Tonight on 60 Minutes.... The RchW Interview

[Lukas Berger]

Originally Posted by joefromchicago

Yes.

Hmm.

The problem here is that I'm somewhat ignorant as to how much correlation we should be expecting to see between the statistical output of rl MLB and a fictional league using historical settings Does anyone know how much correlation is intended to be in OOTP's current fictional statistical model?

I feel like we need some more data here first of all on what's actually happening with this in OOTP14 and how much variance there is in the statistical outputs from sim to sim.

Then perhaps a discussion of how closely we should expect the statistical outputs of fictional leagues using historical settings to correlate with those of the historical MLB would be in order?

[jazzrack]

actually we don't disagree as much as it may seem, but it should be looked at as a pyramid not a curve, bell or tail. and with this viewpoint scale doesn't even matter, any angle is the same viewpoint, there is the handful of the most talented at the top and ever increasing numbers as talent level decreases. but for playability ootp cuts off the top and plays just with that.

it will have a noticeable impact if one looks hard/long/complicated enough. in a scientific endeavor I would agree with you but in a game I for one don't want to see a bunch of players rated 1 & 2 across the board, it is a bad visual.

[Le Grande Orange]

Originally Posted by ukhotstove

Ever get Time Team over your way RchW?

It's aired on TVO, which is the province of Ontario's public broadcasting network. Here's the show page on TVO's web site.

[Bluenoser]

Originally Posted by lukasberger

Hmm.

The problem here is that I'm somewhat ignorant as to how much correlation we should be expecting to see between the statistical output of rl MLB and a fictional league using historical settings Does anyone know how much correlation is intended to be in OOTP's current fictional statistical model?

I feel like we need some more data here first of all on what's actually happening with this in OOTP14 and how much variance there is in the statistical outputs from sim to sim.

Then perhaps a discussion of how closely we should expect the statistical outputs of fictional leagues using historical settings to correlate with those of the historical MLB would be in order?

This.

[The Wolf]

Originally Posted by joefromchicago

No, it's a bell curve - or at least bell-curve-like.

Your model posits that everyone with sufficient talent (in the blue area at the far right end) will become an athlete, while everyone to the left won't. We know, however, that that's not true. Plenty of people who have the talent to be professional athletes don't take that path, while some people, with marginal skills, nevertheless make it into the professional ranks. There isn't, therefore, a straight black line separating the athletes from the non-athletes, but rather a blurred area near the right end of the curve that gets more blue as you go further to the right.

Isolating the athletes, therefore, doesn't result in a small section of a larger bell curve, it results in a bell curve distribution itself. And that's what we should expect. The average professional baseball player is ... well, average. There aren't a lot of players at the low end, as they tend to get weeded out, and there aren't a lot of players at the high end, because that's the way talent works. Where OOTP's player creation model goes wrong is ignoring the long "tail" at the high end, where the truly great players reside.

No way.

[The Wolf]

"Talent in baseball is not normally distributed. It is a pyramid. For every player who is 10 percent above the average player, there are probably twenty players who are 10 pecent below average." - Bill James

[joefromchicago]

Originally Posted by The Wolf

No way.

Well, that settles it. I am convinced by the overwhelming power of your irresistible logic.

[Lukas Berger]

Originally Posted by The Wolf

"Talent in baseball is not normally distributed. It is a pyramid. For every player who is 10 percent above the average player, there are probably twenty players who are 10 pecent below average." - Bill James

I like the quote and think it's accurate, but doesn't that quote actually help to make Joe and Rch's case? Not sure why you're using it as an argument against what they're saying...

[The Wolf]

Originally Posted by lukasberger

I like the quote and think it's accurate, but doesn't that quote actually help to make Joe and Rich's case? Not sure why you're using it as an argument against what they're saying...

If you look at the data on the edge of a bell curve you can represent it it that way. But don't you understand that the pyramid is made up of smaller and smaller layers as talent increases? This is just another way of explaining that it is NOT bell curved. There is no standard distribution of talent once you get into professional baseball - in OOTP terms, there's a lot of 1 star talent, less 2 star talent, even less 3 star talent, still less four star talent, and hardly any five star talent at all - and that is not a bell curve!

[The Wolf]

Originally Posted by joefromchicago

Well, that settles it. I am convinced by the overwhelming power of your irresistible logic.

You should be.

[injury log]

Thanks Rch and Wolf for the feedback on Stephenson.

As for the discussion about ratings distribution in OOTP vs. skill distribution in real life... well, you need to present data if you're going to persuade Markus to change anything. And by data, I don't mean "my career HR leader has 120 fewer HRs than Barry Bonds". For one thing, you can't model ratings distribution based on one outlier (the most extreme power hitter in real life or OOTP). For another, even if there is a problem with outliers in OOTP, knowing that alone doesn't tell you how to fix the issue. It may be that when OOTP generates a guy with once-in-a-lifetime power, that power is not sufficiently, well, powerful. But batting and pitching ratings interact. If you play a guy with 80 Power in OOTP against a lot of pitchers, half of whom are really good and half of whom are really bad, he's going to do worse than if he faces a whole lot of bog-average pitchers. That's just a mathematical consequence of log5 modeling. So if your HR leader doesn't have enough HRs, it might be pitching ratings that are the culprit. And without a lot of data -- crucially, about all the guys in the middle of the ratings spectrum -- it's impossible to really do anything.

