Forced Results: an attempt at a more realistic sim.

[uruguru]

I've done a lot of simming with OOTP and my focus has always been on "what if" scenarios and how history would be different. As a Houston native, my team of choice has always been the Astros and so like many players my favorite franchise is always the most interesting area for me to sim.


What I'd like to do with this thread is document the most realistic sim I can imagine with OOTP. What I will do is start in 1962 and play as manager with "historical transactions" and "historical lineups" turned on (1-year recalc). This means that this won't be an interesting timeline of me (as GM) building a championship team by acquiring star players.


For additional realism beyond the standard league settings, I will be save-scumming all of the AI-AI games so that their game results (win or loss) match actual history. The only difference in this league from actual history should be the results of my managed games. Part of me wonders if these "forced results" would be an interesting feature for OOTP to add in some future release but we'll see if it feels more immersive. This page is what I will use to determine the outcome of AI games: https://www.baseball-reference.com/l...schedule.shtml


For Houston games, I will be creating my own lineups from the players made available to me by the AI GM, but with the 20/20 hindsight of how these players performed in real life. That's sort of the "What If" scenario that will make it different from actual history and hopefully maintain my interest in the scenario.


I'll try to make it interesting, but who knows?

[uruguru]

It's two days before Opening Day, I have a 25-man roster to build and there are 36 players available for me to use. These are all players whose first team in 1962 was Houston so the game gives them to me at the beginning.

However, 2B Billy Goodman was not signed as a free agent until May 15th, and P George Brunet was not acquired from Milwaukee until the day after that. These are minor transaction-related errors that I can easily work around, so that leaves me with 34 players to choose from: 22 batters and 12 pitchers.

Since I will be varying from the historical lineups, this means that some lesser-used players in 1962 will get more playing time from me. As a result, I need to identify which players served time in the minors in 1962 and use their ratings from an equivalent save that includes minor leagues. The idea behind this is to more accurately reflect the ability of these players so that, for example, 1B Dick Gernert is not rated just on his 24 major-league at bats, but instead on the additional 346 at-bats he had in the minors. Players with no minor-league time will have identical ratings in both saves, but players with minor league time will have their milb ratings universally ported to this save.

It turns out that ratings were adjusted in this way for 12 batters and 4 pitchers. Some players improved and some became worse. If not around 50/50, it was pretty close. I would like to tell you that it was a wash, but the 1962 club kept so many atrocious hitters on the roster that any change in lineup composition will improve the team notably.

One final constraint I am going to apply is that any player who did NOT play in the minors in 1962 must stay on the active roster. I am doing this to attempt to model players who could not be demoted without clearing waivers. This means 10 of the 15 position players are non-negotiable, as are 8 of the 10 pitchers.

So let's build our active roster...

CATCHERS
From the uploaded image, you can see that Hal Smith is the only catcher who is average (3/5) defensively. Jim Campbell and Merrit Ranew are both mediocre (2/5), but Campbell is best hitter of the 3. So Smith is the starter and Campbell is the backup. Ranew is demoted to reserve.

INFIELDERS
Now to the infield, where it gets exciting! First base is easy... we need two guys and Norm Larker and Dick Gernert are offensively too much better than Pidge Browne, so he gets booted to the reserve roster. Larker bats left and Gernert bats right and they are close offensively, so they will be platooned to maximize their offensive output.

At SS, the two best defenders are Bob Aspromonte and Ernie Fazio. Fazio is wretched offensively, so he's gone. Aspromonte played 3B for the real 1962 team but he came up in the LA organization as a SS and occasionally played SS for Houston, so I feel justified in considering him for this full-time position. George Williams has better fielding ratings than Aspromonte and would probably be better defensively in OOTP, but he played 2B exclusively in the minors (I checked), so he cannot be considered for it. Don Buddin was the team's real SS and acquired from Boston when the team traded away their #1 expansion pick Eddie Bressoud, but he was a bust offensively and also famously error-prone at the position (his nickname was "Don Booting") so he will not be getting the starting nod from me. Aspromonte it is.

