I still prefer not to restrict playing time for people who performed reasonably close to expectations in limited time. However, if enough people feel strongly enough against that, we can use Greg's suggestion of limiting playing time to 110% of their actual time.

Well, I could go either way, though I'd prefer to set the bar higher than 110%. What happens if, say, a first stringer gets injured and all that's left is a backup with limited time?

I'd also like to take a guy with 107 PA and make him a platoon partner against LHP. I'd think that sort of thing would be kosher.

But we've got Joel Pineiro, too, who had 11 starts at a high-quality level. Of course I'd like to have him for the season.

I was talking about the '89 A's, but that's neither here nor there. The '53 Dodgers had a run differential of 266, the '55 team had a differential of 207. Please tell me you're not basing your entire argument on ordinal rankings.

No, I'm not. If you want a full analysis comparing the two, I can provide it. '53 might be better, might not. And actually, what am I arguing for, since I don't particularly care which Dodgers team is included.

If we nitpicked each and every selection, we'd be here until the end of time.

Well, I could go either way, though I'd prefer to set the bar higher than 110%.

Yeah, if injuries are set to "random" or whatever, we're probably going to need PT maximum
set higher than 110%. We don't have 40-man rosters and a free agent pool to help us with
PT during an injury streak.

But if we set it too high, then some teams can take advantage of a limited-PT guy with a
great season.

I was walking into a Staples Store and Dom Dimaggio was going out

Well, I was at Dunkin Dounuts and I saw Jow DiMaggio there having a dounut and coffee. Well, not really.

Yeah, if injuries are set to "random" or whatever, we're probably going to need PT maximum
set higher than 110%. We don't have 40-man rosters and a free agent pool to help us with
PT during an injury streak

Especially us 15 man roster types.

Karl I was walking into a Staples Store and Dom Dimaggio was going out...

Did Dom throw that speedball by you, make you look like a fool boy?

I don't think their should be playing time maxes... just a reasonableness check on the manager profile.

I don't think their should be playing time maxes... just a reasonableness check on the manager profile.

I agree with this, but we need more clarification about what playing time is trying to do:

Obviously we don't want a guy who was 1 for 1 with a double in the starting lineup. But for the purposes of this exercise, do we want Joel Pineiro pitching a full season? He had 11 starts, because he started the season in the minors.

If you are looking for yet another person with no experience with DM, and are the 1910 A’s are free, and no one else wants em, I'd be happy to lend a hand. Although, I understand if you are looking for someone who saw them play in person and - at least in my current life - I have not.

I don't think their should be playing time maxes... just a reasonableness check on the manager profile.

I'm kind of struggling with that issue. I'm managing the 1905 Giants. Roger Bresnahan is the 2nd or 3rd best hitter on the team, and my leadoff hitter. But he played only 104 games, and caught only 87, not because he was injury prone, but because catchers didn't play near as much back then as they do today. He was by any definition, a regular. While having him catch 75-80% of the games, like modern catchers do, would be a huge boon to this team, I feel would be unrealistic. For some teams, like the 1910 A's and 1912 Red Sox this won't matter much as there's not much difference in the hitting ability of the two catchers. UCCF will face this though with Johnny Kling (136 OPS+)/Pat Moran(82).

I suggested to SG that he set one of the three starting pitchers to have the backup be a personal catcher, and to also have the backup catch the DH games while Bresnahan DH's.

I agree with this, but we need more clarification about what playing time is trying to do:

It's a grey area. Like I've said, I don't want to restrict the playing time too much, although Pineiro starting 35 games helps the Mariners considerably. I'd suggest that for pitchers we reduce their starts to at most double of their actual starts. For position players like Canseco, who had missed a lot of time but had established a level of performance consistent with their typical output, I am fine with letting them start full-time. For Shane Spencer types who greatly outperformed their expectations, I'd want to restrict their playing time to 125% of their actual.

For the smaller rosters, all these considerations can be thrown out IMO. There's just not enough flexibility to be overly restrictive about playing time.

biggs, the 1910 A's are yours. No DMB experience is really needed.

Although, I understand if you are looking for someone who saw them play in person and - at least in my current life - I have not.

The 1910 A's?

I don't think our demographic stretches that far.

As with others, I've never played DMB, but I'm willing to help out. I'll take either the 27 or 39 Yankees (I think they're both still available).

I think SG makes a good case regarding part-time players. You can't let Shane Spencer slug .900 all year, but you can let these teams present their "best reasonable selves." It doesn't necessarily pay to have managers shove the teams exactly into their actual at bat totals. Because after all, I'll bet almost everyone who is following this already knows which couple of teams they think had the best stats.

For players who were injured during the season (Canseco '89), you can just give them high injury ratings in the game (I understand from a previous post--I'm not a Diamond Mind player).

But for players who were traded, I would think you should have to take the "finished" end of the trade. If the '89 A's are leading off with Rickey, they shouldn't have Plunk and Caderet coming out of the pen or Polonia coming off the bench.

But for players who were traded, I would think you should have to take the "finished" end of the trade.

I agree with this completely.

