Stat Question(s)

[jay87shot]

I don't know where to put this but thought maybe more people might have questions about some advanced stats.

Over the years I have liked using the statistic WAR as a good reference on how strong a players overall season has gone. However I have seen some funny things recently, here is an example. I usually use baseball-reference for stats, I know there are a few different versions like fangraphs use FIP for pitchers and other differenes.

Freddy Peralta 6 innings 1 run 1 hit 8 k's 1 bb = 0.2 WAR

Joe Ross 3.2 innings 0 runs 2 hits 3 k's 5 bb= 0.2 WAR

I am sure the ERA plays a bigger part and WAR really judges a longer part of the season. It just seems like Freddy was amazing and Ross wasn't great. Does anyone have an easy explanation on WAR. I get it is based on a bell curve and has like 6 categories that are used to determine the wins above replacement. 

[sveumrules]

Baseball Reference pitcher WAR is the most convoluted. It is based off of runs allowed with adjustments for defense.

FanGraphs default pitcher WAR is based off of FIP (Ks, BBs, HRs and IFFBs), it strips out defense entirely. They also have a runs allowed based WAR available if you toggle over to their value leaderboard.

For me personally I pretty much ignore BRef WAR for pitchers and use a mix of the two FanGraphs versions, while obviously looking at underlying stats as well. Just like two guys with the same OPS might get there in different ways (OBP/SLG heavy or evenly distributed) two pitchers with the same FIP might also get there in different ways.

Position players the main difference is BRef uses DRS for defense, while FanGraphs uses OAA. Personally I tend to come down closer to the larger defensive swings that DRS spits out (+85 to -79 on a team level in 2023) versus OAA (+41 to -50 on a team level in 2023), but I tend to default to FanGraphs WAR because their leaderboard interface is just so much easier to navigate, especially when looking at multiple seasons.

[Thurston Fluff]

The only thing I think WAR is good for is offense. It's hard to quantify the value of defense compared to offense. It's value is also somewhat dependent of the type of pitcher as well.

[brewcrewdue80]

11 hours ago, jay87shot said:

I don't know where to put this but thought maybe more people might have questions about some advanced stats.

Over the years I have liked using the statistic WAR as a good reference on how strong a players overall season has gone. However I have seen some funny things recently, here is an example. I usually use baseball-reference for stats, I know there are a few different versions like fangraphs use FIP for pitchers and other differenes.

Freddy Peralta 6 innings 1 run 1 hit 8 k's 1 bb = 0.2 WAR

Joe Ross 3.2 innings 0 runs 2 hits 3 k's 5 bb= 0.2 WAR

I am sure the ERA plays a bigger part and WAR really judges a longer part of the season. It just seems like Freddy was amazing and Ross wasn't great. Does anyone have an easy explanation on WAR. I get it is based on a bell curve and has like 6 categories that are used to determine the wins above replacement. 

SP WAR is super cringy imo due to the WAR easily gained/lost by RP. There's just some figure with SPs that should increase their values earned. 

I like Fip regarding pitchers usually. Fip gets hit hard on HR rate tied in to Ks and Whip. You can figure out who stands to be volatile. 

I wouldn't discount WAR that Freddy has vs Ross comes to rounding up or down. Freddy having .242 WAR let's say and Ross having .186 WAR. Gets rounded both to .2  yeah that has to be it. You don't see WAR to the hundredths.

 

[SomewhereInTime]

13 hours ago, sveumrules said:

Baseball Reference pitcher WAR is the most convoluted. It is based off of runs allowed with adjustments for defense.

FanGraphs default pitcher WAR is based off of FIP (Ks, BBs, HRs and IFFBs), it strips out defense entirely. They also have a runs allowed based WAR available if you toggle over to their value leaderboard.

For me personally I pretty much ignore BRef WAR for pitchers and use a mix of the two FanGraphs versions, while obviously looking at underlying stats as well. Just like two guys with the same OPS might get there in different ways (OBP/SLG heavy or evenly distributed) two pitchers with the same FIP might also get there in different ways.

Position players the main difference is BRef uses DRS for defense, while FanGraphs uses OAA. Personally I tend to come down closer to the larger defensive swings that DRS spits out (+85 to -79 on a team level in 2023) versus OAA (+41 to -50 on a team level in 2023), but I tend to default to FanGraphs WAR because their leaderboard interface is just so much easier to navigate, especially when looking at multiple seasons.

Outstanding post

[jay87shot]

I usually read most of my league wide baseball on mlbtraderumors and it links to baseball reference, I will have to try to use fangraphs more. Thanks for the help.

[homer]

18 hours ago, Thurston Fluff said:

The only thing I think WAR is good for is absolutely nothin' just say it again... 

FIFY

[JosephC]

I tend to think the opposite way from sveumrules, and prefer Baseball Reference when it comes to looking at pitcher WAR.  Primarily because I think fWAR extremely over-values pitchers that are performing on a mediocre level.  Best example I can quickly give for Brewer fans is Matt Garza during his last year in Milwaukee.  He was awful, everyone who watched the Brewers that season would tell you he was awful and the message boards were just filled with comments that the Brewers should just release him and put a young guy (or even organizational filler) out there because Garza wasn't passing anyone's "eye-test."  He finished the year with a 4.94 ERA and a K/BB ratio well below 2.  FIP = 4.91, xFIP = 5.12, SIERA = 5.25.  Only made 22 starts (24 appearance in all) and Fangraphs came up with a 0.9 WAR (which pro-rates to 1.3 WAR if he had started the full season...32 starts).  Baseball Reference has him at -0.2.  Anybody who watched the Brewers that season would say the -0.2 figure seemed like the more accurate figure.

