Is ER/GS More Appropriate?

[Tyrus4189Cobb]

Starting pitchers often don't throw beyond the sixth inning or seventh inning. I am not an opponent of ERA because everyone is being judged on an even per nine inning basis. If you allow 2 ER in 7 IP, you have a chance next game to pitch 2 scoreless innings to round your ERA back to 2.00.

Yet it's difficult for a starter to keep this number low without the extra few innings in their start to water it down a bit. If you pitch 6 innings, your ERA is only going to be so low after allowing runs. It jumps up to 3.00 once you allow 2 runs, but 2 ER in 6 IP is a good start. Let's take Spaghetti Eddie, who starts five games as an example.

Game 1: 5 ER, 5 IP
Game 2: 3 ER, 7.2 IP
Game 3: 4 ER, 6 IP
Game 4: 2 ER, 6.1 IP
Game 5: 2 ER, 5.2 IP

Eddie's ERA is 4.70. Not exactly glamorous. On the other hand, he allowed 3.2 ER per game started, which really isn't too bad over five games. A good team can get 3 or 4 wins out of that.

Do you think ER/GS (I like to pronounce it "ergs") is more appropriate for 21st century baseball (in addition to considerng their innings pitched)?

[Ben Grimm]

I'm not sure I'm reading it correctly, so feel free to point me in the right direction.

From what I see, it could be somewhat okay as long as everybody has the same amount of starts. If you take a look at the Giants for example, both Cain and Bumgarner have 2.83 ERAs while Vogelsong's is 2.85. But when you break it down ER/GS, Vogelsong is 2.04, Bumgarner 2.16 & Cain is 2.20.

The other part that beomes involved is pitchers who also pitched out of the pen for any games (which is the case of two top-10 ERA leaders in the NL - Cain & Miley). So for these guys, how do you treat that? Do you then have to go thru their splits and only use the ERs they gave up as a starter?

I like ERA a bit better if only becuase it uses a larger quantity sample as the divisor with innings pitched. I'd actually like to see a run of say 100+ random pitchers done your way and see if there are any larger discrepencies along the way, though.

[ipitch]

What purpose would it serve? Isn't it really just a less-accurate way of measuring ERA? It would give the 6-inning guys an advantage over the 8-inning guys. It would be the equivalent of saying that going 2-3 as a batter is equal to going 2-5.

[Tyrus4189Cobb]

I like ERA a bit better if only becuase it uses a larger quantity sample as the divisor with innings pitched. I'd actually like to see a run of say 100+ random pitchers done your way and see if there are any larger discrepencies along the way, though.

I would use only games started.

I may just do that. 100 isn't unreasonable. To clarify, I'm not supporting ER/GS over ERA, I'm thinking aloud over this forum.

[Tyrus4189Cobb]

What purpose would it serve? Isn't it really just a less-accurate way of measuring ERA? It would give the 6-inning guys an advantage over the 8-inning guys. It would be the equivalent of saying that going 2-3 as a batter is equal to going 2-5.

You could say that for any non-counting stat. A 150 OPS+ in 100 PA isn't better than 150 OPS+ in 500 PA. That's why you have to consider the amount of innings pitched over the season and he average of a given start in order to quantify it.

This isn't the equivalent of your example because it's based on percentages and ratios. Going 2-5 or 40-100 is mathematically equivalent because it's 40%, but I've never seen anyone argue that the former is realistically equal. That's why there are required minimums for statistical titles.

Again, I'm not supporting ER/GS over ERA, just throwing out my thoughts.

[JDanger]

On the other hand, he allowed 3.2 ER per game started, which really isn't too bad over five games. A good team can get 3 or 4 wins out of that.

This guy:
Game 1: 5 ER, 8 IP
Game 2: 3 ER, 8.2 IP
Game 3: 4 ER, 9 IP
Game 4: 2 ER, 7.1 IP
Game 5: 2 ER, 8.1 IP

and this guy:

Game 1: 5 ER, 2.2 IP
Game 2: 3 ER, 3.2 IP
Game 3: 4 ER, 4.2 IP
Game 4: 2 ER, 1.1 IP
Game 5: 2 ER, 5.2 IP

both have the same score.

