Starters v Relievers - Page 6

andymac · 05-26-2006, 09:49 AM

Quote:

Originally Posted by dougaiton

Could one problem be that OOTP will make any good reliever with decent endurance a starter? Anyone with an endurance over x will automatically become a starter if his ability is over y

But is that really a problem? If the pitcher is good enough, shouldn't he be starting?

Gastric ReFlux · 05-26-2006, 11:05 AM

We've got a way to compare OOTP versus real life.

http://reports.ootpbaseball2006.com/...00_0_2006.html

We can use the pitching registers to get a decent approximation of starting results versus relief results. Toss out any pitchers who have made both starts and relief appearances, then use the ones who have appeared in strictly one role to get an idea of the results we're seeing.

We'd be interested in the following numbers:

ERA so we need IP and ER's.
Hits allowed
Strikeouts
Walks
HR

Convert IP, Hits, HR, SO, and BB to a rate stat of per 9 innings.

I won't have time this afternoon to work on this, but it could be very helpful for us to crunch these numbers and see how OOTP is performing compared to real-life which can be found at Retrosheet.

andymac · 05-26-2006, 11:26 AM

Quote:

Originally Posted by Gastric ReFlux

We've got a way to compare OOTP versus real life.

http://reports.ootpbaseball2006.com/...00_0_2006.html

We can use the pitching registers to get a decent approximation of starting results versus relief results. Toss out any pitchers who have made both starts and relief appearances, then use the ones who have appeared in strictly one role to get an idea of the results we're seeing.

We'd be interested in the following numbers:

ERA so we need IP and ER's.
Hits allowed
Strikeouts
Walks
HR

Convert IP, Hits, HR, SO, and BB to a rate stat of per 9 innings.

I won't have time this afternoon to work on this, but it could be very helpful for us to crunch these numbers and see how OOTP is performing compared to real-life which can be found at Retrosheet.

Here are the rates I got after eliminating the pitchers who were used in both rolls.

Code:

	RP	SP

H/9	9.04	8.89
ERA	4.30	4.00
K/9	6.92	6.94
BB/9	3.77	2.92
HR/9	0.94	0.91

Joe Morgan · 05-26-2006, 01:06 PM

bump because quite frankly this is about 50 times more important than Laughing Boy and his non-hidden hidden ratings.

tysok · 05-26-2006, 01:13 PM

Quote:

Originally Posted by andymac

Here are the rates I got after eliminating the pitchers who were used in both rolls.

Code:

	RP	SP

H/9	9.04	8.89
ERA	4.30	4.00
K/9	6.92	6.94
BB/9	3.77	2.92
HR/9	0.94	0.91

Take another look at the numbers, I think you may have left a guy in that shouldn't... I made that mistake and got the same numbers, but when games and games started for the "starters" didn't add up I found him.. he has 25 starts and 26 appearances (if you made the same mistake. Numbers change only slightly and of course only for starters:
H/9 8.76
ERA 3.93
K/9 6.87
BB/9 2.85
HR/9 .90

But I wouldn't take these numbers for much of any representation personally.... you're dropping 57 (18%) pitchers. Accounting for 221 (16.9%) wins, 273 (20.9%) losses, 1133 (12.8%) relief appearances, 640 (24.4%) starts, 4543.2 (19.4%) IP, 4617 (19.9%) hits, 2119 (19.9%) ER, 481 (20%) HR, 1636 (19.8%) BB, 3214 (18.2%) Ks with 4.20 ERA.

The assumption in throwing these guys out is that these stats are 50-50 between starters and relievers... which is more than likely not the case. The guys that were cut that were primarily starters (had only 8 or fewer relief appearances compared with at least 9 starts) had an ERA of 4.33 combined.

So you have a representation of 80% of 1 year of the new game, which isn't bad and clearly shows incorrect numbers, but you don't have an accurate years data to drive from... I wish it broke it down so we could get real numbers.

No Pepper · 05-26-2006, 03:31 PM

Here are the rates I got after eliminating the pitchers who were used in both roles, for 2007 the following year.

