So, OK, an official farewell ...

[RchW]

I seriously doubt if that setting would have any effect on a player hitting 0.100 above or below the norm, or any other output. Even if a player had a talent change his actual ratings might take some time to follow and would still be affected only by the batter pitcher matchup from AB to AB.

[TribeFanInNC]

Originally Posted by knockahoma

I may have been guilty of unclear writing. I meant the defensive variation for the same season in results. For instance, Bobby Grich. If he made 17 errors in real life, then he might hit somewhere between in 13-21 replay after replay. With dice, it's easy to see why. A field goal kicker will hit a 25 yarder 2-10, 12. He misses with an 11. So, I wonder if OOTP is more complex in determining outcomes?

I don't think it is more complex necessarily. It's more like OOTP can be more accurate because it isn't constrained by the discrete percentages that dice rolls give. With the miss-on-11 example, the dice are constrained to missing 2/36 times whereas OOTP can miss 2004/36000 times if it needs too. It can be more exact when making it's random number determinations.

I think this is the drawback to Strat. I think even the computer version uses dice rolls.

[robc]

I'm not here to rant at you or criticize your post. It's cool if the game isn't worth it to you, but I would like to offer this bit of info for you. If you already know the below info, then my apologies.

Since the game engine essentially uses probabilities to generate results, some of the variance from real life you see can be explained by the following.

If in real life a player has a .333 batting average, then a simulation based on his stats should produce a batting average between .313 and .353 68% of the time. It should produce a batting average between .293 and .373 95% of the time. As you can see there is still a 5% chance that the batter will be even out of this big range.

So, if you are seeing a ton of players who have large variances than it should be unexpected. But it is within a statistical norm to have 5% of the players to be pretty far of their real life batting averages.

Also remember that if you are playing with the development engine turned on, the players' ratings will start to vary from their initial numbers based on real life.

[knockahoma]

I don't think it is more complex necessarily. It's more like OOTP can be more accurate because it isn't constrained by the discrete percentages that dice rolls give. With the miss-on-11 example, the dice are constrained to missing 2/36 times whereas OOTP can miss 2004/36000 times if it needs too. It can be more exact when making it's random number determinations.

I think this is the drawback to Strat. I think even the computer version uses dice rolls.

That's a good observation. However, the point goes against the grain of my experience with these games. I would call STRAT and ACTION more accurate on replay mode. And I don't mean to criticize OOTP. I'm very happy with how far OOTP has come, not to mention "it strong-like-bull" in customization. Very enjoyable game.

If in real life a player has a .333 batting average, then a simulation based on his stats should produce a batting average between .313 and .353 68% of the time. It should produce a batting average between .293 and .373 95% of the time. As you can see there is still a 5% chance that the batter will be even out of this big range.

So, if you are seeing a ton of players who have large variances than it should be unexpected. But it is within a statistical norm to have 5% of the players to be pretty far of their real life batting averages.

Also remember that if you are playing with the development engine turned on, the players' ratings will start to vary from their initial numbers based on real life.

Yeah, that's it. However, I know that many years ago, STRAT started having split player cards. Both pitcher and batter had lefty/righty readings based on real-life stats. OOTP generates generic splits. I've seen OOTP's genero-split make a big difference time and again.

For example, in the old strat cards, you rarely used Oscar Gamble against lefties. You'd be considering your lineup, glance at Oscar's "against lefties" card and say, " Holy roller, batman, all I can see are a few un-likely homers and a ton of strikeouts. He looked like Ted Williams against righties. If you had to guess by his card (which did not break down BA against lefty and righty pitching), in 500 ABs against lefties, Oscar would hit .200, 10 HRs.

Not so in OOTP's genero-split where Oscar often hits .270 with good power against lefties. So, you gotta figure this genero-split plays against historic replay consistency, too.

[robc]

Originally Posted by knockahoma

So, you gotta figure this genero-split plays against historic replay consistency, too.

