Expected ERA OOTP6

[John Marsh]

For OOTP5, a pitcher’s ability to avoid hits determined (more than any other talent) the number of runs he would prevent. As a result, most of us would not even look at a pitcher if he did not have a “Good” rating in avoiding hits. A pitcher’s ability to prevent home runs, base on balls, and doubles mattered, of course, (in that order in fact), but it did not matter nearly as much as his ability to avoid hits.

With the advent of DIPS, however, all that changes. The question then quickly becomes what talents will matter most for a pitcher given the DIPS system? Obviously, what will matter is a pitcher’s ability to avoid home runs and walks and to strike batters out--but in what order of importance and in what proportion?

I’ve only started to work on that problem, but I wanted to see if what I’ve begun looks right to other people.

If (according to the current OOTP5 rating system) we have a pitcher with Brilliant talent ratings in avoiding home runs, avoiding walks, and striking out batters, we know that he will (if he achieves the minimum numbers for an “8” rating in each category):

(1) surrender 10 home runs per 550 at bats;
(2) give up 42 walks per 550 batters faced;
and (3) strike out 143 batters per 550 at bats.

Because avoiding walks is measured in batters faced and not at bats, we need to reduce the number of home runs and strikeouts to the number of batters faced minus number of walks given up. This would give us the following raw stat line:

AB HR BB K

508 9 42 132

This leaves 367 Balls in Play (Batters Faced-(HR+BB+K). Over the past three years, the average of balls in play that resulted in hits was roughly .2974. If we multiply that number by the number of balls in play, our pitcher will have surrendered roughly 109 hits on balls in play. Over the last three years, .228 of those hits on balls in play went for doubles and .024 went for triples. So that should give our prospect the following raw numbers:

AB H 2B 3B HR BB K

508 118 (109+9HR) 25 3 9 42 132

The estimated number of runs created for such a line is 51.2. The number of innings pitched ((AB-H)/3) is 130.0. If roughly .925 percent of all runs are earned runs, then the number of earned runs should be 47.36, which would translate into an expected ERA of 3.28.

That 3.28 expected ERA should represent the ERA of a pitcher with an “8” rating in avoiding home runs, avoiding walks, and striking out batters who pitches for an average defensive team in a neutral park, right?

The goal, ultimately, will be to figure out which categories (Home Runs, Base on Balls, Strikeouts) have more influence over expected ERA. So, for example, should you draft (all other things being equal) the prospect with a Brilliant ability to strike batters out but only average abilities to prevent home runs and base on balls? Or the prospect with Good ability across all three categories?

But I want to make sure that I understand DIPS sufficiently and that the formula I have laid out here is accurate. Any feedback is welcome.

[andy m]

interesting post. surprised nobody has replied, maybe that's because it's not an april fools joke or written all in caps.

[Skipaway]

There are two issues we need to study.

1. Within the defense independent stats.

We need to figure out how the ratings effect expected defense-independent stats. Then we can multiply and sum up the expected run numbers from linear weight, and we'll get a correlation between run value and ratings.

2. The value of avoiding balls in play.

This would be the tricky part. Some pitchers just put more balls in play, and how good an idea is that depends mostly on defense. Is it better to be high K rate but high BB rate, or low BB rate but low K rate? Perhaps the former should be valued higher since it might allow the team to go with worse defense? However, wouldn't that kind of pitcher tend to have higher pitch counts and pitch less innings?

[Henry]

Originally posted by andy m
interesting post. surprised nobody has replied, maybe that's because it's not an april fools joke or written all in caps.

Most folks don't understand DIPS well enough yet to offer constructive comments. This is apparent following most of its critisizms on this board as to its use in the game. This will take some time

In another thread, I've offered to publish constructive and intelligent articles concerning any facet of baseball simulations at the Baseball Simulations Group website. I hope everyone aspiring to become an author and analyst considers that option

Henry

[I Am The Game]

Originally posted by Skipaway
There are two issues we need to study.

1. Within the defense independent stats.

We need to figure out how the ratings effect expected defense-independent stats. Then we can multiply and sum up the expected run numbers from linear weight, and we'll get a correlation between run value and ratings.

2. The value of avoiding balls in play.

This would be the tricky part. Some pitchers just put more balls in play, and how good an idea is that depends mostly on defense. Is it better to be high K rate but high BB rate, or low BB rate but low K rate? Perhaps the former should be valued higher since it might allow the team to go with worse defense? However, wouldn't that kind of pitcher tend to have higher pitch counts and pitch less innings?

High BB/High K = Kerry Wood

A good example might be someone like Derek Lowe in Boston. I mean, we need to see what happens this year, but I don't know how much of a coincidence it is that the guy had a Cy Young-Caliber season with Rey Sanchez at 2B, and was average with Todd Walker. And all indications are Pokey Reese is going to make this guy a very rich man next year. I would like to see Defense play a role in a pitcher's ability if he has a low K rate (therefore putting more balls into play). It would force you to actually factor in defense, and if you sacrifice the defense to get offense it therefor adversly effects your pitching staff. A team like the Cubs has the ability to sacrifice a little D for the Offense, but a team like the Braves, cannot.

