What is More Important in Historical Leagues: IRL Stats or Ratings?

[Brad K]

Originally Posted by uruguru

I've developed a (non-baseball) strategy game and I can tell you first-hand writing a competent AI is extremely difficult and that RNG elements are the easiest way to conceal AI shortcomings.

What does AI have to do with ratings that prevent historical outcomes?

[uruguru]

Originally Posted by Brad K

What does AI have to do with ratings that prevent historical outcomes?

You said that OOTP devs were "married to randomness" and I was trying to give a reason as to why that might be.

[thehef]

Originally Posted by David Watts

Exactly. That being said, OOTP has 1 year recalc, 3 year recalc and 5 year recalc. For the life of me, I don't know why Maris would be rated in a way to prevent him reaching or even exceeding 61 when using 1 year recalc. Nor, do I see a reason for Mickey Mantle to be rated to hit more home runs in 61 than Maris when using 1 year recalc. 3 year and 5 year recalc should do exactly what you describe by allowing Maris to come closer to his average hr total without maybe reaching the 61 outlier season.

Oh boy... We had a lengthy discussion on this maybe 18 months or so ago in the Historical Forums... I get the argument that if '61 Maris is rated to hit 61 dingers, that - with one-year recalc - he'd be just as like to hit 71 as he would be to hit 51. To that, I say, "so what?" That's one-year recalc, and as you noted, if that's not what you're after then that's what 3- and 5-yr recalc is for...

Just like I don't want 1980 George Brett modified to nearly always hit below .390, I don't Maris to consistently hit fewer than 61 HR's...

Originally Posted by uruguru

It depends. Are you trying to recreate 1961, or are you trying to recreate Roger Maris?

I typically would want to do both, by playing a historical sim over several years using 1-year recalc. If doing so, I should see Maris hit in the vicinity of 61 HR's in 1961, just as (un)likely to hit 41 or 51 as he is to hit 71 or 81... and produce similar-to-real-life numbers in his other seasons, too.

I don't see the problem with that approach. I figure if gamers scream about Maris hitting 71, then they need to be educated as to what 1-year recalc is all about, and what their options are... Maris, with 1-yr recalc, hitting 71 or 51 would not be a flaw in the game. Just like 1980 G Brett hitting .370 or .410 would not be indicate a flaw. Alas, not my call

[uruguru]

Originally Posted by thehef

I typically would want to do both, by playing a historical sim over several years using 1-year recalc. If doing so, I should see Maris hit in the vicinity of 61 HR's in 1961, just as (un)likely to hit 41 or 51 as he is to hit 71 or 81...

I think this is where our opinions differ. In the real world, there is obviously an increasing level of difficulty in hitting more and more home runs that does not exist in a computer simulation. This suggests that the distribution of home run totals in the simulation should be bottom-heavy. What I mean by that is that we can all easily imagine a season with Maris hitting only 51 homers (he did hit 61, after all), but no one has ever hit 71 without the benefit of performance-enhancing drugs. So, from that perspective, a sim where Maris hits 71 home runs in 1961 is *much* less realistic than one in which he hits 51.


Mathematically they may seem equivalent but, from the perspective of trying to simulate a real baseball season, they are not.

[thehef]

Originally Posted by uruguru

I think this is where our opinions differ. In the real world, there is obviously an increasing level of difficulty in hitting more and more home runs that does not exist in a computer simulation. This suggests that the distribution of home run totals in the simulation should be bottom-heavy. What I mean by that is that we can all easily imagine a season with Maris hitting only 51 homers (he did hit 61, after all), but no one has ever hit 71 without the benefit of performance-enhancing drugs. So, from that perspective, a sim where Maris hits 71 home runs in 1961 is *much* less realistic than one in which he hits 51.


Mathematically they may seem equivalent but, from the perspective of trying to simulate a real baseball season, they are not.

Fair enough. But I don't believe there is the same governor (of sorts) applied to George Brett's 1980 or Rod Carew's 1977 batting avg.

Personally, if I were to do 10 replay sims (1-yr recalc, etc.) for 1961, I'd rather see Maris avg around 61 HR's, as opposed to a low-50 number... even if a few of those seasons were 65, 67, 70... Repeatedly not seeing one of the most-memorable seasons in baseball history play out in a sim is less than ideal, IMO.

"The top single-season homerun hitters will generally not approach their real-life HR numbers even if your game is configured for the most-accurate replay," is not a feature to be advertised But it's true.

