Your 2015 Offseason Plan

[cabiness42]

The park adjustment that is made is crude and is based on adding or subtracting runs rather than adjusting for each offensive component. This ends up being okay because the stat is based on linear weights, so even though it won't factor that a guy has, say, a couple more triples than he normally would in a neutral park, it DOES factor in the runs that those triples represent.

 

From FG:

 

 

 

So for example, in 2014, a triple was worth ~1.65 runs on average, so every triple that someone hits gets them that much credit. All of that adds up to a player's Runs Created, and THEN an overall run scoring coefficient gets thrown on. For the Cell, it was 104 this year, which means that the park si expected to produce 4% more runs, so everyone's performances there get a 4% hit.

 

So the TL;DR answer is this: all offensive events get converted to runs first, then the park factors are added. This works out because each type of event is naturally weighted, by runs, in the system to begin with.

 

Seems like it should go in the reverse order. Parks can definitely affect types of hits differently, and can affect left/right handed hitters differently. Triples are clearly harder to come by in the Cell than at Comerica, and homers are clearly much easier to come by at Yankee Stadium for lefties than righties. If we are just blanketly saying that the Cell inflates offense by 4%, then that doesn't give me much confidence in that system. Seems pretty lazy to me.

[oneofthemikes]

I would love to see the Sox go get Sandoval and Markakis. A top half of a line up of :

 

Eaton

Markakis

Abreu

Garcia

Sandoval

 

would be pretty scary. Markakis would see great pitches all the time with Abreu behind him. Hopefully Sandoval would have a nice spike in his power numbers moving to the Cell. Then Gillaspie could either be traded or used as a DH and to spell Sandoval at 3B on occasion.

 

Not sure what I would like to see them do in terms of a RH starter, but I think I'd like to see them go get Sergio Romo and Zach Duke for the pen.

[Chicago White Sox]

Seems like it should go in the reverse order. Parks can definitely affect types of hits differently, and can affect left/right handed hitters differently. Triples are clearly harder to come by in the Cell than at Comerica, and homers are clearly much easier to come by at Yankee Stadium for lefties than righties. If we are just blanketly saying that the Cell inflates offense by 4%, then that doesn't give me much confidence in that system. Seems pretty lazy to me.

See, this was my exact concern. Do slap hitters like Juan Pierre & Scott Podsednik gain as much when playing at the Cell as guys who tend to put more loft on the ball? I would guess not, but it sounds like wRC+ spreads the benefit to all players equally.

 

And to layer in more complexity, there is also some seasonality/variability to park effects. Tempetature, wind, etc. aren't static, meaning it's difficult to measure how much incremental run production in a given period of time was due to favorable weather vs. a team simply having a hot streak.

[SCCWS]

WARNING: OPS Rant

 

Omg why are people still citing OPS and treating it like it's a proxy for a player's overall value? Not only is OPS bad, but everyone is still tuned to the standards of the steroid era, when every outfielder had to be 800+. This is simply NOT representative of baseball today. Let's use numbers that (a) properly weight the constantly changing value relationship between OBP and SLG, and (b) are actually adjusted to the offensive era that we are in.

 

2014 Jason Heyward: 110 wRC+, +24.1 UZR, 5.4fWAR.

 

This is a guy who hit 10% better than league average and played elite defense in 2014 -- in fact, almost as elite as you can possibly be. The result was an all-star caliber player. You would NOT expect the +25 UZR to be sustainable (though he's done it before in 2012), but since he's always been between +10 and +25, it's reasonable to expect a +10-15. And since the wRC+ is actually a touch BELOW his career of 117, there's no reason to believe he won't be sitting at 110-115 next year. And that would be an awesome season and he would be the best position player on our team, save Abreu.

 

Please stop citing OPS, it's problematic. It gives us an inaccurate picture of the truth and we have better tools that are just as easy to find and utilize.

 

Ok, I'm done. I'm sorry, GreenSox. I know I'm a jerk.

 

The point about the one-year of control is huge, but for like three pages here it looks like everyone is talking about a trade being contingent on an extension. The part that makes it interesting is that his value would conceivably be lower than one might think simply because of the lack of control. I'm not sure how it would be affected if all involved parties knew that a handshake extension agreement was being made.