Code:

	RP	SP

H/9	9.18	9.20
ERA	4.27	4.37
K/9	6.56	6.52
BB/9	3.95	3.12
HR/9	0.89	0.91

Looking at players that appeared as both a starter and reliever, there were about 7-10 low endurance guys (~3) that had 20+ starts and less than 100 innings. All of them had ERAs under 4. Jacksonville had 2 of these guys in their rotation (Abernathy and Blezard) and went on to win their division, going 100-62 in 2007. Without split totals for starts it's hard to tell if these guys would change the results.

It must be an AI-related, whether it be the situation relievers enter the game (start of inning or not) or in transactions. I've seen starters dropped from the ML club and relievers on the roster (low endurance CL or MR) taking over the rotation spot, whereas IRL, you promote a AAA SP or make a trade for a SP to fill the void.

All that said, ERA isn't a good measure for the skills of players, if you're talking about ratings boosts.

soundsradio · 05-27-2006, 03:59 AM

Quote:

Originally Posted by Gorilla Shakespeare

"

Exactly what I was thinking on the first point, and you raise an excellent second point. One more thing: it's flat-out not realistic that some pitchers' endurance is 40 pitches as a starter. *Every* pitcher should be able to throw at least 85-90 pitches per game. As a parallel to real life, that's about the point Rick Honeycutt crapped out, and he was about as extreme a situational reliever as you could ever get.

Not quite true. Pitchers have to be built up to 85-90 pitches. If they do not throw much in the spring, they may have their first start limited to 60 pitches, their next start caped at 75 and so on up til around 105 they are built up. Minor league pitchers are more prevelant examples. A guy gets invited to big league camp and mostly pitches mop up, maybe throws 12-15 innings in spring training. This would be a pitcher who does not have a roster spot locked in the big leagues. When he starts the season in the rotation at AA or AAA, he may have his first start with a 60 pitch limit. The same is true for pitchers returning from injury. Unless healing time for an injured player includes time for rehab, that pitchers first start would typically be about 50 pitches. Here are a few examples of pitch counts and their uses.

1. I think it was Kevin Appier years ago with the Royals had a no-hitter after 80 pitches. He had reached his limit and was pulled with the no-hitter intact.

2. John Wasdin pitching in AAA in 2003. He had a perfect game after his pitch limit of 75 pitches (It was his first start of the year and he had not pitched a whole lot in the spring). If he has not had a perfect game, even if he had had a no hitter he would have been pulled. As it was, he had not even had a batter with a three ball count so they let him continue. He threw a perfect game. Struck out 15 and threw exactly 100 pitches.

3. Justin Lehr goes from the bullpen to the rotation in AAA. The Brewers don't think he is a major league starter, but they do consider him an option as a long reliever and he needs to develop more than one pitch to be effective. Thus he moves into the rotation and builds up arm strength.

Joe Morgan · 05-27-2006, 01:20 PM

I definitely agree that this is the way things are done in modern baseball. I also clearly remember seeing a guy pitch a no-no in 1990 right after the regular season started (there was an owner's lockout that year, which didn't affect the regular season schedule but drastically shortened spring training) and get taken out in the 7th for one Mike Witt, who wrapped things up.

Thing is, this is one of these issues that I think would already be taken care of in spring training - you know, the thing that the game doesn't simulate. Historically speaking, it is very, very rare for pitchers to consistently be taken out after 4 innings. It's not so rare for relievers to be tried as starters, and for those relievers to fail miserably, which is why I go back to the simple little boost for anybody who throws in relief. I know, I know, it doesn't completely explain Sparky Lyle or Dennis Eckersley, but it's like 70% of the explanation and IMO that's quite an effective "band aid" if you ask me.

Lidstrom · 05-27-2006, 01:30 PM

Quote:

Originally Posted by Joe Morgan

I definitely agree that this is the way things are done in modern baseball. I also clearly remember seeing a guy pitch a no-no in 1990 right after the regular season started (there was an owner's lockout that year, which didn't affect the regular season schedule but drastically shortened spring training) and get taken out in the 7th for one Mike Witt, who wrapped things up.

There was a game where three Braves' pitchers combined on a no-hitter.