[Craig Scarborough]

Originally posted by I Am The Game
I would like to see Defense play a role in a pitcher's ability if he has a low K rate (therefore putting more balls into play). It would force you to actually factor in defense, and if you sacrifice the defense to get offense it therefor adversly effects your pitching staff. A team like the Cubs has the ability to sacrifice a little D for the Offense, but a team like the Braves, cannot.

This is an automatic - the more balls in play, the more your team's defense factors into the equation. Just a reminder, the defensive impact has been altered a bit in OOTP6 based on real play-by-play data, so what you learned in OOTP5 may not always be the case in OOTP6.

Regarding this discussion in particular, OOTP is getting much closer to the "promised land". Meaning: more like MLB, rather than some kind of "OOTP version" of the real game. Therefore, it's my suggestion you value the same attributes in the game as you would in real life. This all gets very interesting, as the basic strategies in the game more closely resemble the ones you'll see a real GM/field manager make.

As an example, suppose your #3 starter was pretty good last year - had 3.50 ERA, won 15 games for you. He's always had good control and kept the ball in the park, however, he gave up more than a hit per inning. Was this an aberration (sp)? Studies have shown that pitchers that give up more than a hit per inning are more likely to pitch poorly the following year (remember Minnesota's Alan Anderson?). Is this my pitcher's true level of performance? If so, he's certainly more volatile. Do I try to improve the defense around him to lessen the chances of his downfall? Do I consider his 3.50 ERA a "career year"? Or, was the "hit per inning" piece the fluke and he's actually a better pitcher than that? Or, is his control and his ability to keep the ball in the park so good that I accept this "hit per inning" number?

What's interesting to me (so far in the beta) is I'm seeing that pitchers who have excellent control and keep the ball in the park have value. Their career paths seem more volatile (though there's a certain level of volatility in every pitcher), but they have less room for error than the guy with excellent control and 9+ strikeouts per game.

I guess to bottom line it - pitcher strikeouts have much more value in OOTP6 than in OOTP5 (just as in real life), however, just looking at a pitcher's K per 9 innings won't give you the whole story. There are an awful lot of good pitchers out there who can't strike out the side every other inning.

-Craig

[John Marsh]

One of the most interesting things to emerge from this discussion is the impossibility of separating defense from pitching, which, of course, is the whole point of DIPS. What I am after, however, is a way to evaluate pitching talent independent of defense. In other words, when you draft a pitching prospect, presumably you're drafting for a future team whose defensive strengths and weaknesses will have altered significantly. We will of course need to learn much more about the rating system for OOTP6. And, as Skipaway suggests, we will need to learn much more about the OOTP6 engine before we can make many conclusions. For example, Craig suggests that the OOTP6 engine was influenced by play-by-play data, which I assume refers to the study that distinguished between line drives, ground balls, fly balls, and pop outs for purposes of BABIP. Pitchers, if I remember the study right, have *some* influence over preventing line drives. That might mean, even if only in the terms of the limited study posted above, that it is not accurate to assign pitchers league average doubles and triples as a percentage of their number of hits on balls in play. That confirms Oakland's stat wizard Paul de Podesta, who, if I am remembering Moneyball right, believed that pitchers had some influence over preventing extra base hits. It will be quite a task to sort this out--but what fun. And Henry I will gladly submit any final results for the baseball simulations group.

[Henry]

And Henry I will gladly submit any final results for the baseball simulations group.

Excellent !!

Henry

[Craig Scarborough]

John,

Just to clarify...when I spoke of the defense being tweaked in OOTP6, I was referring to defense only, not the ability or inability of pitchers to prevent extra base hits.

In the past, it was found that good defense perhaps had too much influence on balls in play. However, after looking at play-by-play data, Markus was able to tweak individual defensive performance when the ball is hit in the responsible fielder's "zone". Obviously, defense will still play a part in OOTP, but it's now more in line with real MLB performance.

-Craig

[Craig Scarborough]

John,

Just to clarify...when I spoke of the defense being tweaked in OOTP6, I was referring to defense only, not the ability or inability of pitchers to prevent extra base hits.

In the past, it was found that good defense perhaps had too much influence on balls in play. However, after looking at play-by-play data, Markus was able to tweak individual defensive performance when the ball is hit in the responsible fielder's "zone". Obviously, defense will still play a part in OOTP, but it's now more in line with real MLB performance.

-Craig

[rwd59]

Originally posted by Henry
Most folks don't understand DIPS well enough yet to offer constructive comments.

I don't need to understand it. I just want it to work!

