The Sandy Koufax/Nolan Ryan problem

[Syd Thrift]

Originally Posted by uruguru

Pitcher stamina, as used in OOTP, is a usage stat masquerading as an ability rating. This is really the easiest way to describe the problem. The rating doesn't actually reflect the physical stamina of the pitcher but simply how the real-world manager used that pitcher.

As a result, OOTP managers are required to use pitchers exactly how they were used in real life. Want to keep Dennis Eckersley or John Smoltz in the rotation? Too bad. Their stamina gets recalculated to a low value because the real-world manager had different ideas and so you have to put him in the bullpen too (choosing the "career average" for stamina does not solve this problem, btw)

Then, if you have Eckersley in the pen and one of your starters gets injured, you can't put him back in the rotation when obviously the real manager always had that as an option.

This also creates havoc for expansions because players are recalculated based on their next season ratings before the draft. For example, the real 1962 Dodgers did not protect their reliever Turk Farrell. He was drafted by Houston moved into the starting rotation (this happens a LOT in historical sims). But in OOTP, the 1962 Dodgers suddenly see Farrell as a starter (based on his 1962 stamina) and decide to protect him and put him in the rotation. If you are playing without minor leagues, expansion teams in historical sims are basically screwed because there have no real options for drafting starting pitchers. This is why people will run historical sims and see expansion teams with 10-152 records.

I have played this game way too much.

To the Eckersley question, if you’re running a game that started in, say, 1980… well, for one thing Eckersley might never make the transition to the bullpen if you don’t have recalc turned on. In fact, it’s more likely that he won’t than that he will. But if he did, you’d have the ability to switch him back to start. IRL, too, Eck didn’t start a game after 1987 and started only 2 that year. The theoretical starting of him is, well, only in theory, as Tony LaRussa et al never saw him as a potential spot starter for the last decade or so of his career. I’m not really sure what the game should do in these situations (that’s not a way of saying “hey, I guess it’s ok then”; I mean I don’t know what it should do).

All that said, the weirdness between starting and relieving is one reason I prefer fictional play. Even with fictional, there are guys who will never ever be effective as starters no matter what their stamina or whatever because they only throw 2 pitches. Conversely though there are a whole bunch of guys who can do both if needed and who might be better if you had them concentrate on their top 2-3 pitches in relief (which is sort of what Eckersley did IRL I guess). There’s one guy in particular in the league I’m doing who was a mediocre to bad starter who is now ranked as the 4th best closer in all of baseball. Maybe he’ll get another crack at the rotation eventually but I feel like this is his career now.

I feel like overall stamina isn’t perfect by any means. One issue I seem to see a lot, for another example, is low control / high stuff guys with low endurances but enough pitches to start. What are you really supposed to do with those guys? They aren’t the classic sinker/slider or fastball/splitter style relievers so they get starts but then they can’t really give you innings even when they are effective because 6 innings causes them to throw like 120 pitches. I wind up manually upping stamina but of course that’s a workaround. Maybe a better way of handling this is to not have stamina do a lot to ratings when pitchers get tired but massively increase injury chances instead. The Nolan Ryan example here is kind of telling; even though there may be studies indicating his K rate remained the same throughout games, he was also rather famously placed on strict pitch counts throughout his 30s to prevent injury.

[Syd Thrift]

Oh right… one big thing I’ve seen, speaking of injuries, and perhaps this is why you don’t see great pitchers achieving true greatness, is that injuries compound injuries and after like 15 years in the game practically every starting pitcher who throws a lot winds up Wrecked. The fact is that some pitchers IRL never seemed to get truly dinged, or came back from injuries to still be fine and not injury prone. I don’t know that you can predict who’s going to be a Randy Johnson in advance vs. who’s going to be a Teddy Higuera, but there’s something possibly still unknown, that causes some guys to fall apart by the time they’re 30 and others to… not.

[Rain King]

Originally Posted by uruguru

Pitcher stamina, as used in OOTP, is a usage stat masquerading as an ability rating.

It should be noted that this is only the case when looking at "real" players. The game has to figure out how to assign Stamina and in order to have the players represent their "real-life" selves goes by their Usage.

I think there would be far more complaints if a different formula was used and it caused pitchers to be used differently than they were historically.

In fictional leagues you will see players of wide-ranging stamina in all roles with their number of quality pitches playing a much larger role than stamina in their usage.