At 2B, the real team played Joey Amalfitano, who was a premium pick in the expansion draft. However, his defense was average at best and rookie George Williams is both great defensively and comparable offensively. Therefore, Williams gets the starting spot from me and Amalfitano will back him up. Williams is going to strike out a lot, and that's probably what slowed his progress in real life, but I'm a sabermetric time traveler from 2023 so I'm less concerned about strikeouts. Honestly, this decision is not that consequential since the GOAT at second base will be arriving next season and rendering all of this moot.

Finally at 3B, you'll notice that catchers Ranew and Campbell are on the list. Ranew played the position sparingly, but Campbell started his professional career as a third baseman (moving to C later) and played 3B extensively in 1962 for the minor-league club. Slotting him as a 3B is not a stretch by any means and his bat is too good to be used solely as a backup C. So Campbell is the unicorn who will be starting games at both 3B and C, even though, like Williams, he is going to rack up the strikeouts.

Buddin will start when the C, 3B or SS need a rest so he will get a lot of playing time. Amalfitano will sub for Williams at 2B on his rest days. Bob Lillis is an adequate emergency backup at 2B/3B/SS and better overall than J.C. Hartman, so he gets the final infield spot. Normally that's one infield spot too many but our 3B will be catching 40 games a season so we need the additional depth.

OUTFIELDERS
Now to the outfield! Al Spangler is clearly the leader of the pack here, and only one of two outfielders who is good defensively in center. Slotting him in at CF is an easy pick. I believe he played CF for historical team, but moved to LF when they acquired Carl Warwick. Spangler has a weak arm and so never played much RF.

Roman Mejias is possibly the best hitter on the team (4/5 contact) but is atrocious with the glove, so we will hide him in LF where he is at least average. He played RF for the historical team.

The next two best hitters are mirror images of each other. LHB Dave Roberts and RHB Johnny Weekly will platoon in RF for the time being. Roberts was a Panamaniam player with a long productive career in the minors before playing in Japan. Weekly was a Rule 5 minor league pick from San Francisco who ultimately did not pan out. These guys will be holding down the spot until Rusty Staub arrives in a year or two.

Veteran Jim Pendleton will provide backup with the ability to play respectably at any of the three outfield positions. Al Heist takes the final outfield spot not by virtue of his ratings, but because he did not play in the minors at all in 1962 so I will need to carry him on the active roster. Unfortunate.

PITCHERS
Now comes the easy part, the pitchers. We need 10 pitchers and 8 are sort of required to be on the active roster including the entire starting rotation: Ken Johnson, Turk Farrell, Dean Stone, Bob Bruce and Hal Woodeshick. Veteran All-Star Bobby Shantz will be the swingman (spot starter) out of the bullpen and will be used to draw a crowd by starting on Opening Day, just as in the real world. Bobby Tiefenauer and Jim Golden round out the required eight. Jim Umbricht will be the stopper as he is clearly the best reliever on the team and will get the call anytime he's rested and the game is close. He's a sidearmer, so he's especially effective against RHB. Al Cicotte will round out the bullpen.

The only other two pitchers available, Dave Giusti and Dick Drott, represented a considerable dropoff in pitching quality so the pitching decisions were all pretty straightforward. Fortunately, the pitching is a real strong point for this expansion team.

The hardest part of all of this was the typing. Time to finally start the season!

[uruguru]

ok, Opening Day was kind of hilarious. In the real world, the Colt .45s throttled the Cubs 11-2 en route to a three-game sweep. In OOTP, I actually somehow topped that as Houston strolled to a 13-0 beatdown. When we went up 9-0 after 6 innings, that triggered my internal "mercy rule" and I sent out reserves for 5 of the 8 defensive positions (not catcher, nor the two platoons). The idea of this was to ensure my starters stay rested as much as possible.