Updating post #88:

1905 Giants - Miserlou
1906 Cubs - UCCF
1909 Pirates - Greg Tamer
1910 A’s - biggs
1912 Red Sox - kevin
1927 Yanks
1929 A’s - Gaedel, Elster, Backe
1939 Yanks - Dan Turkenkopf
1942 Cards - Patrick W (e-mail sent to him, unsure if it's confirmed)
1946 Red Sox - karlmagnus
1948 Indians
1953 Yanks - Potch1214
1954 Indians - Ernie C's Elbow
1955 Dodgers - Screw All
1961 Yanks - SJ
1967 Cards
1968 Tigers - fra paolo
1970 Orioles - Rauseo
1974 A’s
1975 Reds - Tamer
1978 Yanks - RB in DC
1984 Tigers - Brady Anderson (them or the '98 Braves, but I scrolled to this one first & so am slotting him here)
1986 Mets - Howard Megdal (asked about them, but hasn't officially claimed them)
1989 A’s - DCA
1995 Indians - Bernal Diaz
1998 Braves
1998 Yanks - SG in Atl
2001 Mariners - Werr

I reckon' four teams totally unaccounted for with a few just penciled in.

I'd suggest that for pitchers we reduce their starts to at most double of their actual starts.

That's fair enough.

For the smaller rosters, all these considerations can be thrown out IMO. There's just not enough flexibility to be overly restrictive about playing time.

So are you suggesting that UCCF and I should play our better catchers most of the time?

I am really surprised that the 1998 Braves remain unclaimed. I remember quite a bit of support for them in the voting threads.

But for players who were traded, I would think you should have to take the "finished" end of the trade. If the '89 A's are leading off with Rickey, they shouldn't have Plunk and Caderet coming out of the pen or Polonia coming off the bench.

Great point. I didn't think about this for some reason.

Perez-to-Walker-to-Lee, not in time (p8p) has already done the '27 Yanks, so we can cross them off the list above. Dan T, I'll let you take the '39 Yanks.

So are you suggesting that UCCF and I should play our better catchers most of the time?

Sure, I think that's fine.

OK. In that case, ignore the "When Mathewson pitches" lineups I sent you this morning. I'll rest Roger by having him DH.

I am interested in managing one of the remaining teams. As a Pirates fan, I have no real rooting interest in any of the 4 that are left, but I'd claim the '48 Indians, if possible.

SG,

I assume each team gets a 25-man roster? Does the manager also select the roster?

I was thinking full rosters actually, although Chris Derosa's point about post-trade teams is good. I can inactivate players who were not in the organization at the end of the year, and players who got less than say 10 AB or pitched fewer than 5 innings I guess (except for teams with fewer than 25 players).

Another DMB question:

For pitchers, what gives the best idea of their Diamond Mind performance: DIPS/Peripherals, ERA, OPS against, something else?

Since other DMB newbies have chimed in, I'll take the 1948 Indians if they're still available. My dad was 13 years old and living in Cleveland at the time, so maybe he can help me set them up.

You can probably ignore ERA almost completely, and look at component ERA:

Component ERA (CERA) = (D*9)/IP

If CERA≥2.24 than CERA-.56
If CERA<2.24 than CERA*.75

* D = (A*B)/C
* A = H+BB+HBP
* B = (((H-HR)*1.255)+4*HR)*.89)+((BB-IBB)*.56)+(HBP*.56)
* C = BFP (Batters facing Pitcher)

In other words, look at the DIPS peripherals but also hits allowed.

Kiko, too late. KevinApps already took the '48 Indians. '67 Cards, '74 A's and '98 Braves are still available.

Thanks, SG.

I'll take the 67 Cards. I was just learning to walk.

If managerial duties are restricted to setting up the team and you can send me the template, I'd be happy to do the 90-win, Alvin Dark-skippered, overrated-by-Pythagoras 1974 Oakland A's.

But for the purposes of this exercise, do we want Joel Pineiro pitching a full season? He had 11 starts, because he started the season in the minors.

I wasn't around for the original thread, but it seems to me that the point was to try to determine a ranking of the teams as they actually were, not as they could have been.

In other words, it seems to me that if a manager/GM made some of what you would consider questionable lineup choices, you should stick with them. When you push the envelope of "what if Pitcher X" pitched the full season or "what if Hitter Y" batted 7th instead of 1st, at that point you are talking about a different team than the one that actually played.

Of course, this is DMB and it is by definition different, but still I would think that the goal should be to try to replicate what was actually done, not necessarily injecting your subjective opinions on how the teams could be better.

Of course, this is DMB and it is by definition different, but still I would think that the goal should be to try to replicate what was actually done, not necessarily injecting your subjective opinions on how the teams could be better.

Well, I think that's a subjective question. Is the greatest team the greatest collection of personnel or the greatest combination of personnel and decisions?

By the way, I wouldn't mind seeing it simmed both ways.

By the way, I wouldn't mind seeing it simmed both ways.

That's a really good idea actually. I can set up another run that strictly matches all actual playing time. Then we can argue about which is the "official" run.

Then we can argue about which is the "official" run.

I like this a lot. Then we can also figure out which of us are better or worse than the real managers. :-)

Well, I think that's a subjective question. Is the greatest team the greatest collection of personnel or the greatest combination of personnel and decisions?

Put it this way -- the 2001 Mariners (like every other team) made the list of "greatest teams" because of both the players and the decision-process.

If you try to divorce the two, you have some entity that isn't what the 2001 Mariners actually were. That entity might have played better than the actual 2001 Mariners . . . or they might have played worse . . . but they would have been a different team.