[uruguru]

Originally Posted by Hoiles

Do you play with recalc on? If not, historical players who had cups of coffee or struggled during their first season (like those two) start out with low ratings/high potentials, but often don’t pan out later in their careers. The random “dice rolls” with the development engine don’t give them the needed boosts to get them to their real-life peaks.

You pretty much have to play with recalc on (I use 5-year averages, doubling the current season) if you want any semblance of historical accuracy. If you don't care, that's fine, too.

If you don't have recalc on, then you have to deal with issues like HOF starter Phil Niekro spending his career as a bullpen specialist.

[David Watts]

Originally Posted by uruguru

You pretty much have to play with recalc on (I use 5-year averages, doubling the current season) if you want any semblance of historical accuracy. If you don't care, that's fine, too.

If you don't have recalc on, then you have to deal with issues like HOF starter Phil Niekro spending his career as a bullpen specialist.

If you want to avoid Niekro spending his career as a bullpen specialist, set the "base pitcher stamina on" to Entire Career.

[Hoiles]

Just to reiterate my previous point, without recalc on, even if Koufax and Ryan are imported as five star potential starters, they start out as 1 star pitchers considered by OOTP as essentially a Todd Van Poppel or Brien Taylor "can't miss" prospect who could easily miss due to bad development/coaching/injuries etc. Either that, or you'd need to turn talent randomness completely to zero or something.

[BarneyRubble]

OOTP probably doesn't use 'outcomes' when playing historical. I think OOTP uses stats to generate 'ratings', and those ratings generate 'fictional stats', which generate 'outcomes'. This is fine and dandy, but that might not mirror historical outcomes. I don't think it's supposed to. I could be wrong.

Pitching statistics in baseball aren't always going to create accurate 'ratings'. Each one is like a piece to a puzzle and all are useful to get a clear picture. Jim Palmer is a good example. I've read articles calling him 'overrated'. One article said his control wasn't great. They concluded this by a single 'stat'. These are the kind of guys that name their kids FIP and OPS. They apparently never saw Palmer pitch or heard him talk about pitching. And that's the problem with leaning on too few statistics, and not on outcomes. All statistics (minus OPS) are good in some aspect, and all statistics are flawed in some aspect. Palmer didn't try to K everyone, he wanted them to hit the ball (there goes his stuff and FIP). He'd walk a batter if need be (there goes his control rating). However, his K rate went up significantly in the playoffs (so did his BB rate). So creating 'ratings' from statistics is NOT always going to be accurate. It's just not. Palmer pitched to WIN and no team ever lost because they gave up too many FIPs. Palmer started over 250 less games than Nolan Ryan, yet their CG and SHO are nearly identical. Let that sink in.

Growing up, I marveled at guys like Fergie Jenkins, Bob Gibson, Tom Seaver, Jim Palmer, etc... These guys knew how to pitch. I suppose OOTP could have a 'Cy Young' rating (make it hidden) that replicates how well a pitcher knows how to pitch. This might help those who play historical and it might be fun trying to find the next Randy Johnson.

As for stamina, I've always thought it was a decent way to simulate the pitching. IRL, teams 'stretch out' their starters. Power pitchers don't necessarily have less 'stamina' than a finesse pitcher, they just expend more energy per pitch. Relievers obviously expend more energy as they typically don't pitch more than an inning or 2. I believe OOTP adds 'stuff' to relievers. It gets the job done. However, OOTP could just build it into the game engine and not mess with the 'stuff rating'. They could then equalize the 'stamina' ratings (not by RP and SP) and leave the 'stuff rating' alone. Pitchers could then be used in one of three modes, starting, long relief, short relief. The pitcher would know, coming into the game, how he is to pitch. Relievers know how many innings they are expected to pitch. The engine would adjust 'stamina' and 'stuff' (internally) accordingly. Just a thought.

[uruguru]

Originally Posted by David Watts

If you want to avoid Niekro spending his career as a bullpen specialist, set the "base pitcher stamina on" to Entire Career.

That's really a band aid to the actual problem, though. If you have a starter who spent half of his career as a reliever, then you end up with a guy who can barely go 6 innings.

[uruguru]

Originally Posted by Hoiles

Either that, or you'd need to turn talent randomness completely to zero or something.

