Offseason rumors/discussion thread

[fatbeard]

I dont see how anyone can argue over wins above replacement, since replacement is a variable. A replacement might hit a walk off grand slam on sunday night baseball.

If only the people who actually invented the statistic explained how they come up with the replacement value...

https://www.fangraphs.com/blogs/unifying-replacement-level/

 

[anotheridiot]

If only the people who actually invented the statistic explained how they come up with the replacement value...

https://www.fangraphs.com/blogs/unifying-replacement-level/

Now you are just trying to confuse me.

Actually, we know it isn’t always accurate because depending on your source — FanGraphs or Baseball-reference.com — you can get wildly different WAR scores… For example:

Does (Jack) Morris, in fact, belong in the Hall of Fame? No, he doesn’t, according to baseball-reference.com, which gives him a WAR score of 39.3, tied for 145th all time among pitchers. Maybe he does, according to FanGraphs, which gives him a 56.9 WAR, 75th all time.

 

[CSF77]

Now you are just trying to confuse me.

Actually, we know it isn’t always accurate because depending on your source — FanGraphs or Baseball-reference.com — you can get wildly different WAR scores… For example:

Does (Jack) Morris, in fact, belong in the Hall of Fame? No, he doesn’t, according to baseball-reference.com, which gives him a WAR score of 39.3, tied for 145th all time among pitchers. Maybe he does, according to FanGraphs, which gives him a 56.9 WAR, 75th all time.

bWAR and fWAR use diffrent equations that give diffrent values.

Baseball ref uses RA9 in their equation. Fangraphs uses FIP

 

[beckdawg]

bWAR and fWAR use diffrent equations that give diffrent values.

Baseball ref uses RA9 in their equation. Fangraphs uses FIP

I believer there's also difference with regard to fielding. I know fWAR uses UZR. I don't know off hand how bWAR figures it but i know it used to give players like Darwin Barney rather ridiculous values based on purely defense. They may have changed it since then but that's when I moved to be a fWAR fan.

 

[fatbeard]

I believer there's also difference with regard to fielding. I know fWAR uses UZR. I don't know off hand how bWAR figures it but i know it used to give players like Darwin Barney rather ridiculous values based on purely defense. They may have changed it since then but that's when I moved to be a fWAR fan.

I believe bWAR uses DRS for their defensive calculations.

 

[beckdawg]

I believe bWAR uses DRS for their defensive calculations.

Could be. His DRS in 2012 was a lot higher than his UZR. I think another issue is they don't seem to use the same scales. For example, they list 5+ bWAR as All-star level. For fangraphs 4-5 is all-star level and 5-6 is superstar and 6+ is MVP.

 

[CSF77]

I believer there's also difference with regard to fielding. I know fWAR uses UZR. I don't know off hand how bWAR figures it but i know it used to give players like Darwin Barney rather ridiculous values based on purely defense. They may have changed it since then but that's when I moved to be a fWAR fan.

The formula
For position players: (The number of runs above average a player is worth in his batting, baserunning and fielding + adjustment for position + adjustment for league + the number of runs provided by a replacement-level player) / runs per win

 

[beckdawg]

The formula
For position players: (The number of runs above average a player is worth in his batting, baserunning and fielding + adjustment for position + adjustment for league + the number of runs provided by a replacement-level player) / runs per win

My point is that 2012 darwin barney is vastly different in terms of bWAR vs fWAR. He was 4.6 bWAR which is basically all-star level where he was 1.9 fWAR which is below league average. The only reason that makes any sense is they are valuing defense far more heavily than fWAR as he was a 75 wRC+.

 

[Castor76]

Now you are just trying to confuse me.

Actually, we know it isn’t always accurate because depending on your source — FanGraphs or Baseball-reference.com — you can get wildly different WAR scores… For example:

Does (Jack) Morris, in fact, belong in the Hall of Fame? No, he doesn’t, according to baseball-reference.com, which gives him a WAR score of 39.3, tied for 145th all time among pitchers. Maybe he does, according to FanGraphs, which gives him a 56.9 WAR, 75th all time.

Which is why I think WAR is a flawed stat and never rely on it.

 

[fatbeard]

Which is why I think WAR is a flawed stat and never rely on it.