Of course, my comments don't just apply to you, Mug; the same goes for everyone else who seems to be tweaking their teams. If you want to run the team as you would like (rather than they way it actually was), it's sure to be fun, but it loses the point of a "sim."

BTW, I should ask SG: Am I correct in thinking that the only thing you want folks to do is set up manager profiles (as opposed to being involved in the game play itself)?

Well, manager's profiles, lineups, and pitching staffs. If someone wants to manually play 1000 seasons they have my blessing, but that was not the plan.

If you want to run the team as you would like (rather than they way it actually was), it's sure to be fun, but it loses the point of a "sim."

I'm not sure I agree. I mean, you're already changing what actually was by running this sim. Facing different teams would change the decision-making process.

I don't think it's any less valid to treat a team as a collection of players than it is to consider them a collection of players and playing time, etc. But I think both are good questions which is why I'd like to see it both ways.

I'm not sure I agree. I mean, you're already changing what actually was by running this sim.

Sure, but when you start saying things like "I'm going to assume Joel Piniero was called up in April, rather than June [or whenever]," that isn't really far from:

* the '68 Tigers saying "I'm going to assume that Mickey Stanley played SS the entire season, rather than Ray Oyler";

* the '86 Mets saying "I'm going to assume that Kevin Mitchell played SS instead of Rafael Santana"; or even

* the 1989 A's saying "I'm going to assume that Jose Canseco never got hurt."

These may be interesting questions, but if the point is to sim the teams as they actually were, it ruins the exercise.

Well, we're not assuming Pineiro played the full season. I think we all agreed that was no good.

These may be interesting questions,

Actually, to be more precise, I *do* think they are interesting questions. I just think that if you construct a sim based on those assumptions, you're no longer simming "the 1968 Tigers against the 1989 A's" -- you're simming "the 1968 Tigers with Mickey Stanley as a full-time SS against "the 1989 A's with Jose Canseco healthy all season."

I don't feel strongly about it, but I guess I'd lean toward a little discretion for the manager. A point in favor of discretion is that all these teams will be losing more often than they did in real life. The 1998 Yankees, going into the year knowing what they knew about Chuck Knoblauch and Derek Jeter, quite sensibly led off with Chuck Knoblauch. Knoblauch struggled to hit .260 all year and Jeter blossomed, but when you come out .730 ball, you don't mess around with stuff like that.

Now put the team in a league of elite teams, and the '98 Yanks come out playing .530 ball. Knoblauch would be on a shorter leash--the whole batting order would be open to question. Chad Curtis's playing time would certainly be scrutinized more closely, and perhaps Joe Girardi's too.

I guess my point is that keeping the team profile historical but making the competition fictional just potentially creates a more illogical fiction than is necessary.

I guess my point is that keeping the team profile historical but making the competition fictional just potentially creates a more illogical fiction than is necessary.

Yes, but there's a big difference between deciding lineup construction and pretending a major injury to a major player doesn't happen.

Miserlou,
It was my understanding that the game had an injury rating to input for each player. Any player injured during the season in real life ought to be given a proper injury rating, so that the game will injure them from time to time. Hopefully the game is sophisticated enough that a guy who missed half the season (Canseco '89 for instance)can be given a rating that will injure him for about half the season, because that is certainly a real part of the player's performance you shouldn't wish away.

Karl,
on the Slaughter play you reference with Walker up; Dimaggio,by his account and corroborated by Halberstam"The Teammates",was frantically trying to move Culberson over toward Left from the top step of the dugout...Cronin is never mentioned with regard to outfield positioning...it turned out that Dimaggio was correct...the best video of this play is espn classic's,the video narrative synopsis of the Halberstam book..the camera POV is to the right of Homeplate,10 to 15 feet high with a clear view of Culberson to Pesky. If you see this video it may further confirm your contention with regard to Pesky. Simply,the team was not ready to attack Slaughter's attack...

Did you notice Cronin used four different leadoff hitters in the 46' World Series...

I'm proud to say NONE of Cronin's 4 leadoff hitters are the ones I've proposed to use, which would be Pesky and DiMaggio, as I presume you'd agree. Cronin was a good SS but a SERIOUSLY stubborn and unresponsive manager, from what one's read -- if one couldn't beat his performance in a sim one would be in real trouble.

I don't think restricting to exact actual playing time is remotely realistic, particularly for those with restricted rosters (having Mathewson pitch every 3rd start against the 1998 Yankees isn't realistic either; pitch counts in 1905 were about 2/3 those of 1998.) How could you possibly allocate innings correctly for the pitchers, for example? -- you'd end up either with unused aces or with dummies pitching in pennant situations. The manager in the real season didn't KNOW how much he could get out of each player, and had the flexibility to use more if he felt like it.

I do agree that a player who was injured in the "real" season like Canseco 1989 should have a high injury factor in the simulation. Sometimes he'll get injured, sometimes he won't. You're thus playing the '89 As without the "luck" of the actual season pre-programmed. The simulation will be a little unrealistic in using more heavily those who had good seasons; perhaps one could program a 50% reversion to career mean when usage went beyond the historical pattern, but you will always have the situation where a rookie is brought up in September, gets a few at-bats and later turns into Ted Williams -- whatever decision one makes on usage of that rookie will be historically realitic. It should all even out between teams.