Quote:

September 11, 1991 - Kent Mercker (6 inn.), Mark Wohlers (2 inn.) and Alejandro Pena (1 inn.), Atlanta Braves 1 San Diego Padres 0

It's a definite exception to the norm, but demonstrates that it CAN happen in MLB.

Joe Morgan · 05-29-2006, 01:10 PM

Bump! Can I get a TRU DAT in here?

kageboshi · 05-29-2006, 02:03 PM

Quote:

Originally Posted by Joe Morgan

I definitely agree that this is the way things are done in modern baseball. I also clearly remember seeing a guy pitch a no-no in 1990 right after the regular season started (there was an owner's lockout that year, which didn't affect the regular season schedule but drastically shortened spring training) and get taken out in the 7th for one Mike Witt, who wrapped things up.

Thing is, this is one of these issues that I think would already be taken care of in spring training - you know, the thing that the game doesn't simulate. Historically speaking, it is very, very rare for pitchers to consistently be taken out after 4 innings. It's not so rare for relievers to be tried as starters, and for those relievers to fail miserably, which is why I go back to the simple little boost for anybody who throws in relief. I know, I know, it doesn't completely explain Sparky Lyle or Dennis Eckersley, but it's like 70% of the explanation and IMO that's quite an effective "band aid" if you ask me.

Yeah, I agree with that, the boost would help somewhat this year, but I also thing that the boost shouldnt be a constant, that some guys should have a higher boost than others and some might perhaps... pitch worse as relievers.

tysok · 05-29-2006, 11:04 PM

At this point I wouldn't agree with anything.

This thread has sparked my curiosity enough that the first thing I do with my test league is to run some seasons and get some break downs with true complete data over a few years time.

Then I have to pull like data from reality to compare it to. A cursory look into the relief vs starter question prompted me to want more information, which will take a lot of time that I haven't had. I pulled off 15 pitchers from 2005 that pitched 15+ starts with a relatively equal amount of relief appearances... only 1 of them fit every assumption we had (relievers give up less hits, less HRs, and have less ERA, while giving up more walks and getting more Ks). Only 6 of them fit 3 of the 5, and only 2 fit 2 of the 5.

(Those numbers aren't exact, I'm pulling it from memory - it may be way off, but it was not nearly a pretty picture)

At the same time, a low number were traditionally relievers, a slightly higher number were traditionally starters, but most were rookies. That's a very bad sample size to draw any conclusion whatsoever from. But that's an important data set to analyze in the question of whether any pitcher in relief should have a ratings boost.

Basically for this question to get a resolution it's going to take a lot of work with the in game data to discover where the problem is, or if there is a problem. Cursory examinations and small sample sizes lead to nothing.

That being said, I can give you a TRU DAT that if a pitcher is used as a starter (given more than just a few uses), it is rare for them to consistently pitch only 4 innings the majority of the those starts.

I have a hard time believing (hopefully) that the reports put up are from a real league running full tilt with the game we'll be getting on the 31st. at least one of those teams had only 1 pitcher shown as being in a starter role.

Syd Thrift · 05-30-2006, 10:16 PM

Can we at least get a final "no, we don't care about this" or "yes, we think that TangoTiger might have a modicum of authority on the matter and we'll at least look at it" about this issue? Again, I understand that my post count is such that I'm just going to get flamed for this, but this is a legitimate issue. Are you going to wait for Puresim to fix this for you to do the same?

andymac · 05-30-2006, 11:26 PM

Why must people continually **** up any point they are trying to make by bringing PureSim into the discussion.

Syd Thrift · 05-30-2006, 11:32 PM

Quote:

Originally Posted by andymac

Why must people continually **** up any point they are trying to make by bringing PureSim into the discussion.

Why is it a bad point? Puresim is a competitor. It comes up short in many areas compared to OOTP, I will grant you, but it's still out there. It's not like I brought up Mogul or Baseball Stars or something.

Cras · 05-30-2006, 11:35 PM

Quote:

Originally Posted by Syd Thrift

Why is it a bad point?

You are comparing apples to oranges, that is why.

This apple sucks, there is no citric acid in it, orange, a competitor, has citrus acid in it, why doesn't my apple have citric acid in it. This is bull ****, this must be fixed now.