Here's how this all worked, mechanically. First, I save-scummed all of the AI-AI games by going to the MLB Scores page and "Quick-Playing" each game. When the result matched the real world... save. If not, kill OOTP and restart to try again. This process surprisingly did not take very long at all, compared to the actual game played, so I think it's a workable plan for players who are simming game-by-game and want that extra accuracy.

Once I entered the game, I switched "substitution" control to me and left the rest to the AI manager. I then loaded in my desired schedule (vs the RHP), my starting pitcher and then started the game. I then just played it out in one-pitch mode until I wanted to make a substitution. Pretty straightforward.

Obviously, my roster construction is so amazing and dominant that I can just declare victory now and dispense with the rest of the sim. (joke!)

Our offense is on a pace to score 2106 runs and allow none, so I can safely project a 162-0 season before getting swept by the Yankees in the World Series.

[Brad K]

"all of the AI-AI games so that their game results (win or loss) match actual history."

I had no idea such a thing was possible.

[uruguru]

Originally Posted by Brad K

"all of the AI-AI games so that their game results (win or loss) match actual history."

I had no idea such a thing was possible.

Well, I have to save-scum to get it. Here's how I described it in a subsequent post:

Here's how this all worked, mechanically. First, I save-scummed all of the AI-AI games by going to the MLB Scores page and "Quick-Playing" each game. When the result matched the real world... save. If not, kill OOTP and restart to try again.

The idea, of course, is to make the rest of the league more closely mirror their actual performance. If the Yankees won the 1962 AL pennant, then why should my managing of an NL team change that?

[Brad K]

Oh, I assumed you were doing some kind of editing.

[uruguru]

Originally Posted by Brad K

Oh, I assumed you were doing some kind of editing.

That would be cool, but I suspect that the way the game works is that the internal database of player data (AB, H, etc) is updated real-time as games are played. These database updates are kept in memory until you save, at which point the database is written to the file system. So once a game is simmed, there's no way to edit a game result because then all of the individual player stats from that game would be inconsistent with the results. It's also why I suspect having the "forced results" feature would be really difficult for the dev team to implement since they don't have the awkward luxury of making it work by killing and restarting the game.

[thehef]

Originally Posted by uruguru

Well, I have to save-scum to get it. Here's how I described it in a subsequent post:


The idea, of course, is to make the rest of the league more closely mirror their actual performance. If the Yankees won the 1962 AL pennant, then why should my managing of an NL team change that?

Man, that's a lot of killing & re-starting of OOTP. I'm not sure I'd have the patience to do that...

One thing I've done to get a certain postseason matchup or force a tie (at the end of the season, for tiebreaker testing or whatever) is just make sure that the team I want to lose starts (and leaves in) their worst pitcher and sits all their best hitters. And if that doesn't work, at some point I'll start intentionally walking enough guys to make sure the team I want to lose, loses. But I don't think that would work for you.

I guess if I was attempting something like what you're doing... well, I'll just use an example:

- Let's say I've save-scummed to the point where 10 of the 14 games played that day are complete and have the results I want. At this point, I would stop save-scumming and just roll with the results that the game gives me. And let's say that two teams that were supposed to win that day - let's say Boston and Texas - lost in OOTP. What I would do is then make sure that the next time they lost IRL, I made sure in OOTP that they were first two teams to lose. A bit less save-scumming. And in the end, does it really matter - for your purposes whether Boston & Texas won IRL on 6/29 but lost in OOTP, but then lost IRL on 7/1 but won in OOTP? Same difference...

Or another thing you could do: Sim out all of April (except for your team's games) and then see where things stand. If some teams need to lose a little more or win a little more to be closer to where their RL record was, then start making adjustments in May. Then when you get things to where all teams are pretty close to their RL record, start simming again for awhile until you see a few teams that need some adjustments... And keep going like that for awhile... Just a thought to maybe save you some time...