"realitic" should be "unrealistic." Sorry!

Are active rosters restricted to 25 players? If so, how do we farm and activate players throughout the season? If not, isn't that an unfair advantage for teams with more players, especially with regards to arms in the bullpen?

I guess what I'm trying to avoid is a situation where, say, Willie Bloomquist of 2002 becomes a starting player all year (or even, to a lesser degree, Pineiro of '01)... but still maintain a situation where a team can platoon Ed Sprague with John Olerud against LHP.

I feel that using *all* the same decisions in a completely different league is like following directions to the grocery store in a completely different city. With a completely different situation, the decisions can't be the same--nor should they, if you feel that a team is a collection of talent and not of decisions.

Nevertheless, I'm strongly in favor of running it both ways. This is a situation where we can have our cake and eat it too.

Are active rosters restricted to 25 players? If so, how do we farm and activate players throughout the season? If not, isn't that an unfair advantage for teams with more players, especially with regards to arms in the bullpen?

In 124, SG says he planned on having everyone available except post-trade players and people with less than 10 PA or 5 IP.

The trouble is, with 25 players, injuries could be a real problem. I wonder if you could fix it by capping rosters at 28-30 or so. And maybe limit the number of pitchers.

Mug, the "same innings as in real life" sim can't be done if you have a random element for injuries in there. Even without that, if you get more extra inning games than in the original season (or the original season was 154 and you're playing 162) you will end up without any players at all for the last few games.

Are active rosters restricted to 25 players? If so, how do we farm and activate players throughout the season? If not, isn't that an unfair advantage for teams with more players, especially with regards to arms in the bullpen?

I wanted to run with full rosters and then allow the managers to set the playing time for one set of runs. I think we should do two sets of runs. One with the rules/setup we've discussed in this thread (increased PT at manager's discretion and within reason, custom manager's profiles, custom lineups and rotations, etc.,) and with injuries on, then one using auto-generated depth charts and rotations/pitching staff usage based on actual playing time and no injuries.

This would answer two questions.

1) Which team had the most talent when at their best
2) Which team had the best "season" when factoring in actual playing time and performances.

We can set up the teams that made mid-season trades with actual transactions so that they would have the players that were on the roster at the time they were on. It'd be a bit more work to set up, but the payoff is probably worth it.

That's a good point. I'm assuming if we do the same-playing-time sim, we'll turn off random injuries, pro-rate the 154-game players, and add about 5 or 10 percent to everyone.

Even without that, if you get more extra inning games than in the original season (or the original season was 154 and you're playing 162) you will end up without any players at all for the last few games.

My understanding is that in Diamond Mind using depth charts based on actual playing time, playing time would be handled on a percentage basis. It is not going to be exact. Take a team like the 1927 Yankees for example. Pat Collins was the starting catcher in 46% of the games, with Johnny Grabowski in 34% and Benny Bengough in 20%. Diamond Mind would take this into account while running, and try to make sure that each gets in that percentage of games, whether you played 100 games, 154, 162, or 300. It's not on a games/ab/innings basis, but it is pretty close to it.

Does DM substitute within games to maintain the percentages, or does it use a composite player that is 46% of player A and 54% of player B? Or does it play whole games only and then use all 25 players on the 162nd game to make the numbers work?

The other problem with this one as a '46 RedSox trainer is that I'm stuck with dozy managerial decisions (other than on batting order) of the quality that used 4 leadoff hitters in the WS, none of which were among his 2 obvious leadoff hitters. Teams that historically had competent managers who figured out quickly who could play and who couldn't will be at a competitive advantage. I'll bet the '46 Sox do relatively a LOT worse with Cronin player selection than with optimised or near-optimised player selection.

Does DM substitute within games to maintain the percentages, or does it use a composite player that is 46% of player A and 54% of player B? Or does it play whole games only and then use all 25 players on the 162nd game to make the numbers work?

I'm pretty sure percentages are based on starts. In-game strategy is based on what makes baseball sense I think (pinch-hitting for platoon advantages, etc.,) and not on just getting the playing time in.

Now put the team in a league of elite teams, and the '98 Yanks come out playing .530 ball. Knoblauch would be on a shorter leash--the whole batting order would be open to question. Chad Curtis's playing time would certainly be scrutinized more closely, and perhaps Joe Girardi's too.

The other problem with this one as a '46 RedSox trainer is that I'm stuck with dozy managerial decisions (other than on batting order) of the quality that used 4 leadoff hitters in the WS, none of which were among his 2 obvious leadoff hitters. Teams that historically had competent managers who figured out quickly who could play and who couldn't will be at a competitive advantage. I'll bet the '46 Sox do relatively a LOT worse with Cronin player selection than with optimised or near-optimised player selection.

In other words, it seems to me that if a manager/GM made some of what you would consider questionable lineup choices, you should stick with them. When you push the envelope of "what if Pitcher X" pitched the full season or "what if Hitter Y" batted 7th instead of 1st, at that point you are talking about a different team than the one that actually played.

These are three different posters. I don't have a strong preference here, but I'm going to set the '89 A's profile such that they are most like the actual team. Lineups will be consistent with TLR's. Storm Davis is going to be in the top 4 of a 5-day rotation, not Curt Young. And Weiss will get plenty of PT. Other than that, though, they're pretty well optimized though.