[Tyrus4189Cobb]

This guy:
Game 1: 5 ER, 8 IP
Game 2: 3 ER, 8.2 IP
Game 3: 4 ER, 9 IP
Game 4: 2 ER, 7.1 IP
Game 5: 2 ER, 8.1 IP

and this guy:

Game 1: 5 ER, 2.2 IP
Game 2: 3 ER, 3.2 IP
Game 3: 4 ER, 4.2 IP
Game 4: 2 ER, 1.1 IP
Game 5: 2 ER, 5.2 IP

both have the same score.

This guy:
Game 1: 3 ER 6 IP
Game 2: 4 ER 7 IP
Game 3: 1 ER 7 IP

Has the same ERA as:

Game 1: 2 ER 9 IP
Game 2: 2 ER 7 IP
Game 3: 5 ER 4 IP
Game 4: 1 ER 5 IP

But ERA is still used not only by itself but for ERA+, even if IP don't match (99% of the time).

See post #5

[ipitch]

This guy:
Game 1: 3 ER 6 IP
Game 2: 4 ER 7 IP
Game 3: 1 ER 7 IP

Has the same ERA as:

Game 1: 2 ER 9 IP
Game 2: 2 ER 7 IP
Game 3: 5 ER 4 IP
Game 4: 1 ER 5 IP

They deserve to have the same ERA.

[drstrangelove]

Starting pitchers often don't throw beyond the sixth inning or seventh inning. I am not an opponent of ERA because everyone is being judged on an even per nine inning basis. If you allow 2 ER in 7 IP, you have a chance next game to pitch 2 scoreless innings to round your ERA back to 2.00.

Yet it's difficult for a starter to keep this number low without the extra few innings in their start to water it down a bit. If you pitch 6 innings, your ERA is only going to be so low after allowing runs. It jumps up to 3.00 once you allow 2 runs, but 2 ER in 6 IP is a good start. Let's take Spaghetti Eddie, who starts five games as an example.

Game 1: 5 ER, 5 IP
Game 2: 3 ER, 7.2 IP
Game 3: 4 ER, 6 IP
Game 4: 2 ER, 6.1 IP
Game 5: 2 ER, 5.2 IP

Eddie's ERA is 4.70. Not exactly glamorous. On the other hand, he allowed 3.2 ER per game started, which really isn't too bad over five games. A good team can get 3 or 4 wins out of that.

Do you think ER/GS (I like to pronounce it "ergs") is more appropriate for 21st century baseball (in addition to considerng their innings pitched)?

I'm sure I misunderstand your post.

A pitcher who ONLY throws complete games and averages 2.5 earned runs per games is better than a pitcher who ONLY throws 6 innings per game and averages 2.4 earned runs per game, although this stat would unequivocally imply the opposite. Adjustments for PF and league would not correct this misleading result.

A number of per game stats were common in the 19th century when players (including pitchers) typically played the entire game. The value of this type of metric is lost when players do not play full games which, among pitchers, became more common after 1900, and became the norm after the 1950's.

The reason per game stats are not used is because they almost always provide inaccurate results.

[Brooklyn]

Eddie's ERA is 4.70. Not exactly glamorous. On the other hand, he allowed 3.2 ER per game started, which really isn't too bad over five games. A good team can get 3 or 4 wins out of that.
(in addition to considerng their innings pitched)?

ER/GS has no value at all. As already pointed out, a pitcher going 5 and 9, giving up the same runs would rate the same, and there is clearly not the same value

In your example, 4.70 is not exactly glamorous, but 3.2 per game start isn't good either, with an average of just over 6 innings. I think you are forgetting that the bullpen is likely to give up run, too. The team isn't any more likley to win these games. In fact, a pitcher with a 4.7 ERA pitching 5 complete games is much more valuable then your pitcher above

[Tyrus4189Cobb]

I'm beginning to see the dissension. After looking at 50 pitchers in 2011, I found that those with 3.00 or more ERGS tend to be the losers. At first I thought this metric may be more appropriate because of the amount of IP today. I see it has too man problems. Live and learn, that was the point of this thread.

ERGS was just a curiousity thing. I'm working on another stat right now. I hope to harass the BBF community with it soon in the .1% chance that I one day make a viable one. My goal is to create one and set up a website where people will frequently visit for reference.

I appreciate everyone's feedback.