That is a good point!

[Killing Time]

Longest. Farewell. Thread. Ever.

[Syd Thrift]

Is the standard deviation for a ~550 at bat season 20 points of BA? I thought it was closer to 35-40 points. If so, I take back some of what I said. Still, up to the 3rd standard deviation is a range that covers 99.8% of a given population (given normal distribution, of course), if you have a 30 team league in which, say, 4 guys on each club are regulars, that's *still* room for there to be approximately 2 guys a year who are more than 3 standard deviations from the norm - in this case, hitting 60 points above or 60 points below their "real" ability. *And* about 2 guys a year will be more than 3 standard deviations off their HR totals (although since homers have a baseline at zero that BA really does not for all intents and purposes, distribution's not exactly normal). If you *don't* have a couple .250 hitters hitting .310 or a .330 hitter or two hitting .270 every year on average (meaning you could have a year with 4 or 5 guys outperforming and another with none), it's a sign that the game might be rigging results.

Which gets back to my point about the way humans perceive things. We see these outliers first because we are wired to do so. To compound the fact, if we've already sort of come to the conclusion that the game "isn't accurate" in this case, confirmation bias is extremely powerful: it's human nature to see the big misses but not the big hits. Even thinking about reading a post in which you comment that in your 1993 replay Ken Griffey Jr. hit .315 with 44 homeruns sounds kind of silly, doesn't it? It's not newsworthy when a player performs the way you'd expect him to and thus you ignore the 2/3rds of the league that does that.

If you drop it down to 200 at-bats, the standard deviation gets higher and the population does as well, so if you include platoon partners (and I think many players would if they were wont to complain about the game's "lack of accuracy"), you ought to end up with a dozen players or more every year missing their real life totals. Again, if you *don't* see that, you ought to be concerned that the game that you're playing isn't producing statistically genuine results. Dial it down to 50 at-bats and... well, Dave Cameron of fangraphs.com is fond of saying that you can't predict anything based on 60 ABs. There are just too many Shane Spencers and Scott Spiezios at that level.

FWIW I played a lot of Strat before OOTP came out as well. And before Strat, Statis Pro, and a little bit of APBA and, just before I switched over to OOTP, Diamond Mind. I *do* remember seeing weird outlier seasons every now and then. They were rather rare, as you'd expect them to be if you have a card and dice game and are replaying a single team's season and really, therefore, only seeing their stats (as was the norm back before computer games - could you imagine playing out an entire season of 1000+ games, each game taking half an hour to play, and then spending half an hour every game day compiling stats? Speaking as someone who tried to do that several times and failing miserably before SOM for Windows came out, that's a lot of time just to produce a single season's worth of stats). The one I remember in particular was the aforementioned 1984 Gorman Thomas, he of the .155 BA (not 1982 Gorman of course; that year he went to the World Series). Gorman hit in the .260s for me that year and I just couldn't bring myself to bench him, even as he went way over his PA limits. IIRC his power numbers were horrible but hey, he hit a lot more singles than IRL. I also remember Phil Niekro having an annoyingly good 1987 season for the Indians in a replay I did of that year.

[RchW]

Originally Posted by Killing Time

Longest. Farewell. Thread. Ever.

Well Elvis left the building after the 19th post. Ahh, thank you very much.

[Killing Time]

Originally Posted by RchW

Well Elvis left the building after the 19th post.

So Elvis left and you're just all hanging out now? Funny.

[Elvis]

Originally Posted by Killing Time

So Elvis left and you're just all hanging out now? Funny.

Now everyone can leave.

[silvam14]

I'm not 100% certain, but I don't think anyone who plays fictional (like I do) has ever started a thread like this one.

[Left-handed Badger]

I do have to admit I find the pitching side to be a total crapshoot in historicals. (but, I only tried it a couple of times in historicals, and my theory was that it probably had something to do with a significantly lower frequency of Ks back then, probably fouled up the conclusions).