It's just that I don't think that it's realistic -- or fair -- to use the players' actual stats to make managerial preferences, because the real manager would not have access to these, and might not realisticly make the same judgements about the value of OBP vs speed, for example. Instead I strongly suggest that you use a qualitative metric of "how good the player is" including reputation and prior years' stats and I will very disappointed if we use Bellhorns and Giambis as leadoff hitters, when in actuality there is a convential wisdom leadoff hitter, who actually hit leadoff, on the team. In other words, I agree 100% with the last paragraph.

But for players who were traded, I would think you should have to take the "finished" end of the trade. If the '89 A's are leading off with Rickey, they shouldn't have Plunk and Caderet coming out of the pen or Polonia coming off the bench.

Good point. I might have missed this ... but luckily I still have Eck, Honeycutt, Burns, and Nelson. This is another argument against strict PT limits ... if/when Eck or another RP gets hurt, the A's will need Corsi to step up if they don't have Plunk or Caderet.

I'd suggest that for pitchers we reduce their starts to at most double of their actual starts.

Perhaps for someone like Pineiro (a mid-season callup who was very good), we should have them available starting July 15 (or some other appropriate date). This is different from the trade acquisition, because we have only partial year stats for Pineiro (i.e. can't add his Tacoma MLE's to his 2001 line the way we can add Rickey's NYY stats). However the 2x starts (I would prefer 1.5x starts, except when necessitated by injury) seems like it might be better. Just throwing this idea out for consideration though.

For 60 years, fans of the '46 Sox have wondered how good the team could have been if it had a grownup as manager. Cronin was SPECTACULARLY bad, and was only there because he was part of Yawkey's racist Irish mafia. It's bad enough not having Jackie Robinson and Campanella because of the fool, but to subject such a wonderful team to Cronin's ideas of who should bat leadoff, when the right answer is so obvious to any normal observer that 95% of managers, however non-sabermetric, would have got it correct is simply a travesty.

It's just that I don't think that it's realistic -- or fair -- to use the players' actual stats to make managerial preferences

This is a valid point. On the other hand, adjustments are made based on the year's stats. I'm sure the Mariners didn't plan to start 2001 giving Bret Boone 200 AB in the 3 spot.

I don't have a strong preference here, but I'm going to set the '89 A's profile such that they are most like the actual team. Lineups will be consistent with TLR's.

I really think that if we're going to do two sets of sims, you shouldn't do that. We'll have one set that mimics TLR and reality and another that mimics our best conception of what's optimal. If both questions are worth answering, let's answer both. If you play a realistic lineup against a bunch of optimal lineups, we end up answering neither question.

but to subject such a wonderful team to Cronin's ideas of who should bat leadoff, when the right answer is so obvious to any normal observer that 95% of managers, however non-sabermetric, would have got it correct is simply a travesty.

I think you're making too much of this. It's not like he had Pesky batting cleanup and DiMaggio batting 7th or something. They were 2-3 respectively for every game of the series. Does it matter that much that they were 2-3 instead of 1-2?

OK, here's what I think we should do.

1) An official run, which will use the actual percentages of playing times of all players. Oakland '89 will get about 30 innings of Greg Cadaret, 250 PA of Jose Canseco, and 400 PA of Rickey Henderson. Pitching staffs will be set up to mimic actual usage in that season. Depth charts for each position will be set up so everyone plays the same amount that they did in the season in question. Injuries will be turned OFF, so they will not be a factor. Roster size will not be an issue as players will only appear in roughly the same amount of time that they did during that season.

2) An experimental run, using the manager's profiles that have been/are being provided to me, and with increased playing time for part-time players at the manager's discretion, within reason. Injuries will be on, teams will be set up as post-trade (ie, Henderson on the A's, Cadaret, Plunk and Polonia are not).

Does anyone have any issues with this dual-setup?

Forgot to add a little results based analysis. For the series, Red Sox leadoff hitters compiled a .322 OBP (admittedly, boosted tremendously by Moses' 4-5 in a losing effort in game 4), compared to .258 for Pesky and .310 for DiMaggio.

I suggest the manager's profiles run should be official and the other experimental. If others disagree I suppose we'd better take a vote on it.

I think the actual percentages run is historically very dubious indeed (some of these teams clinched about mid August for example, so played all kinds of scrubs the last few weeks.)

Miserlou, based on actual outcomes Bobby Doerr had an OBP of ,.458 in the Series, but Cronin batted him sixth -- he wasn't one of the 4 leadoffs either.

So, when does the stimulation begin?

I think the actual percentages run is historically very dubious indeed (some of these teams clinched about mid August for example, so played all kinds of scrubs the last few weeks.)

That's an excellent point. I'd imagine the '95 Indians played a lot more scrubs in the final weeks than the '42 Cardinals or the '78 Yankees

I suggest the manager's profiles run should be official and the other experimental. If others disagree I suppose we'd better take a vote on it.

Voting is going to extend this too long I fear. Spring training starts in 3 weeks. How about we don't classify either as official or un-official, and people can make their own determinations to the validity of the two sets of runs?

So, when does the stimulation begin?

I'll kick off the actual PT/no injuries run today. I'm still waiting to hear back from manager volunteers for some of the teams for the customized run. I'm guessing it will take about a week to run through the first 1000. Someone suggested looking at one of the runs in detail so I will probably do that as well.