If you are interested in accurate historical sims, you should generally turn player development completely off (if players are forced to retire in their historical years), or keep it on with TCR to 1, development rate to 0, and aging rate to at least 1 if players are not forced to retire.

If you turn development off, you obviously have to force players to retire or else some guy who had a great final year in the majors (i.e. Koufax), will pitch like that until you end the sim.

[MathBandit]

I'm certainly glad that OOTP bases it's algorithms and ratings on what actually happened and not try and decide which pitchers had a magical ability to WIN without having to make good pitches. If we're going down that route we may as well just play The Show where some guys have magical 'Clutch' abilities because I guess they just slack off the rest of the time or something.

[uruguru]

Originally Posted by MathBandit

I'm certainly glad that OOTP bases it's algorithms and ratings on what actually happened and not try and decide which pitchers had a magical ability to WIN without having to make good pitches. If we're going down that route we may as well just play The Show where some guys have magical 'Clutch' abilities because I guess they just slack off the rest of the time or something.

You do realize that Baltimore in the early 70s (e.g. when Palmer pitched) focused on having a great defense (Brooks at 3B, Belanger at SS and Blair in CF) and encouraged their pitchers to put the ball in play.

Strikeout rates then were not what they are now, so the idea of dominating the mound as a pure FIP pitcher was literally something pretty much only Nolan Ryan could accomplish.

[MathBandit]

Originally Posted by uruguru

You do realize that Baltimore in the early 70s (e.g. when Palmer pitched) focused on having a great defense (Brooks at 3B, Belanger at SS and Blair in CF) and encouraged their pitchers to put the ball in play.

Strikeout rates then were not what they are now, so the idea of dominating the mound as a pure FIP pitcher was literally something pretty much only Nolan Ryan could accomplish.

I do. And that by definition means the pitchers succeeded in part due to having an elite defense behind them and so weren't as good as pitchers who had better ability to miss bats. It also means that compared to pitchers who weren't as fortunate to have a great defense behind them, their surface numbers (Wins, ERA, etc) were likely inflated relative to their actual pitching value/performance.

[uruguru]

Originally Posted by MathBandit

I do. And that by definition means the pitchers succeeded in part due to having an elite defense behind them and so weren't as good as pitchers who had better ability to miss bats.

But that's literally not it. It's not like these pitchers couldn't dominate hitters. It's because the balance of the game was different back then. Dominating hitters was no more a viable strategy then than getting them to put the ball in play if you had a good defense behind you.

Batters weren't swinging for the fences then and willing to take a strikeout and walk back to the dugout like they do now. They would choke up to avoid the strikeout so it was counterproductive to attempt to strike out every batter. Nolan Ryan was able to do this only because he was extremely wild (so batters would not dig in) and he could literally throw 100+ mph for 9 innings. He was a freak of nature, tbh.

Players naturally respond to the metrics that are used to measure their performance. When statistical analyses came out that implied a batter striking out was no worse than any other out, GMs and scouts shifted their thinking and batters over time responded by not caring as much if they struck out.

The way we measure the game and what we value in it changes the way players play it. That doesn't mean that historical players who played a different game were lesser players simply because we now value their performances differently.

[1991Twins]

I'd argue that the park's factors AND defensive ratings are equally important to having players perform at their peak in a historical sim. Since OOTP can be played in so many various ways, and is not meant to simply recreate ratings, if the above environment is different from history (ie: a trade, drafted to a different team) then statistics should vary (ie: Greg Maddux playing his prime years in Colorado).

[kq76]

Originally Posted by Syd Thrift

Maybe a better way of handling this is to not have stamina do a lot to ratings when pitchers get tired but massively increase injury chances instead. The Nolan Ryan example here is kind of telling; even though there may be studies indicating his K rate remained the same throughout games, he was also rather famously placed on strict pitch counts throughout his 30s to prevent injury.

I generally welcome injuries, but this is a scary proposition even to me.

Just thinking about what probably happens IRL, it must be a combination of the two: performance goes down and injury risk goes up. And I can only imagine that's how OOTP simulates it too.

What stamina ratings are we talking about for these pitchers (Eck and Smoltz) anyway? My (fictional) league routinely tells me pitchers with 40+ stamina can start, 30- can't, it's only 35 stamina that's iffy.

[uruguru]

Originally Posted by Syd Thrift

he [Ryan] was also rather famously placed on strict pitch counts throughout his 30s to prevent injury.