You realize that the poster in question here is comparing two different WAR stats with two completely different calculations, and that's why there's a difference...right?

 

[Castor76]

You realize that the poster in question here is comparing two different WAR stats with two completely different calculations, and that's why there's a difference...right?

Yes, and that's part of the flaw I find with it. BA, SLG, OPS, et al are the same because a singular formula used by everyone. You can explain any of those a 9 year old and they get it. Algorithms can be be made to whatever you want it to be.

 

[beckdawg]

Yes, and that's part of the flaw I find with it. BA, SLG, OPS, et al are the same because a singular formula used by everyone. You can explain any of those a 9 year old and they get it. Algorithms can be be made to whatever you want it to be.

You can explain simple addition to a 3rd grader but do you want a third grader doing your taxes? As I said before I don't really follow bWAR because I don't like the way it's implemented. That being said, fWAR really isn't that complicated unless you want to understand the minute details of how every individual piece is calculated. Simply stated you get a runs value for your offense your defense and your base running. Every 10 runs is a win. Like I said if you want to dig down into exactly how UZR works it can get complicated but that's because measuring defense is complicated. Offense isn't really that complicated. It's based on wOBA which is just a weighted version of OBP.

 

[fatbeard]

Yes, and that's part of the flaw I find with it. BA, SLG, OPS, et al are the same because a singular formula used by everyone. You can explain any of those a 9 year old and they get it. Algorithms can be be made to whatever you want it to be.

That's not a flaw. They are literally two different stats that happen to share part of a name. You might as well complain that "Runs" and "Runs Batted In" are flawed because they both contain "Runs" in the names.

 

[Castor76]

You can explain simple addition to a 3rd grader but do you want a third grader doing your taxes? As I said before I don't really follow bWAR because I don't like the way it's implemented. That being said, fWAR really isn't that complicated unless you want to understand the minute details of how every individual piece is calculated. Simply stated you get a runs value for your offense your defense and your base running. Every 10 runs is a win. Like I said if you want to dig down into exactly how UZR works it can get complicated but that's because measuring defense is complicated. Offense isn't really that complicated. It's based on wOBA which is just a weighted version of OBP.

Except for there's no rhyme or reason to the multipliers used for wOBA. It's an opinion. It's like those Strength of Schedule formulas that use your opponents' SOS within the formula. You have to determine your SOS by using your opponents SOS, but that can't be determined until your SOS is determined, buuuuuuuuuuuuut here's your SOS.

There's no opinion in a lot of numbers. A lot of the saber stats, good and useful as they are, have opinion in those formulas. Doesn't mean they can't be used or they have no value. I just don't put them on an alter.

 

[Castor76]

That's not a flaw. They are literally two different stats that happen to share part of a name. You might as well complain that "Runs" and "Runs Batted In" are flawed because they both contain "Runs" in the names.

The flaw I'm speaking of is that each WAR stat depends upon an algorithm where a human assigns an arbitrary value to a condition. Just like wOBA does.

 

[fatbeard]

The flaw I'm speaking of is that each WAR stat depends upon an algorithm where a human assigns an arbitrary value to a condition. Just like wOBA does.

You're just flat wrong here. Take a gram, for instance. There is no objective definition--free from human minds--of what a gram "is". It's just a label we as a species have decided to apply to an arbitrary and finite amount of mass. It doesn't matter if that amount of mass is just something we all decided to agree on, as long as it is consistently applied we can make reliable comparisons between the masses of different objects using grams as a baseline. In the exact same way, it doesn't matter if values within the algorithm are arbitrary (they're not really, there is a logic behind them); as long as the value is consistent across each calculation (bWAR or fWAR), the results will be consistent. And if the results are consistent, we can reliably use WAR to compare player value, as long as we're sticking with one form of WAR.

 

[Castor76]

You're just flat wrong here. Take a gram, for instance. There is no objective definition--free from human minds--of what a gram "is". It's just a label we as a species have decided to apply to an arbitrary and finite amount of mass. It doesn't matter if that amount of mass is just something we all decided to agree on, as long as it is consistently applied we can make reliable comparisons between the masses of different objects using grams as a baseline. In the exact same way, it doesn't matter if values within the algorithm are arbitrary (they're not really, there is a logic behind them); as long as the value is consistent across each calculation (bWAR or fWAR), the results will be consistent. And if the results are consistent, we can reliably use WAR to compare player value, as long as we're sticking with one form of WAR.