[Lukas Berger]

Originally Posted by The Wolf

If you look at the data on the edge of a bell curve you can represent it it that way. But don't you understand that the pyramid is made up of smaller and smaller layers as talent increases? This is just another way of explaining that it is NOT bell curved. There is no standard distribution of talent once you get into professional baseball - in OOTP terms, there's a lot of 1 star talent' less 2 star talent, even less 3 star talent, still less four star talent, and hardly any five star talent at all - and that is not a bell curve!

Ok, good point. That's true enough. I think you're right on the bell curve part, but that wasn't the main thrust of the original argument. I don't know why the talk about bell curves started as it doesn't really seem that applicable to the main point of the discussion.

Their original argument, if I understand it correctly, was that those 5 star players aren't as dominant in OOTP as in real life since there are too many players rated at three and four stars in the game and the stats are too evenly distributed. Thus that 5 star type talent needs to be rarer and also that those 5 star players should get a higher share of the stats generated in a given league. A pyramid, in other words. Which seems to be supported by your Bill James quote and seems to be what you're arguing as well.

As to whether it's actually an issue in OOTP or not is something I have no opinion on. I do think it's worth discussing and gathering data on though.

[Lukas Berger]

Originally Posted by injury log

Thanks Rch and Wolf for the feedback on Stephenson.

As for the discussion about ratings distribution in OOTP vs. skill distribution in real life... well, you need to present data if you're going to persuade Markus to change anything. And by data, I don't mean "my career HR leader has 120 fewer HRs than Barry Bonds". For one thing, you can't model ratings distribution based on one outlier (the most extreme power hitter in real life or OOTP). For another, even if there is a problem with outliers in OOTP, knowing that alone doesn't tell you how to fix the issue. It may be that when OOTP generates a guy with once-in-a-lifetime power, that power is not sufficiently, well, powerful. But batting and pitching ratings interact. If you play a guy with 80 Power in OOTP against a lot of pitchers, half of whom have ridiculously good Movement, and half of whom are really good and half of whom are really bad, he's going to do worse than if he faces a whole lot of bog-average pitchers. That's just a mathematical consequence of log5 modeling. So if your HR leader doesn't have enough HRs, it might be pitching ratings that are the culprit. And without a lot of data -- crucially, about all the guys in the middle of the ratings spectrum -- it's impossible to really do anything.

Exactly Hence my calls for more data in my posts above.

[The Wolf]

Originally Posted by lukasberger

I like the quote and think it's accurate, but doesn't that quote actually help to make Joe and Rich's case? Not sure why you're using it as an argument against what they're saying...

They're saying that MLB talent is normally distributed - a whole bell curve of it's own - and I know that it's not.

[Lukas Berger]

Originally Posted by The Wolf

They're saying that MLB talent is normally distributed - a whole bell curve of it's own - and I know that it's not.

I agree with you. But my understanding is that's what they were saying too. They're saying that in OOTP talent is too normally distributed, but that in RL MLB it isn't.

Oh well, no point arguing about it as they can simply clarify any wrong impressions themselves...

[injury log]

Originally Posted by joefromchicago

No, it's a bell curve - or at least bell-curve-like.

[...]

Isolating the athletes, therefore, doesn't result in a small section of a larger bell curve, it results in a bell curve distribution itself.

And this argument made no sense to me. If you have a certain distribution of data, then remove data points randomly, the remaining data points should still be distributed in the same way as the original data. So if you think athletic skill is distributed on a bell curve, so that MLB players are the extreme right tail of that curve, the fact that some random collection of MLB-calibre athletes instead take up golf or accounting or cinematography doesn't change the expected distribution of skills at all. You'd still expect to see the tail end of a bell curve.

[The Wolf]

Originally Posted by lukasberger

Their original argument, if I understand it correctly, was that those 5 star players aren't as dominant in OOTP as in real life since there are too many players rated at three and four stars in the game and the stats are too evenly distributed. Thus that 5 star type talent needs to be rarer and also that those 5 star players should get a higher share of the stats generated in a given league. A pyramid, in other words. Which seems to be supported by your Bill James quote and seems to be what you're arguing as well.

I don't argue that. But within the MLB, relative scarcity of players increases proportionately with talent level. That's my point, and Bill James' point too.

[Lukas Berger]

Originally Posted by The Wolf

I don't argue that. But within the MLB, relative scarcity of players increases proportionately with talent level. That's my point, and Bill James' point too.

[ukhotstove]

Originally Posted by Le Grande Orange

It's aired on TVO, which is the province of Ontario's public broadcasting network. Here's the show page on TVO's web site.

Just wondering with his interests if he watched it, they've stopped filming it but have done a couple of specials.

[RchW]

Originally Posted by ukhotstove

Ever get Time Team over your way RchW?

I wasn't aware of that show but have seen other shows on other channels. There was one on Pompeii several years ago that reconstructed the events in 79CE.