[uruguru]

I appreciate the advice however I am of the mindset that if I am going through the trouble of doing it 90% of the way, I might as well do the last 10% as well. The plan is to finish April and re-assess.

[uruguru]

Thanks to a Greenspanian episode of irrational exuberance, I have already finished the 15 games for the month of April. Here is a recap of the games for those interested, but I still want to do comparison of my lineup changes vs. the real world in a followup post sometime tomorrow.

No transactions to report for April.

[darkcloud4579]

Originally Posted by uruguru

I've done a lot of simming with OOTP and my focus has always been on "what if" scenarios and how history would be different. As a Houston native, my team of choice has always been the Astros and so like many players my favorite franchise is always the most interesting area for me to sim.


What I'd like to do with this thread is document the most realistic sim I can imagine with OOTP. What I will do is start in 1962 and play as manager with "historical transactions" and "historical lineups" turned on (1-year recalc). This means that this won't be an interesting timeline of me (as GM) building a championship team by acquiring star players.


For additional realism beyond the standard league settings, I will be save-scumming all of the AI-AI games so that their game results (win or loss) match actual history. The only difference in this league from actual history should be the results of my managed games. Part of me wonders if these "forced results" would be an interesting feature for OOTP to add in some future release but we'll see if it feels more immersive. This page is what I will use to determine the outcome of AI games: https://www.baseball-reference.com/l...schedule.shtml


For Houston games, I will be creating my own lineups from the players made available to me by the AI GM, but with the 20/20 hindsight of how these players performed in real life. That's sort of the "What If" scenario that will make it different from actual history and hopefully maintain my interest in the scenario.


I'll try to make it interesting, but who knows?

Interesting premise and I like the idea of making it as relevant for yourself to maintain interest in playing it. Good luck!

[uruguru]

brief reminder of the effect of forcing results

[uruguru]

Attached are the April stats for position players and pitchers.

Let's compare the two specific positional substitions I have made from the real team:

Real: 2B Joey Amalfitano: 620 OPS for 1962 season
OOTP: 2B George Williams: 800 OPS through 15 games

Real: SS Don Buddin: 629 OPS for 40 games in Houston
OOTP: 3B Jim Campbell: 479 OPS through 15 games (ugh)

Conclusion: a wash.

So why is the OOTP team three games ahead of real life? The pitching. The starters threw 5 shutouts in the first 15 games, a pace that is both ridiculous and unsustainable. However, let's not be too hasty by ruling out that it could also be due to my elite managing skills. Maybe I missed my true calling?

[uruguru]

I just realized after playing the May 1st game that "Position Player Fatigure" is set to "None" if you playing with real-life lineups. Makes a lot of sense! Fortunately, I had so many rain days in April that I never played more than 5 days in a row so fatigue was never an issue. However, going forward I need to make a homebrew system to avoid an unfair advantage. Here it is:

Starting a game at any position besides C or 1B costs 1 point of fatigue. Catching costs 1.5. First base costs 0.75. If playing both games of a doubleheader, the second game is double cost.

Players start with a pool of fatigue points based on their age:

under 25: 10 points
25-29: 9 points
30-34: 8 points
35-39: 7 points
40+: 5 points

Players cannot start a game if that start would push them below 0. Resting a complete game resets a player back to full. Pinch-hitting does not disrupt this rest as long as the player doesn't stay in the game and play a position. This means that reserve players and platoon players should basically never get fatigued so I dont have to worry about them.

My starting catcher Hal Smith can start 5 games in a row. He has 8 points, and each game costs 1.5 points.

2B George Williams, SS Bob Aspromonte, and 3B Jim Campbell have 10 points and can start 10 games in a row, although Campbell will have to catch as well since he is the backup catcher.

CF Al Spangler has 9 points and can start 9 games in a row.

LF Roman Mejias has 8 points and can start 8 games in a row.

My 1B and RF are platoon so they should be fine unless we have to face a long stretch of RHP.