Except there is a defined way of determining a gram in weight. It's the weight of one cubic centimeter of pure water. You are correct in that it's something the rest of us just said, "Sure." and went from there.

Why I can't actually be wrong in my opinion when it comes to WAR, is it's an opinion, just like the values assigned in war. If the person who designed the formulas had different opinions on the various parts, then the pieces put into it would have a different value giving a different results. A bias, even when applied evenly, is still a bias.

My point being, you tell 10 people who have no concept of what baseball is to come up with a formula to represent the percentage rate a which a person comes to the plate at gets safely on base with no out being made through their own action and they come up with what we have already for OBP. Tell those same people to determine what any one player's value over a readily available replacement would be and you'll get 10 formulas.

My real point is this. You tell me a player was a 3.5 WAR, you may have told me he's a good player, but you haven't told me what makes him good. You tell me that same guy has a slash line .290/.360/.480 and a RISP w/ 2 Outs line of .320/.410/.560 and I know I got a guy who's a solid bat, good at getting on base and delivers fairly well when an inning is tight. But according to WAR, Kolten Wong was worth the same as Joey Votto.

 

[CSF77]

Except there is a defined way of determining a gram in weight. It's the weight of one cubic centimeter of pure water. You are correct in that it's something the rest of us just said, "Sure." and went from there.

Why I can't actually be wrong in my opinion when it comes to WAR, is it's an opinion, just like the values assigned in war. If the person who designed the formulas had different opinions on the various parts, then the pieces put into it would have a different value giving a different results. A bias, even when applied evenly, is still a bias.

My point being, you tell 10 people who have no concept of what baseball is to come up with a formula to represent the percentage rate a which a person comes to the plate at gets safely on base with no out being made through their own action and they come up with what we have already for OBP. Tell those same people to determine what any one player's value over a readily available replacement would be and you'll get 10 formulas.

My real point is this. You tell me a player was a 3.5 WAR, you may have told me he's a good player, but you haven't told me what makes him good. You tell me that same guy has a slash line .290/.360/.480 and a RISP w/ 2 Outs line of .320/.410/.560 and I know I got a guy who's a solid bat, good at getting on base and delivers fairly well when an inning is tight. But according to WAR, Kolten Wong was worth the same as Joey Votto.

It is a body of work. You can single out any instance. 1-2 with man on 1st. 2-2 bases empty. and so on. They have splits for that. But players are paid for a body of work. Players are paid for a career. No a single moment. Kerry had a great game but that will never get him into the hall.

 

[Castor76]

It is a body of work. You can single out any instance. 1-2 with man on 1st. 2-2 bases empty. and so on. They have splits for that. But players are paid for a body of work. Players are paid for a career. No a single moment. Kerry had a great game but that will never get him into the hall.

Doesn't change that while WAR can tell you a player is good, but it doesn't tell you what part of their game is making them that kind of good. Wong and Votto had the same WAR last season. It doesn't tell you that Wong was almost all defense and is probably a 6 hitter at best while Votto was even defensively and very easily a heart of the lineup guy. I know Votto's had bigger years and more of them. But for the whole of 2018, that algorithm said those two guys had the same value.

Slash lines aren't single moments. They're the accumulations of particular instances over time. They tell you a lot more about a player.

 

[CSF77]

Doesn't change that while WAR can tell you a player is good, but it doesn't tell you what part of their game is making them that kind of good. Wong and Votto had the same WAR last season. It doesn't tell you that Wong was almost all defense and is probably a 6 hitter at best while Votto was even defensively and very easily a heart of the lineup guy. I know Votto's had bigger years and more of them. But for the whole of 2018, that algorithm said those two guys had the same value.

Slash lines aren't single moments. They're the accumulations of particular instances over time. They tell you a lot more about a player.

It is not supposed to. It is a condensed value that shows over all. If you want to see his Value as a hitter you look at wRC+ and as a defender DRS.

These values are tools. You can also look up statcast info.