is exactly what a fanpost should be.
I'm not sure where you got this info
But do they have any predictive conclusions based on past data? For example, do certain types of hitters consistently outperform their PrOPS?
"I just wish that the late Harry Caray were still around so I could hear him mispronounce 'Kosuke Fukudome' every fukun' night" -- Dennis Miller
No, they just list them.
I get it from the Hardball Times stat library. Go here: http://www.hardballtimes.com/thtstats/main/?view=props&league_filter[]=2
From there you can isolate years, positions, teams, etc; you just have to draw your own conclusions.
Bradbury did create some conclusions in his introductory article, which is listed near the top of my fanpost. There are some other articles floating around on the net by him that explain more, one of which can be found here: http://www.sabernomics.com/sabernomics/index.php/2007/03/props-questions/.
Interestingly, he dispels the notion that speed significantly influences the “luck” factor.
interesting to see
that Greene and Mather will likey do better in ’09, and that Ludwick and Ankiel is not a fluke
by Cards Fan in Chitown on Jan 14, 2009 3:10 PM EST reply actions
Big ol' FWIW: I'm growing skeptical of Bradbury
FIRST AND FOREMOST: thanks for the great fanpost! I hadn’t looked into PrOPS myself before reading it.
http://www.sabernomics.com/sabernomics/index.php/2008/12/ibanez-versus-burrell/
…I don’t have time to walk through individual estimates at this time…When my next book is published the changes to the model will be explained in even much greater detail than I they were in my previous book.
The SABR community seems to disagree, so…buy my books if you want to know why I’m an outlier?
While I personally disagree with his conclusions, it’s the (lack of) response to criticism and apparent proprietary leaning which bug me. It’s quite a sense of entitlement open source stats can give a person, though, and I think we’re beyond fortunate to have places like Fangraphs collecting it all.
That has nothing to do with PrOPS, though, just my dubious ramblings.
Over at The Book Blog we’ve got Tangotiger, Rally (CHONE), and MGL weighing in on PrOPS itself:
Rally: I don’t like Props much, because it isn’t predictive. In a very basic way it could be useful, in that a player with a higher props than real production will tend to improve, but we have tested it around here and it’s less predictive than a Marcel. Ground balls, Flyballs, and Linedrives among different players are not created equal, and that is the flaw. (emphasis mine)
Tangotiger: Right, I agree with Rally. It’s too dangerous to treat a David Eckstein and Ryan Howard line drive as the same thing. This is another case of where doing too much is (likely) worse than doing nothing at all.
If that’s any indication it looks like it’s still cool as a post-analysis, as you collected very well, but it doesn’t look like it’s necessarily indicative of future performance. Another more recent stat to consider might be “new” xBABIP. I think it looks promising, but we’ll need the model :)
xBABIP seems very much like PrOPS
It can let you know how “lucky” someone was in any given year, but they don’t necessarily predict anything.
by vivaelpujols on Jan 14, 2009 4:41 PM EST up reply actions
All good points.
I don’t think that PrOPS by itself is a good projection tool; I think that it can be useful in conjunction with real projection systems to tune the forecasts. Seeing if a very good/bad season was “real” or not could be very helpful in predicting the future performance of the player. It would be much better if the batted balls were adjusted for the player’s power and swing type, like Tangotiger says.
I haven’t looked into xBABIP at all; I’ll have to take a look at it.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
In Association With
Big League Stew - MLB - Yahoo! Sports
- Have you criticized Hanley Ramirez on Twitter? Your name may be on his bat!
- U.S. senator cries foul over Nationals plans to take back park from Phillies fans
- Photo: Comerica Park readies itself for Hockeytown
- Podcast: Dirk Hayhurst talks about Italy, the reality of the big leagues and his new book
- Dennis ‘Oil Can’ Boyd claims he used cocaine before a majority of his games
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