This system is simple enough for me to track and plan around. Since the OOTP fatigue system is not available, it's all I have.

edit: extra-inning games longer than 12 innings will count as 2 games for players who play at least 13 innings.

[Brad K]

That's good. Innovation on the fly!

[uruguru]

We've won 3 of the first 4 in May and are sitting at 13-6, 5 games ahead of the real-world performance. Pitching is still going gangbusters.

Either RNG is still really favoring us right now, the advantage of using lineups based on ratings vs. historical is bigger than I expected, or OOTP secretly puts the thumb on the scales for player teams (something I've never actually tested). I still think it's RNG. In a great RNG season, this team would be at best .500.


Ok, here's the shutout countdown, something I enjoy doing.

Only two teams in MLB history have ever gone an entire season without being shut out. Here's how the shutout countdown is going in this sim:

8 teams still alive: Astros 13-6, Athletics 11-12, Angels 9-9, Dodgers 14-10, Yankees 12-6, Mets 4-14, Giants 19-5, Cardinals 12-7

The rest:

9. 5/2 Twins, by H.Brown of BAL
10. 4/29 Indians, by J.Kralick of MIN
11. 4/28 Pirates, by S.Koufax of LAD
11. 4/28 Red Sox, by R.Herbert of CWS
13. 4/27 Senators, by B.Stafford of NYY
13. 4/27 Orioles, by E.Rakow of KCA
15. 4/23 Braves, by J.Moeller+bullpen of LAD
16. 4/22 White Sox, by E.Rakow of KCA
17. 4/21 Phillies, by B.Bruce of HOU
18. 4/14 Reds, by G.Perry+bullpen of SFG
19. 4/10 Cubs, by B.Shantz of HOU
20. 4/9 Tigers, by B.Daniels of WAS

[uruguru]

It's May 9th and there were two significant transactions.

On 5/7, we traded P Bobby Shantz to St. Louis in exchange for OF Carl Warwick and P John Anderson. irl, Houston used Shantz as a starter (3 games, all starts) and St. Louis will use him exclusively as a reliever. Warwick is a good bat upgrade and the real team used him in CF. I will keep Spangler in CF and use Warwick in RF (ending the platoon there). It's the best solution since Spangler doesn't have the arm for RF and I want to continue to hide Roman Mejias' defensive shortcomings in LF.

Anderson looked like a good arm for the bullpen until I incorporated his minor-league ratings. Suddenly he looked mediocre and will be kept on the reserve roster. But who will replace Shantz in the bullpen?

On 5/9, we purchased the contract of P Don McMahon from the Braves. That's who.

Combined, these two moves are really good because we clearly upgrade our RF spot and get a younger quality arm in the bullpen (Shantz was 36). Kudos to the GM!

edit: hmmm, Al Cicotte is on the DL. No notice of that. Well, I guess Anderson will be in the bullpen!


============================

also, I think I figured out what is going on with the better than expected record. Perhaps I have some good RNG. Perhaps playing by ratings is a real advantage when the AI has to play with RL lineups.

But honestly, I think maybe it's the defensive changes I made. In the real world, SS Don Buddin and 2B Joey Amalfitano both turned out to be average defensively, 3 on a scale of 1 to 5. In their place, I put Bob Aspromonte at SS (4/5) and George Williams at 2B (5/5), and neither of them came with an offensive downside.

It's true that we human players have the advantage of hindsight, but I think the defensive improvements combined with the already-good pitching is currently giving me a bigger than expected advantage. Or maybe it's just early and everything will revert to the mean. /knock on wood

Attached is the team information page to give you an idea of how the team is doing now.

[darkcloud4579]

You can also use the simulation engine to simulate games though I realize it's not the same thing, it'd prevent you from having to constantly do this if you're just trying to produce results and see what happens.

But yeah OOTP not having a "force win" button (which is understandable given how it's used) makes this really tough and slow, but it seems to be working for you so far.