Miserlou, based on actual outcomes Bobby Doerr had an OBP of ,.458 in the Series, but Cronin batted him sixth -- he wasn't one of the 4 leadoffs either.

The point is, the obvious leadoff hitters were Pesky or DiMaggio, and they were outperformed buy the imfamous 4. So had Cronin did what is obviuos to many, it likely wouldn't have made a bit of difference. Doerr batting 6th is probably only one slot worst than optimum.

You obviously know more about this team than I, but looking at both in-season performance, and established ability, the top half of the order should have been:

Pesky
DiMaggio
Williams
York
Doerr

in reality it was:

Imfamous 4
Pesky
DiMaggio
Williams
York
Doerr

Given that the 4 outperformed both Pesky and DiMaggio, I see no reason to blame the series loss on Cronin's lineup choices.

I don't have a strong preference here, but I'm going to set the '89 A's profile such that they are most like the actual team. Lineups will be consistent with TLR's.

I really think that if we're going to do two sets of sims, you shouldn't do that. We'll have one set that mimics TLR and reality and another that mimics our best conception of what's optimal. If both questions are worth answering, let's answer both. If you play a realistic lineup against a bunch of optimal lineups, we end up answering neither question.

Dan, if it's going to be two sets of sims -- one with actual PT fixed and another with us free to manage the teams however we see fit, someone else can do the 89 A's, because I don't want to. I think the first is ridiculous, for the reason that karlmagnus posted in #65, without even mentioning that I think it's a lot less interesting to observe, and the second I think defeats the purpose of the excercise, since we're introducing a very significant component of DMB user skill into the sim. If we're going to be free to juggle rotations, lineups, etc, then one person has to do all of them, or we're really just have a DMB tournament with an all-time list of teams. Neither of these options is what I thought I was getting into.

Personally, I took on the 1989 A's with the expectation that I would be trying to reproduce TLR's actual managerial tendencies as best I could, with a few judgement calls on how to adapt to other styles of play, like no DH, and maybe cutting off the bottom 10% of his quirky moves as a *slight* accomodation for a higher standard of competition. But even that might be best gleaned from playoff lineups and usage patterns.

If this -- a good faith effort to present the actual 89 A's for across-era elite competition -- is not what's wanted, then someone else can take the managerial profile. I have absolutely no interest in using this tournament to showcase my DMB prowess and personal lineup-making and reliever-usage philosophy, or to prove that I can out-manage the certified genius. And I fear that's what this is turning into.

we're really just have a DMB tournament

well duh ... what I meant is that it would be like me vs SG, not the A's vs the Yanks.

On the other hand it's just, you know, fun.

Why don't you guys use these teams' postseason rosters/lineups as the benchmark, with the understanding that:

(1) You'll allow for players who play most of the season but are injured immediately before the playoffs (e.g. Jim Rice in 1975)

(2) You'll use pitching rotations that reflect regular season as opposed to postseason practice (i.e. modern teams will have 5 man rotations, not 3 or 4)

And then turn injuries off.

Component ERA (CERA) = (D*9)/IP

If CERA≥2.24 than CERA-.56
If CERA<2.24 than CERA*.75

* D = (A*B)/C
* A = H+BB+HBP
* B = (((H-HR)*1.255)+4*HR)*.89)+((BB-IBB)*.56)+(HBP*.56)
* C = BFP (Batters facing Pitcher)

You don't even need to make it that difficult. CERA = 31 * OBP * SLG is much easier to calculate and works even better (since it includes more data). Of course, if you really want to go after accuracy, look at base runs allowed.

For relievers, I tend to correct for platoon differences (facing one side a lot more in real life than they will in DMB) and intentional walks (some situational guys give up a ton of IBB, throwing off their OBP allowed).

I have absolutely no interest in using this tournament to showcase my DMB prowess and personal lineup-making and reliever-usage philosophy, or to prove that I can out-manage the certified genius. And I fear that's what this is turning into.

For Christ's sake, I don't think that's what anyone is talking about. I, for one, have never played DMB in my life. Nor do I have any interest in proving myself to be a super-genius. I think you'd be a lot more persuasive if you accepted that people might have legitimate differing points of view about how this exercise should be run and we aren't all just here to feed our massive egos.

Other than that, I think you make some good points. But some of us are saying that we want to see the 1989 A's as a collection of talent, independent of TLR's managing--because that's what question is more interesting, not because TLR is a moron.

Frankly, I think the collection of talent on these teams is going to be the primary factor and you'd see similar results any of the three proposed ways--especially between yours and the actual playing time version. I already suggested bumping the playing time numbers 5 or 10%--that should be *plenty* to accomodate playing scrubs at the end of the year.

Why don't you guys use these teams' postseason rosters/lineups as the benchmark, with the understanding that:

(1) You'll allow for players who play most of the season but are injured immediately before the playoffs (e.g. Jim Rice in 1975)

(2) You'll use pitching rotations that reflect regular season as opposed to postseason practice (i.e. modern teams will have 5 man rotations, not 3 or 4)

And then turn injuries off.

That's a really good idea.

Presumably you'd use regular season stats for your postseason rosters, right? I'm not playing the 1939 Yankees with Ted having an elbow injury and batting .205!

Presumably you'd use regular season stats for your postseason rosters, right? I'm not playing the 1939 Yankees with Ted having an elbow injury and batting .205!