[Henry]

Originally posted by rwd59
I don't need to understand it. I just want it to work!

That's certainly the intention

Henry

[nestajones]

Originally posted by John Marsh
what talents will matter most for a pitcher given the DIPS system? Obviously, what will matter is a pitcher’s ability to avoid home runs and walks and to strike batters out--but in what order of importance and in what proportion?

If (according to the current OOTP5 rating system) we have a pitcher with Brilliant talent ratings in avoiding home runs, avoiding walks, and striking out batters, we know that he will (if he achieves the minimum numbers for an “8” rating in each category):

(1) surrender 10 home runs per 550 at bats;
(2) give up 42 walks per 550 batters faced;
and (3) strike out 143 batters per 550 at bats.

Because avoiding walks is measured in batters faced and not at bats, we need to reduce the number of home runs and strikeouts to the number of batters faced minus number of walks given up. This would give us the following raw stat line:

AB HR BB K

508 9 42 132

This leaves 367 Balls in Play (Batters Faced-(HR+BB+K). Over the past three years, the average of balls in play that resulted in hits was roughly .2974. If we multiply that number by the number of balls in play, our pitcher will have surrendered roughly 109 hits on balls in play. Over the last three years, .228 of those hits on balls in play went for doubles and .024 went for triples. So that should give our prospect the following raw numbers:

AB H 2B 3B HR BB K

508 118 (109+9HR) 25 3 9 42 132

The estimated number of runs created for such a line is 51.2. The number of innings pitched ((AB-H)/3) is 130.0. If roughly .925 percent of all runs are earned runs, then the number of earned runs should be 47.36, which would translate into an expected ERA of 3.28.

That 3.28 expected ERA should represent the ERA of a pitcher with an “8” rating in avoiding home runs, avoiding walks, and striking out batters who pitches for an average defensive team in a neutral park, right?

tangotiger developed a simplified version of calculating a DIPS ERA, called FIP (fielding independent pitching, i think). if i remember correctly, the formula is:

((13*HR)+(3*BB)-(2*K))/IP) + 3.20 (I think the final constant might change in different eras, not sure...but 3.2 is good for now.)

So your above pitcher would have a fielding-independent ERA of 3.04.

In the ballpark with your estimate, but a fair amount lower. Not sure why this would be...I'll take a closer look at your method when I have time later.

[nestajones]

Following up...I plugged our imaginary 8/8/8 pitcher into the DIPS ERA calculator (available at http://yankeefan.blogspot.com/files/DIPSWorksheets.ZIP). His DIPS ERA (assuming no HBP) is 3.07.

Still not sure why there's a discrepancy between the FIP ERA and DIPS ERA, and the estimated ERA by components. It's not huge, but it puzzles me.

By the way, that constant ("+ 3.20") in the FIP ERA formula I gave above does not work for all eras. So you can use that formula to compare one pitcher to another, but to actually calculate their FIP ERA, you need to determine the right constant for the league and year.

[Rob Fenix]

Originally posted by Craig Scarborough
What's interesting to me (so far in the beta) is I'm seeing that pitchers who have excellent control and keep the ball in the park have value. Their career paths seem more volatile (though there's a certain level of volatility in every pitcher), but they have less room for error than the guy with excellent control and 9+ strikeouts per game.
-Craig

So my 1 Star SP (who happened to win 19 games, along with the Cy Young last year) should continue to flourish in OOTP6? He keeps the ball in the park, avoids walks, and posted about a 4.5:1 K:BB ratio over 253.2 IP. The defense that played behind him last season was in the bottom half of the league also. At 31 yrs old last season, he actually bumped his Runs and Hits ratings up to a whopping 3. This guy has not only defied his scouting reports, but defied our league averages and many of our ballpark factors in the process.

[chrisj]

Somewhat changing the subject...

Originally posted by John Marsh
For OOTP5, a pitcher’s ability to avoid hits determined (more than any other talent) the number of runs he would prevent. As a result, most of us would not even look at a pitcher if he did not have a “Good” rating in avoiding hits. A pitcher’s ability to prevent home runs, base on balls, and doubles mattered, of course, (in that order in fact), but it did not matter nearly as much as his ability to avoid hits.

Which gave people like me a huge advantage in online leagues. (and an even bigger advantage in solo play as computer teams wouldn't touch players with say a poor in avoid hits).

http://jobl.end-on-end.com/FTP/p1310.html

Take Willams Raglin for example. In the JOBL, no one touched him until I traded for him last year when he was 32. Brought him to the majors and he went 10-9 with a 2.80 ERA, and this season so far he is 4-1 with a 2.09 ERA.

With a poor in avoiding hits (and an overall avoid hits of 2). His key, by far, is the brilliant in avoid homeruns and avoid walks - which made him a star pitcher.

I'm going to miss players like this come OOTP6.