Syd Thrift · 05-30-2006, 11:56 PM

? I have delivered no ultimatums, nor have I called OOTP trash. Surely, however, you must appreciate that TangoTiger knows what he's talking about, and in fact probably knows more than you do about baseball from a sabermetrician's standpoint. I know he knows a lot more than I do, anyway.

Anyway, to add something more than infighting... others have been claiming that this is entirely or mostly due to the AI's poor use of bullpens. However, a look at starter v reliever stats throughout the years indicates that this is not the case. Here are the ERA differences between reliever and starter. A + before the number indicates that relief ERA is lower, and the number is a percentage:

2005: +6.1%
2000: +6.8%
1995: +5.6%
1990: +10.9%
1985: +6.2%
1980: +13.1%
1975: +10.8%
1970: +4.0%
1965: +8.2%
1960: +6.6%
1957: +1.1% (this is the earliest consecutive year Retrosheet has this data available)

A brief look shows that since 1960, at least, relievers have consistently outperformed starters by about 6% or so. Going all the way back to 1922, when relievers were generally only used to mop up for bad outings by starters, the split is 4.9% in favor of starters (4.30 ERA for relievers vs. a 4.10 ERA for starters) - which is *still* a smaller split than what OOTP metes out.

tysok · 05-31-2006, 12:04 AM

I'm not surprised to see no answer from any high authority here. There's yet to be proven that a problem exists, for right now it's just people discussing an issue that may or may not even exist.

I put out the numbers for a season from OOTP5, there was no problem in that season. By the way, my cousin did another 3 seasons the same as I did, 10 seasons from the one I analyzed, 5 years later, and then the year after that... they all came up correct. That doesn't mean a great deal since OOTP2006 is a 'complete rewrite', and that's data from 2 (or 3 if you count 6.5) versions ago.

The data from the new game that we could get on the reports (ignoring anyone pitching in both rolls) was not pretty, but several things lead one to believe it's not an anylization of good data.

After we've had the game (finished version) a few days and looked over several seasons of data we can all speak on this subject better... and if that data shows inaccuracies I'll jump on with these other guys... then we can worry about why the higher powers are or are not bothering to answer. Until then this thread is just conjecture, and if I were the developer I wouldn't bother ansering or posting about conjecture either when so many people are waiting for my game, some are even staying up until release (and at least one is even staging a breathing strike until release)

Syd Thrift · 05-31-2006, 12:17 AM

I disagree. You showed that in your v5 league that the issues Gastric Reflux brought up were not causing starters to get fewer decisions. Perhaps (at least in that one league) the reliever messiness wasn't acting as a "comeback code". You never showed that relievers had lower ERAs and peripheral stats than starters do as in real life.

Good work, though. The discussion only benefitted from your study.

andymac · 05-31-2006, 12:40 AM

Quote:

Originally Posted by Syd Thrift

Why is it a bad point? Puresim is a competitor. It comes up short in many areas compared to OOTP, I will grant you, but it's still out there. It's not like I brought up Mogul or Baseball Stars or something.

Because they are two different games with two different code bases (neither of which are open source). They have nothing to do with each other when it comes to functionality. You closed your post with "are you going to wait for PureSim to do this first for you to do the same?". That just makes no sense, it was essentually wadding up your post and throwing it in the trash.

Yes, there are issues with the way pitchers are created and yes there are issues with the way they are used, but nothing PureSim has or doesn't have effects this.

You can talk about what you like & don't like from PureSim, you can talk about what it has that you would like OOTP to have but the assumption that something PureSim has or could eventually have should automatically be implemented into OOTP is an attitude/opinion that ticks me off because it shows a complete lack of knowledge in regards to how a game works or is developed.

05-26-2006, 01:06 PM	#104
Joe Morgan Minors (Triple A) Join Date: Oct 2003 Posts: 273	bump because quite frankly this is about 50 times more important than Laughing Boy and his non-hidden hidden ratings. __________________ Clutch ability and veteran leadership are two of the most important intangibles that it takes to win.