 

 

Cancel the rant. Actually the Top 10 MLB players in WRC are the same top 10 in OPS. Guys like Eaton, DV, amd Conor all fall into the same range in both categories. I am sure there are some odd instances both ways but quick look tells you both stats are very similiar.

[Eminor3rd]

Seems like it should go in the reverse order. Parks can definitely affect types of hits differently, and can affect left/right handed hitters differently. Triples are clearly harder to come by in the Cell than at Comerica, and homers are clearly much easier to come by at Yankee Stadium for lefties than righties. If we are just blanketly saying that the Cell inflates offense by 4%, then that doesn't give me much confidence in that system. Seems pretty lazy to me.

 

But you have to remember that the whole system sits on linear weights, so any blanket run change you make automatically affects all events in proportion to their value.

 

Also, remember what question we're trying to answer. wRC+ is the answer to: How valuable was player x's overall contribution, at the plate, in comparison to those of his peers? This only works if you take steps to equalize the difference between things like triples/stolen bases and homers/walks. A common denominator is essential to this; you wouldn't want it any other way.

 

If you were trying to decide how a player will FIT in a particular ballpark, or how his contributions will translate, you definitely need to dig into the park factors for different offensive events. But if you want to measure how VALUABLE a player was in comparison to another, who might have played in a different park or league, you want to reduce your measurement as much as possible.

 

EDIT: I feel like I'm not explaining myself well. Maybe someone else can try.

Edited October 27, 2014 by Eminor3rd

[Eminor3rd]

Cancel the rant. Actually the Top 10 MLB players in WRC are the same top 10 in OPS. Guys like Eaton, DV, amd Conor all fall into the same range in both categories. I am sure there are some odd instances both ways but quick look tells you both stats are very similiar.

 

OPS has no adjustment for context and inaccurately assumes that a point of OBP is equally as valuable as a point of SLG.

 

Case in point: GreenSox cited Heyward's OPS to display that he was a mediocre hitter in 2014, when in reality he was an above-average hitter over that span.

Edited October 27, 2014 by Eminor3rd

[witesoxfan]

Cancel the rant. Actually the Top 10 MLB players in WRC are the same top 10 in OPS. Guys like Eaton, DV, amd Conor all fall into the same range in both categories. I am sure there are some odd instances both ways but quick look tells you both stats are very similiar.

 

But what does an OPS of .735 tell you? To me, a .735 OPS only tells me that his slugging and on base percentage added up to .735. 10 years ago, a .735 OPS would have been below average. In today's environment, a .735 OPS CAN be fairly good.

 

OPS is also a more severely flawed statistic. It's oft-used, but if a team had an OBP of 1.000, the game would never end because they'd never stop getting on base. The same cannot be said for a 1.000 SLG, unless they were all singles, in which case their OBP would also be 1.000, thus verifying the results. Through simple deduction we reason that OBP is more important than SLG.

[cabiness42]

But you have to remember that the whole system sits on linear weights, so any blanket run change you make automatically affects all events in proportion to their value.

 

Also, remember what question we're trying to answer. wRC+ is the answer to: How valuable was player x's overall contribution, at the plate, in comparison to those of his peers? This only works if you take steps to equalize the difference between things like triples/stolen bases and homers/walks. A common denominator is essential to this; you wouldn't want it any other way.

 

If you were trying to decide how a player will FIT in a particular ballpark, or how his contributions will translate, you definitely need to dig into the park factors for different offensive events. But if you want to measure how VALUABLE a player was in comparison to another, who might have played in a different park or league, you want to reduce your measurement as much as possible.

 

EDIT: I feel like I'm not explaining myself well. Maybe someone else can try.

 

I get what you're trying to say, but I don't agree with it. A right-handed Yankee getting the same park adjustment as a left-handed Yankee is getting f***ed right up the ass.

[cabiness42]

OPS has no adjustment for context and inaccurately assumes that a point of OBP is equally as valuable as a point of SLG.

 

Case in point: GreenSox cited Heyward's OPS to display that he was a mediocre hitter in 2014, when in reality he was an above-average hitter over that span.

 

I agree that OPS is flawed because OBP is a bit more valuable than SLG, but I also think that is context dependent on how you are trying to fit a guy in with the rest of the team. If you already have a team with decent OBP but lacking pop, you might favor SLG even though OBP is generally move valuable.