But, I havent had any problems doing any thing with fictional stuff.

[robc]

Originally Posted by Killing Time

Longest. Farewell. Thread. Ever.

Farewell thread. OOTP accuracy debate. Life of its own.

How did I do?

[actionjackson]

Originally Posted by robc

Farewell thread. OOTP accuracy debate. Life of its own.

How did I do?

Bingo.

[RchW]

Interesting article that I've not read/understood completely yet.

The decline of the .400 hitter: an explanation and a test | Journal of Sport Behavior | Find Articles at BNET

It's not an easy read but it may add something to this discussion.

[Killing Time]

Originally Posted by robc

Farewell thread. OOTP accuracy debate. Life of its own.

How did I do?

Too well.

[dcpcarmona]

So long, farewell, auf weiderseshen good-bye
[Flan.Joe]
I hate [for you] to go and leave this pretty site...

The sun...has gone...to bed and so must [you]

So long...farewell...auf weiderseshen good-bye...
Good-bye...
Good-bye...
Good-bye...
[OOTP members]
Good-bye...

[Malleus Dei]

This is like a funeral. The reason you all gathered here is gone, so you're standing around making conversation. Go home!

[knockahoma]

A funeral. Been to my fair share of those. Apt description. Long after this funeral's conclusion, some of us have been scuffing our shoes against the perfectly cut grass, spitting, and recollecting old times, figuring most everybody else had already gone home. To our surprise, a few were listening to the chat from the bushes, yet to go home themselves.

Hey! I think I see someone in them darn bushes as I speak! C'mon out you!

[knockahoma]

FWIW I played a lot of Strat before OOTP came out as well. And before Strat, Statis Pro, and a little bit of APBA and, just before I switched over to OOTP, Diamond Mind. I *do* remember seeing weird outlier seasons every now and then.

You bring up a point I've been meaning to mention. With Strat, I never finished an entire season with all teams-- are you kidding? However, I was prolific at one-team seasons. In fact, I played all the teams that came with the game, and played my favorite teams innumerable seasons. Basically, I'd start a new season with a trade and get 3 or 4 new players to make it all fresh. So, in a way, playing one team over and over is an advantage. You get to see the variance in each card.

I think of cards like Vic Harris, the infielder for the Rangers, Cards and Cubs. Decent speed, bad fielder (in the minors, he was actually a .300 type hitter). But his 1976 Strat card was a .225 hitter with a couple of homers in 250-ish ABs.

My God! Rolling dice and looking for a hit out of Vic's card was like trolling for a date at a convent. I must have played 50 seasons with Vic if I played 1. I don't ever remember the fella coming close to .300. I'd root for Vic to hit .250 and if he did, wow!

But seems to me that OOTP is no slouch at historicals. I'm especially pleased with the improved defensive ratings. Used to spend an afternoon editing a new league. Now, I just do some tweaking.

And I just ran 1974, my favorite season on OOTP. Replay mode. No trades. Injuries on low.

REAL LIFE

NL EAST

Pitt 88 W
STL 86
PHI 80
MON 79
NY 71
CHI 66


NL WEST

LA 102
CIN 98
ATL 88
HOU 81
SF 72
SD 60


AL EAST

BAL 91
NY 89
BOS 84
CLE 77
MIL 76
DET 72


AL WEST

OAK 90
TEX 84
MIN 82
CHI 80
KC 77
CAL 68


And Oakland beats the Dodgers in the WS.



REPLAY

NL EAST

Pitt 90
MON 87
STL 87
NY 83
PHI 76
CHI 58


NL WEST

LA 100
CIN 91
HOU 85
ATL 79
SF 74
SD 62


AL EAST

BOS 95
BAL 93
NY 80
MIL 74
CLE 71
DET 65

AL WEST

Oak 103
KC 88
CAL 84
TEX 75
CHI 73
MIN 72


LA defeats OAK in the WS 4-2.

Pretty Cool.