05-26-2006, 03:31 PM	#106
No Pepper All Star Starter Join Date: Jan 2003 Location: Milwaukee, WI Posts: 1,570	Here are the rates I got after eliminating the pitchers who were used in both roles, for 2007 the following year. Code: RP SP H/9 9.18 9.20 ERA 4.27 4.37 K/9 6.56 6.52 BB/9 3.95 3.12 HR/9 0.89 0.91 Looking at players that appeared as both a starter and reliever, there were about 7-10 low endurance guys (~3) that had 20+ starts and less than 100 innings. All of them had ERAs under 4. Jacksonville had 2 of these guys in their rotation (Abernathy and Blezard) and went on to win their division, going 100-62 in 2007. Without split totals for starts it's hard to tell if these guys would change the results. It must be an AI-related, whether it be the situation relievers enter the game (start of inning or not) or in transactions. I've seen starters dropped from the ML club and relievers on the roster (low endurance CL or MR) taking over the rotation spot, whereas IRL, you promote a AAA SP or make a trade for a SP to fill the void. All that said, ERA isn't a good measure for the skills of players, if you're talking about ratings boosts.

05-27-2006, 01:20 PM	#108
Joe Morgan Minors (Triple A) Join Date: Oct 2003 Posts: 273	I definitely agree that this is the way things are done in modern baseball. I also clearly remember seeing a guy pitch a no-no in 1990 right after the regular season started (there was an owner's lockout that year, which didn't affect the regular season schedule but drastically shortened spring training) and get taken out in the 7th for one Mike Witt, who wrapped things up. Thing is, this is one of these issues that I think would already be taken care of in spring training - you know, the thing that the game doesn't simulate. Historically speaking, it is very, very rare for pitchers to consistently be taken out after 4 innings. It's not so rare for relievers to be tried as starters, and for those relievers to fail miserably, which is why I go back to the simple little boost for anybody who throws in relief. I know, I know, it doesn't completely explain Sparky Lyle or Dennis Eckersley, but it's like 70% of the explanation and IMO that's quite an effective "band aid" if you ask me. __________________ Clutch ability and veteran leadership are two of the most important intangibles that it takes to win.

05-29-2006, 01:10 PM	#110
Joe Morgan Minors (Triple A) Join Date: Oct 2003 Posts: 273	Bump! Can I get a TRU DAT in here? __________________ Clutch ability and veteran leadership are two of the most important intangibles that it takes to win.

05-30-2006, 11:26 PM	#114
andymac Hall Of Famer Join Date: Feb 2002 Location: Effingham, IL Posts: 5,725	Why must people continually **** up any point they are trying to make by bringing PureSim into the discussion. __________________ June Madness: Links FTB: andymac

05-26-2006, 11:05 AM	#102
Gastric ReFlux Hall Of Famer Join Date: Oct 2002 Location: Member #3409 Posts: 8,350	We've got a way to compare OOTP versus real life. http://reports.ootpbaseball2006.com/...00_0_2006.html We can use the pitching registers to get a decent approximation of starting results versus relief results. Toss out any pitchers who have made both starts and relief appearances, then use the ones who have appeared in strictly one role to get an idea of the results we're seeing. We'd be interested in the following numbers: ERA so we need IP and ER's. Hits allowed Strikeouts Walks HR Convert IP, Hits, HR, SO, and BB to a rate stat of per 9 innings. I won't have time this afternoon to work on this, but it could be very helpful for us to crunch these numbers and see how OOTP is performing compared to real-life which can be found at Retrosheet.

05-29-2006, 11:04 PM	#112
tysok Hall Of Famer Join Date: Aug 2003 Posts: 4,925	At this point I wouldn't agree with anything. This thread has sparked my curiosity enough that the first thing I do with my test league is to run some seasons and get some break downs with true complete data over a few years time. Then I have to pull like data from reality to compare it to. A cursory look into the relief vs starter question prompted me to want more information, which will take a lot of time that I haven't had. I pulled off 15 pitchers from 2005 that pitched 15+ starts with a relatively equal amount of relief appearances... only 1 of them fit every assumption we had (relievers give up less hits, less HRs, and have less ERA, while giving up more walks and getting more Ks). Only 6 of them fit 3 of the 5, and only 2 fit 2 of the 5. (Those numbers aren't exact, I'm pulling it from memory - it may be way off, but it was not nearly a pretty picture) At the same time, a low number were traditionally relievers, a slightly higher number were traditionally starters, but most were rookies. That's a very bad sample size to draw any conclusion whatsoever from. But that's an important data set to analyze in the question of whether any pitcher in relief should have a ratings boost. Basically for this question to get a resolution it's going to take a lot of work with the in game data to discover where the problem is, or if there is a problem. Cursory examinations and small sample sizes lead to nothing. That being said, I can give you a TRU DAT that if a pitcher is used as a starter (given more than just a few uses), it is rare for them to consistently pitch only 4 innings the majority of the those starts. I have a hard time believing (hopefully) that the reports put up are from a real league running full tilt with the game we'll be getting on the 31st. at least one of those teams had only 1 pitcher shown as being in a starter role.