 

 

[shysocks]

You can always use wOBA. Corrects the OBP/SLG imbalance problem of OPS, without the park adjustment. It lacks my favorite feature and main advantage of wRC+, which is the scaling to 100.

[witesoxfan]

I get what you're trying to say, but I don't agree with it. A right-handed Yankee getting the same park adjustment as a left-handed Yankee is getting f***ed right up the ass.

 

I follow your logic and understand what you're saying. I was thinking they did have the breakdown somewhere of park factors for both RH's and LH's. If it's not included, I agree they should do more to neutralize for that as well. Maybe it's something one of us can ask to the brains over at FGs to see if they incorporate that into their statistic or not.

 

I agree that OPS is flawed because OBP is a bit more valuable than SLG, but I also think that is context dependent on how you are trying to fit a guy in with the rest of the team. If you already have a team with decent OBP but lacking pop, you might favor SLG even though OBP is generally move valuable.

 

Right, but that really doesn't have much to do with OPS itself, just general roster construction and a way of favoring what type of player you're looking for based on that criteria.

 

I'm not saying OPS is wrong, just that it's something I try to not to reference on a regular basis because it is a flawed statistic.

[cabiness42]

You can always use wOBA. Corrects the OBP/SLG imbalance problem of OPS, without the park adjustment. It lacks my favorite feature and main advantage of wRC+, which is the scaling to 100.

 

Could you not invent wOBA+ and scale it to 100?

[Eminor3rd]

I get what you're trying to say, but I don't agree with it. A right-handed Yankee getting the same park adjustment as a left-handed Yankee is getting f***ed right up the ass.

 

Maybe by a percentage point or two, and really only if you're a homerun hitter. here's the table with handedness park factors: http://www.fangraphs.com/guts.aspx?type=pf...amp;season=2014

 

Regardless, it's substantially MORE accurate than OPS and OPS+. And if you think the park factors throw it off, you can always cite wOBA, like shysocks said. But even then, you're guessing as to the adjustment that's proper. In cases of extreme handedness advantages, such as homers in Yankee stadium, the overall park adjustment is still much closer to the truth than any other number out there. I could see it being an issue if there was some park where any particular offensive event was substantially above average from one side and substantially below average from the other, but like in Yankee Stadium, it's homer-happy from both sides, it's just extra homer-happy from the left side.

Edited October 27, 2014 by Eminor3rd

[Dick Allen]

I follow your logic and understand what you're saying. I was thinking they did have the breakdown somewhere of park factors for both RH's and LH's. If it's not included, I agree they should do more to neutralize for that as well. Maybe it's something one of us can ask to the brains over at FGs to see if they incorporate that into their statistic or not.

 

 

 

Right, but that really doesn't have much to do with OPS itself, just general roster construction and a way of favoring what type of player you're looking for based on that criteria.

 

I'm not saying OPS is wrong, just that it's something I try to not to reference on a regular basis because it is a flawed statistic.

You can make a case for every statistic being a flawed statistic. One thing I really like about OPS isn't necessarily the number, although if you are getting into the low 700s, obviously, there is diminishing returns. But it can also be used as a gauge as to what direction a hitter is heading.

Edited October 27, 2014 by Dick Allen

[witesoxfan]

You can make a case for every statistic being a flawed statistic.

 

I agree, and I don't think it'd be that hard to do so. At times, you just have to find the least flawed statistics and use more than one of them to prove a point.

[Eminor3rd]

I agree that OPS is flawed because OBP is a bit more valuable than SLG, but I also think that is context dependent on how you are trying to fit a guy in with the rest of the team. If you already have a team with decent OBP but lacking pop, you might favor SLG even though OBP is generally move valuable.

 

Also, I'm not sure OBP points ARE more valuable than SLG points at this juncture. At the very least, they're much closer. Which is part of the issue.

 

Check out this table, noting that run values for HR have steadily increased since the steroid era, whereas run values for singles have decreased: http://www.fangraphs.com/guts.aspx?type=cn

[Dick Allen]

I agree, and I don't think it'd be that hard to do so. At times, you just have to find the least flawed statistics and use more than one of them to prove a point.

One issue I have is with ballpark adjustments. IMO, you are just taking someone's word those adjustments are accurate, and as been pointed out, I think that varies greatly from player to player.