I can live with a 0.00 staff ERA

I know Brady is managing the 84 Tigers, but if he'd be still willing to take the 98 Braves, I'd love to set up the Tigers (with a little tutoring from a DMB-savvy Primate). Let me know either way...

Other than that, I think you make some good points. But some of us are saying that we want to see the 1989 A's as a collection of talent, independent of TLR's managing--because that's what question is more interesting, not because TLR is a moron.

But Cronin is?

Anyway, my complaint is that neither of the options represents "what would happen if the XXX team entered this 162 game season against 27 of the best baseball teams ever."

Ignoring the manager replaces the actual manager of the team -- and I feel a bit like Backlasher here, I think we're in danger of forgetting -- or at least not recognizing -- that we have collections of performance, not talent, and that the performance was sampled from talent with the manager very much involved in that process and we can't just assume that the performance would be the same, given the same talent, if usage differs from what the actual manager did.

The hard PT limits take all the luck out wrt to injuries. The 89 A's -- continually using my team as an example because I know them best -- would not go into the season knowing they'd only get a 1/2 season out of Canseco (well, actually they would, because he broke his wrist in the spring, but during the winter offseason they'd expect him to be playing every day). Making injuries random allows a strong when healthy team to dominate in some seasons, and fall short in others. That's interesting. I'd love to see a "best case" for each of the 28 teams, where everything went right for them. It also brings out the value of a strong bench and redundancy that can cover for injuries, even if they didn't have to in the actual regular season.

It seems to me that in each of the two suggested bunches of seasons we're taking something out from the obvious natural case of "what if the teams just played?" and if we're going to be doing multiple sets of runs, why leave out the obvious one?

Dan -- your explorations of how to deal with Olerud/Edgar in 60-some non-DH games is an example of where judgement is needed. But redoing a batting order or a bullpen because the manager didn't use his optimally seems to me to be adding too much of "us" to the sim. If a guy sucked and got a lot of PT, you shouldn't be able to just hide him, if you have two catchers who split time and one was better than the other, you shouldn't be able to play him every day. And lastly I don't want my DMB skills or superior knowledge (or just as likely, lack of these) to affect the results. I've not played DMB either.

DCA, the "as it was played" sim gives you the effect of Tony LaRussa's decisions -- if you want to take that one as "real" you're free to. Personally I prefer to go the other way, because I want to see how the wonderful '46 Sox would do without the miasma of drunken Celtic racism that infested the Boston front office at that era. If I was managing the '67 Sox, yes I'd probably be happy to go with whatever Dick Williams did. But a Cronin run '46 team is almost as handicapped as if you made it play with Ted Williams' '46WS wrist injury.

Running both sims allows you to measure two things (i) you vs. actual manager -- probably humbling against LaRussa, hopefully positive against Cronin and (ii) how good was the talent on the team if injuries had been random, rather than exactly as they were. Supernaturally healthy teams (like the '01 Mariners IMHO) will lose out on the second sim compared to your '89 As or any other team that won in spite of some injuries.

It seems to me that in each of the two suggested bunches of seasons we're taking something out from the obvious natural case of "what if the teams just played?" and if we're going to be doing multiple sets of runs, why leave out the obvious one?

I guess I don't feel it's any more obvious or correct than the others. I'm fine doing three (God knows how SG feels). I'm fine omitting the "true playing time" too. I do think if we want to do it your way, then basing things mainly on the postseason lineups, as suggested, makes the most sense. That's what the managers did when they were playing big games against good opponents and had a decent handle on their players' capabilities.

But whatever we do, we're not going to end up with "what really would happen." And I do find it interesting to think... Hey, the 2001 Mariners probably would have been a better team if they'd platooned Olerud and Sprague. I wonder what that group of players would be capable of, best case. Now I'm not Lou Piniella and I don't have to worry about Olerud complaining and so on and so forth, and I'm not saying I'm smarter than a manager, but the fact is that we're running a sim that only considers real, isolated baseball performance anyway--so how good can these teams be from a flat-out isolated performance standpoint? We still won't get a perfect answer, of course, but I think it will be interesting nonetheless.

I'm fine doing three (God knows how SG feels).

SG has a headache.

Basically, there's no perfect way to run these, and running 50 different versions of the same thing cheapens it in my mind. There's good and bad in every idea that's been suggested here, so I'll let you all hash it out. I'm definitely running the actual playing time sim, since that is the easiest one to set up. As for running one or two or three more, here's what I see on the table:

1) Use the real-life manager's decisions as closely as possible, but with random injuries
2) Optimize the talent of the team disregarding how they were actually used
3) Playoff rosters with allowances for late season injuries (basically best 25 man roster for each team), with no injuries and with regular season pitching

Did I miss anything?

4) Actual PT limits +10%, no injuries -- or is this part of #3?

4) Actual PT limits +10%, no injuries -- or is this part of #3?

Not really. Make that # 4.

I think we should also run a sim where the front office realizes they're playing against such good teams. For example, the 2001 Mariners probably would have acquired Bonds to play LF if they knew this would be their competition. And they would have re-signed A-Rod instead of letting him go to Texas.

But a Cronin run '46 team is almost as handicapped as if you made it play with Ted Williams' '46WS wrist injury.