05-30-2006, 10:16 PM	#113
Syd Thrift Hall Of Famer Join Date: May 2004 Posts: 10,655	Can we at least get a final "no, we don't care about this" or "yes, we think that TangoTiger might have a modicum of authority on the matter and we'll at least look at it" about this issue? Again, I understand that my post count is such that I'm just going to get flamed for this, but this is a legitimate issue. Are you going to wait for Puresim to fix this for you to do the same?

05-30-2006, 11:56 PM	#117
Syd Thrift Hall Of Famer Join Date: May 2004 Posts: 10,655	? I have delivered no ultimatums, nor have I called OOTP trash. Surely, however, you must appreciate that TangoTiger knows what he's talking about, and in fact probably knows more than you do about baseball from a sabermetrician's standpoint. I know he knows a lot more than I do, anyway. Anyway, to add something more than infighting... others have been claiming that this is entirely or mostly due to the AI's poor use of bullpens. However, a look at starter v reliever stats throughout the years indicates that this is not the case. Here are the ERA differences between reliever and starter. A + before the number indicates that relief ERA is lower, and the number is a percentage: 2005: +6.1% 2000: +6.8% 1995: +5.6% 1990: +10.9% 1985: +6.2% 1980: +13.1% 1975: +10.8% 1970: +4.0% 1965: +8.2% 1960: +6.6% 1957: +1.1% (this is the earliest consecutive year Retrosheet has this data available) A brief look shows that since 1960, at least, relievers have consistently outperformed starters by about 6% or so. Going all the way back to 1922, when relievers were generally only used to mop up for bad outings by starters, the split is 4.9% in favor of starters (4.30 ERA for relievers vs. a 4.10 ERA for starters) - which is still a smaller split than what OOTP metes out.

05-31-2006, 12:04 AM	#118
tysok Hall Of Famer Join Date: Aug 2003 Posts: 4,925	I'm not surprised to see no answer from any high authority here. There's yet to be proven that a problem exists, for right now it's just people discussing an issue that may or may not even exist. I put out the numbers for a season from OOTP5, there was no problem in that season. By the way, my cousin did another 3 seasons the same as I did, 10 seasons from the one I analyzed, 5 years later, and then the year after that... they all came up correct. That doesn't mean a great deal since OOTP2006 is a 'complete rewrite', and that's data from 2 (or 3 if you count 6.5) versions ago. The data from the new game that we could get on the reports (ignoring anyone pitching in both rolls) was not pretty, but several things lead one to believe it's not an anylization of good data. After we've had the game (finished version) a few days and looked over several seasons of data we can all speak on this subject better... and if that data shows inaccuracies I'll jump on with these other guys... then we can worry about why the higher powers are or are not bothering to answer. Until then this thread is just conjecture, and if I were the developer I wouldn't bother ansering or posting about conjecture either when so many people are waiting for my game, some are even staying up until release (and at least one is even staging a breathing strike until release)

05-31-2006, 12:17 AM	#119
Syd Thrift Hall Of Famer Join Date: May 2004 Posts: 10,655	I disagree. You showed that in your v5 league that the issues Gastric Reflux brought up were not causing starters to get fewer decisions. Perhaps (at least in that one league) the reliever messiness wasn't acting as a "comeback code". You never showed that relievers had lower ERAs and peripheral stats than starters do as in real life. Good work, though. The discussion only benefitted from your study.