 

Jim Thome had no problem with Target Field. Target Field freaks Joe Mauer out.

Edited October 27, 2014 by Dick Allen

[Eminor3rd]

You can make a case for every statistic being a flawed statistic. One thing I really like about OPS isn't necessarily the number, although if you are getting into the low 700s, obviously, there is diminishing returns. But it can also be used as a gauge as to what direction a hitter is heading.

 

But you can do the same thing, but better, with wOBA or wRC+.

 

The thing I hate so much about OPS is that it's simply obsolete. It actually has NO use anymore. Like even RBI is a better stat because it still answers a simple question better than any other stat. There's no "sabermetric equivalent" of RBI. The SABR problem with RBI wasn't that it was inaccurate (it's extremely accurate, in fact), but rather that it was being used as a proxy to answer the wrong questions (i.e. who is the best hitter?).

 

Analogy: wRC+ is an ATM card, OPS is a checkbook, RBI is coins. The checkbook is completely obsolete compared to the ATM card; the checkbook is a hindrance. But while the ATM is much more useful than coins, there are still a handful of situations where you need coins instead.

Edited October 27, 2014 by Eminor3rd

[Eminor3rd]

One issue I have is with ballpark adjustments. IMO, you are just taking someone's word those adjustments are accurate, and as been pointed out, I think that varies greatly from player to player.

 

Jim Thome had no problem with Target Field. Target Field freaks Joe Mauer out.

 

Well, in terms of their on-field performance, this is exactly why we NEED ballpark adjustments. Both players turn out fine -- homers are hard to hit there, so they are worth more for both players. Thome gets extra credit for being a monster that no park can contain, and Mauer gets padded for not being able to hit as many out.

 

As far as the mental aspect goes: that's one thing that statistics cannot account for across the board. You'll always have to make that subjective adjustment, regardless of whether or not a park adjustment is applied statistically.

[shysocks]

Could you not invent wOBA+ and scale it to 100?

You sure could, and somebody probably has. I'm not aware of where to find it though, so looking up the league average wOBA is just an extra step. And I think it's necessary to have some type of park factor. Even if the method isn't perfect, penalizing hitters in Coors and rewarding those in Seattle is helpful.

[Dick Allen]

Someday, just like they told us the old stats were stupid, we will be told all of these are dumb too.

 

 

[Dick Allen]

You sure could, and somebody probably has. I'm not aware of where to find it though, so looking up the league average wOBA is just an extra step. And I think it's necessary to have some type of park factor. Even if the method isn't perfect, penalizing hitters in Coors and rewarding those in Seattle is helpful.

I agree there is some kind of park factor. It just isn't the same for everyone.

 

When they did the renovation at USCF and the fence was moved in about 15-20 feet down the lines, I think the first year they kept track and IIRC there were 4 home runs that wouldn't have been home runs if they had kept the same dimensions. It is easy to go overboard on the factor. Again, I think it is more mental that physical. I believe hitters change their approach a lot based on the ballpark. Specifically their home park. The reality is, guys like Thome and Dunn and Thomas didn't have a huge drop off of their home run rates when they played their home games for different teams.

Edited October 27, 2014 by Dick Allen

[Eminor3rd]

Someday, just like they told us the old stats were stupid, we will be told all of these are dumb too.

 

Always a by-product of progress. The key is being more interested in the truth than you are in the metric.

[Jerksticks]

But you can do the same thing, but better, with wOBA or wRC+.

 

The thing I hate so much about OPS is that it's simply obsolete. It actually has NO use anymore. Like even RBI is a better stat because it still answers a simple question better than any other stat. There's no "sabermetric equivalent" of RBI. The SABR problem with RBI wasn't that it was inaccurate (it's extremely accurate, in fact), but rather that it was being used as a proxy to answer the wrong questions (i.e. who is the best hitter?).

 

Analogy: wRC+ is an ATM card, OPS is a checkbook, RBI is coins. The checkbook is completely obsolete compared to the ATM card; the checkbook is a hindrance. But while the ATM is much more useful than coins, there are still a handful of situations where you need coins instead.

 

I like your point but I disagree with the analogy. The checkbook is the single most important thing to American business

[Eminor3rd]

I like your point but I disagree with the analogy. The checkbook is the single most important thing to American business

 

Haha, it didn't feel like a bulletproof analogy even as I was typing it.