Ryan had a tear in an elbow ligament and was placed on a 110 (iirc) pitch count limit in 1987 to protect his elbow, when he was 40. That was the only time he was ever on a pitch count limit. It was not uncommon for him to throw 125+ pitches in starts as a Ranger well into his 40s.

He was a beast.

[Reed]

In OOTP each at bat is an independent event unlike real life. IRL when a pitcher is having one of his better games, it carries over to the following batters.
I guess one solution would be to randomly (like maybe once every 5 games) the pitcher has a better rating than they normally have and then randomly (maybe again once every 5 games) have a worse rating than they normally have. But I am ok with the way it is.

[Syd Thrift]

Originally Posted by uruguru

Ryan had a tear in an elbow ligament and was placed on a 110 (iirc) pitch count limit in 1987 to protect his elbow, when he was 40. That was the only time he was ever on a pitch count limit. It was not uncommon for him to throw 125+ pitches in starts as a Ranger well into his 40s.

He was a beast.

We don’t have average pitch totals but he pitched to a similar number of batters per start in 1984 and 1986 and only faced a couple more per game in 1985. I’ll walk back “throughout his 30s”, and yes, once he left Houston he pitched more (ISTR consternation over the pitch counts contributing to why he chose to leave), but whatever he was doing in 1987, he was probably doing the previous 3 years as well.

[uruguru]

Originally Posted by Syd Thrift

We don’t have average pitch totals but he pitched to a similar number of batters per start in 1984 and 1986 and only faced a couple more per game in 1985. I’ll walk back “throughout his 30s”, and yes, once he left Houston he pitched more (ISTR consternation over the pitch counts contributing to why he chose to leave), but whatever he was doing in 1987, he was probably doing the previous 3 years as well.

Well, as a Houston fan who just happened to move to Dallas in the same year that Ryan signed with the Rangers (!) I am reciting from memory so take everything with an appropriate grain of salt. Ryan was put on a pitch count limit in 1987 because of an elbow injury. He was pitching in pain a lot in 86 and was taken out of a lot of games early that year. He was too old to want the off-season surgery so pitch count limit it was. He had an amazing season in 87, leading the league in K's and ERA, but terrible run support landed him an 8-16 record. He ended up with no complete games because of the pitch limit. By 1988 his elbow was back to normal and then ownership tried to lowball him when his contract was up so he went to Dallas.

[polydamas]

There should be a hidden consistency rating like Work Ethic, Leadership, ect.


I propose the way to balance it:


Look at the streakiest, hot/cold players of all-time and determine what their peak, low and average ratings would be. Then make the 0 consistency replicate their variability.



An example of a streaky player: JD Martinez

in 2020: 237 PA .213/.291/.389 79 OPS+

in 2022: 173 PA .379/.439/.588 189 OPS+

From 2019-2022 (to establish baseline): .292/.363/.522 131OPS+

That's a huge variation. +58 when hot and -52 when cold.


For Pitchers, a slightly less good example (due to cheating ect.): Dallas Keuchel


in 2015: 911 BF 2.48 ERA 1.017 WHIP 2.91 FIP 157 ERA+
in 2016: 701 BF 4.55 ERA 1.286 WHIP 3.87 FIP 86 ERA+
from 2014-2017: 3.15 ERA 1.140 WHIP 3.38 FIP 125 ERA+

So, +32 when hot and -39 when cold.


Another Hitter Example: Justin Upton

in 2016: 626 PA .246/.310/.465 106 OPS+
in 2017: 635 PA .273/.361/.540 137 OPS+
from 2015-2018: .257/.338/.481 121 OPS+

+16 when hot, -15 when cold.


A Historical example: Reggie Sanders

in 2000: 377 PA .232/.302/.403 76 OPS+
in 1999: 550 PA .284/.376/.527 134 OPS+
from 1998-2001: .264/.344/.478 109 OPS+

+25 when hot, -33 when cold


How about Chris Davis?

in 2014: 525 PA .196/.300/.404 96 OPS+
in 2013: 584 PA .286/.370/.634 168 OPS+
from 2012-2015: .256/.342/.534 136 OPS+

+32/-40


Joey Gallo

in 2020: 226 PA .181/.351/.458 87 OPS+
in 2019: 297 PA .253/.389/.598 145 OPS+
from 2018-2021: .208/.338/.485 117 OPS+

+28/-30