But in the end, it doesn't take that much time to look at both sites and see both numbers.  Whenever I try to figure out trade value or something along those lines, I always look at both sides, take the average and call it the bWAR/fWAR value. 

[sveumrules]

38 minutes ago, jay87shot said:

I usually read most of my league wide baseball on mlbtraderumors and it links to baseball reference, I will have to try to use fangraphs more. Thanks for the help.

I’ve found a good way to calibrate the different WARs is by comparing some of the more divisive players.

Like Burnes last year. He threw 193 IP of 3.39 ERA and 3.81 FIP. FG valued that at 4.8 runs WAR and 3.4 FIP WAR while BRef had him at 3.5 WAR. The perception of Corbin is closest to the rWAR, but the fWAR and bWAR both thought he needed some help from his defense to get there.

Miley was even more extreme at 120 innings of 3.14 ERA and 4.33 FIP which came out at 3.4 rWAR, 1.1 fWAR and 2.6 rWAR. His FIP isn’t as good because he doesn’t strike anybody out, but the run prevention is still top notch so the question comes down to how much credit goes to Miley and how much goes to the defense to account for that difference.

That’s essentially the crux of why pitching WAR is so tricky, we know ERA isn’t solely up to the pitcher. Their defense, ballpark and some degree of “luck” plays into it, whether in the form of batted ball results or sequencing. 

On the position player side Turang came in at 1.6 bWAR but only -0.2 fWAR because DRS had him at +12 on the season compared to OAA having him even with +2 at 2B but -2 in his SS innings. I think most would side with DRS on the eye test, but still have a hard time buying that was enough to make him essentially an average player  with such a poor batting line. Like with most things, the truth is probably somewhere in the middle.

Another one would be Contreras, 5.7 fWAR last year, but only 3.8 bWAR. This one essentially boils down to FG giving credit to William for his framing while BRef doesn’t. The opposite end of the spectrum is a guy like Sal Perez who has 33.1 career bWAR but only 15.4 fWAR because he’s graded out at -115 framing runs.

[nate82]

I believe WPA is a better stat to use than WAR for pitching.  I believe it correlates better than WAR does at least for pitchers on determining who is having a good season. 

[owbc]

1 hour ago, JosephC said:

I tend to think the opposite way from sveumrules, and prefer Baseball Reference when it comes to looking at pitcher WAR.  Primarily because I think fWAR extremely over-values pitchers that are performing on a mediocre level.  Best example I can quickly give for Brewer fans is Matt Garza during his last year in Milwaukee.  He was awful, everyone who watched the Brewers that season would tell you he was awful and the message boards were just filled with comments that the Brewers should just release him and put a young guy (or even organizational filler) out there because Garza wasn't passing anyone's "eye-test."  He finished the year with a 4.94 ERA and a K/BB ratio well below 2.  FIP = 4.91, xFIP = 5.12, SIERA = 5.25.  Only made 22 starts (24 appearance in all) and Fangraphs came up with a 0.9 WAR (which pro-rates to 1.3 WAR if he had started the full season...32 starts).  Baseball Reference has him at -0.2.  Anybody who watched the Brewers that season would say the -0.2 figure seemed like the more accurate figure.

But in the end, it doesn't take that much time to look at both sites and see both numbers.  Whenever I try to figure out trade value or something along those lines, I always look at both sides, take the average and call it the bWAR/fWAR value. 

My interpretation of Garza's WAR is that replacement level pitching is always worse than we think it is. 

In 2017, it looks like replacement level xFIP was around 5.4. So closer to 2017 Junior Guerra, who started 14 games and racked up -0.1 bWAR, -0.7 fWAR. So as bad as Garza was, the 2017 Brewers still had to cover about 25 starts with guys who were worse than him. 

[sveumrules]

1 hour ago, nate82 said:

I believe WPA is a better stat to use than WAR for pitching.  I believe it correlates better than WAR does at least for pitchers on determining who is having a good season. 

Definitely better for relievers since how good they are typically correlates with what kind of leverage situations they are brought into, but it's kind of noisy for starters since so much of their WPA is tied up in what kind of run support they get and it also doesn't really account for differences in volume.

Like last year Miley (+1.98) came out ahead of Peralta (+1.71) and Burnes (+1.63) by WPA. But all that's really telling us is Wade did a better job of stranding the extra base runners he put on and doesn't give Freddy or Corbin any credit for the combined 117 additional innings they pitched.

[Thurston Fluff]

Silly question, how often does the replacement level get updated? By definition it's based on stats accumulated by freely available talent but that changes year to year.

[Playing Catch]

Is there any reason that we would WANT to credit cagey pitchers that go 6+IP with low pitch counts, few strike outs, and the ability to pitch around dangerous hitters? Maybe relying on FIP doesn't really tell the whole story, either.

[Thurston Fluff]

14 minutes ago, Playing Catch said:

Is there any reason that we would WANT to credit cagey pitchers that go 6+IP with low pitch counts, few strike outs, and the ability to pitch around dangerous hitters? Maybe relying on FIP doesn't really tell the whole story, either.

FIP was never meant to tell the whole story. I think there's a legitimate argument that it's overvalued though. Or, more accurately, it doesn't have the same value for every team. The value of a high K, low walk pitcher with the Brewers defense isn't nearly as high as it is for the Nationals.