You keep saying this. I don't mean to be argumentative here, and I'm sure you can point to other things, within the scope of this sim, that you or any manager could have done better, but if you're just talking about lineup construction, I don't see it. How much worse can:

Random RF
Pesky
DiMaggio
Williams
York
Doerr

as a top 6 be than

Pesky
Dimaggio
Williams
York
Doerr
Randon RF

which is presumably the optimum top 6?

Sure it's worst, but since the random RF actually performed better than Pesky or DiMaggio, I don't think you can blame the series loss on that decision.

Are there other aspects of Cronin that can be improved upon (again, within the scope of this sim)? Did he ph for Williams a lot? Did he play Rudy York at SS? Did he sacrifice in the first inning all the time?

4) Actual PT limits +10%, no injuries -- or is this part of #3?

To me that's part of #0 (The one SG is definitely running), but maybe he wants to do that one exactly. I just thought it would cover the effects of late-season callups and what-not, but I don't think it's a big deal at all.

Miserlou, my preferred order is:

Pesky or DiMaggio
Williams
York
DiMaggio or Pesky
Doerr
etc.

One of the joys of this simulation would be to see if this would in fact be better.

SG, I'm happy with your alternative 2 or 3 (prefer 2, with simple rule on e.g 1 for 1 not being allowed as a perfect player) and will feel free to ignore the others. For the '46 Sox I'm not competent to set up 1 and wouldn't want to. The actual PT one is presumably manager free (or will you use our lineups and pitching usage?)

4. looks to me like an accident wating to happen, because the machine won't know when to pick which player. It is realistic only if you can choose players for each game, which might be interesting but I can't possibly commit to -- would be 1/2 hour per day for 162 games.

To me that's part of #0 (The one SG is definitely running), but maybe he wants to do that one exactly. I just thought it would cover the effects of late-season callups and what-not, but I don't think it's a big deal at all.

Okay that makes sense ... from his post, I thought SG was dropping that, and while it's not my favorite for "official" -- that would be option 1 or option 3 -- and I wanted to make sure it hadn't fallen off the table.

1) Use the real-life manager's decisions as closely as possible

I think it bears pointing out that I am doing this for a team that I followed (although I was only 10, so not with as critical an eye as I do now), while others for teams they don't know as well. But that's fine, a good faith effort is good enough -- look at some playoff lineups and B-R's games by position on the team page and try it from there. I wouldn't expect Ron Hassey as Bob Welch's personal catcher detail from those who didn't watch the team play.

OK, then, SG, I'll send you two MPs for my two teams. One set of MPs for second sim and the other set for the third sim, as you described in post 2-84 above.

Miserlou, my preferred order is:

Pesky or DiMaggio
Williams
York
DiMaggio or Pesky
Doerr
etc.

One of the joys of this simulation would be to see if this would in fact be better.

Well, that's all well and good, but I doubt you'd find one manager in history who would have batted Williams 2nd and Pesky 4th. And IMO, such a radical lineup goes beyond what is trying to be accomplished here. Addressing the obvious anomaly that is "Random RF" batting leadoff is one thing, a radical re-thinking of the traditional lineup, one that has rarely if ever been seen in baseball history, is quite another.

Alou this year wants to bat Bonds second; the two cases are pretty well exactly analogous, IMHO.

I'll take either the '97 or '03 Marlins.

Wink, wink.

Well, that's all well and good, but I doubt you'd find one manager in history who would have batted Williams 2nd and Pesky 4th. And IMO, such a radical lineup goes beyond what is trying to be accomplished here. Addressing the obvious anomaly that is "Random RF" batting leadoff is one thing, a radical re-thinking of the traditional lineup, one that has rarely if ever been seen in baseball history, is quite another.

I agree 100%. This is what I was worried about.

Alou this year wants to bat Bonds second; the two cases are pretty well exactly analogous, IMHO.

Alou thought aloud that he might do it. That's a lot different than actually doing it, much less keeping it up for a full season.

One of the joys of this simulation would be to see if this would in fact be better.

I agree with Miserlou. Here's the problem -- the point of the exercise (at least as I have read it) isn't to determine whether you can manage the '46 Red Sox better than they were managed in real life; it's to compare the '46 Red Sox to their counterparts as they were actually managed.

Maybe your lineup will be more optimal for the '46 Red Sox, but it doesn't really answer the question on whether the actual team was better/worse than the others. (Putting aside the fact that it's DMB, of course.)

it's to compare the '46 Red Sox to their counterparts as they were actually managed.

Then the actual playing time simulation should suffice for this part of the exercise. The other simulation(s) will be for those who want to answer different questions. Those questions are no more or less valid than anyone else's.

to determine whether you can manage the '46 Red Sox better than they were managed in real life; it's to compare the '46 Red Sox to their counterparts as they were actually managed.

I think we're doing both.

And Cronin might have been a lousy manager, but I'd guess that who he batted first or sixth is way less significant than all the myriad other ways he might have screwed up the team. Batting order doesn't make that much difference, remember? It's not like he was batting the pitcher 4th and Williams 9th.

I see these little tweaks we're making to the teams as just that. It's a fun "what if" experiment, and nothing more.

About 10 years ago USA Today ran an all-time greatest teams tournament using Diamond Mind. (I'm sure some others like it have also been done.) I believe the '39 Yankees beat the '53 (yes, '53) Dodgers in the final.