[David Watts]

Originally Posted by thehef

Fair enough. But I don't believe there is the same governor (of sorts) applied to George Brett's 1980 or Rod Carew's 1977 batting avg.

Personally, if I were to do 10 replay sims (1-yr recalc, etc.) for 1961, I'd rather see Maris avg around 61 HR's, as opposed to a low-50 number... even if a few of those seasons were 65, 67, 70... Repeatedly not seeing one of the most-memorable seasons in baseball history play out in a sim is less than ideal, IMO.

"The top single-season homerun hitters will generally not approach their real-life HR numbers even if your game is configured for the most-accurate replay," is not a feature to be advertised But it's true.

There's Carew and Brett, but again, what about 1961 Norm Cash? Dude hit .243 in 62 and never hit higher than .283 over the next 13 years of his career. In 61 he hit .361 with 41 bombs.

[Rain King]

I'd like to report a thread hijacking.

[uruguru]

Originally Posted by thehef

"The top single-season homerun hitters will generally not approach their real-life HR numbers even if your game is configured for the most-accurate replay," is not a feature to be advertised But it's true.

I honestly don't know if that's fair. In my quest to create better sims for my gameplay, I have played the early 1960s seasons more times than I would like to admit.

I have seen Harmon Killebrew hit 50+ homers more often than anyone else and he even had one 60-homer season. Mantle, Maris and Mays consistently drop 50+ bombs in random seasons. Maris did hit 62 once! I also saw Al Kaline hit 50+ in a 1962 season when I had injuries turned down.

But by far the craziest I ever saw was Jimmie Hall hitting FIFTY-SEVEN (57) homers in a 1963 sim (he hit 33 irl). It was so shocking that, when I used to trade with the AI, I was always irrationally trying to trade for him to play SS for me. Of course he never came close to that in any other sim.

So the potential for those guys to do it is always there even if we don't see it as much as we'd like.

[Brad K]

Originally Posted by Rain King

I'd like to report a thread hijacking.

How much can be said about the rookie level question that started this thread?

If user consensus is to discuss something else that is the defacto topic.

[Brad K]

Originally Posted by thehef

"The top single-season homerun hitters will generally not approach their real-life HR numbers even if your game is configured for the most-accurate replay," is not a feature to be advertised But it's true.

If that's the attitude then its not a feature. Its a defect.

[Brad K]

Originally Posted by uruguru

You said that OOTP devs were "married to randomness" and I was trying to give a reason as to why that might be.

I don't see how AI affects that. Does AI perhaps have something to do with their adultery?

[Brad K]

Norm Cash, 1961, single season ratings, OOTP 21, 22, 23. Anyone have him hit .400? Cut the homeruns though. That's important.

[Brad K]

In my current save the all time batting average leader is Lonnie Smith who hit .386 in 1980. That's 65 points over his real life best year and 98 points over his real life career average.

[Rain King]

I'm not sure what point you are making.

The "Resulting Stats" in those screenshots are not calibrated to the 1961 League Totals so they are telling us very little here about the actual expected results. What numbers does Cash show when running the Preseason Prediction?

[David Watts]

Originally Posted by Rain King

I'm not sure what point you are making.

The "Resulting Stats" in those screenshots are not calibrated to the 1961 League Totals so they are telling us very little here about the actual expected results. What numbers does Cash show when running the Preseason Prediction?

I just did a couple 1961 real lineups/real transaction sims. Can't find the Pstrickert settings so I just winged it. In the first one Maris hit 53 in the second 48. In the one he hit 53, he was predicted to hit 62 in the preseason predictions. Cash hit .335 in one and 331 in the other. In the first he had 44 home runs.

[uruguru]

Originally Posted by Brad K

I don't see how AI affects that. Does AI perhaps have something to do with their adultery?

Less randomness exposes AI deficiencies because the game becomes more deterministic and predictable to the human player. Writing a competent AI is extremely difficult and even moreso when you have so many optional features as in OOTP. This means that devs of games like this are encouraged to add more randomness to their games in order to hide those deficiencies.

[Brad K]

Originally Posted by Rain King

I'm not sure what point you are making.

The "Resulting Stats" in those screenshots are not calibrated to the 1961 League Totals so they are telling us very little here about the actual expected results. What numbers does Cash show when running the Preseason Prediction?

That's what they say but under close questioning they can't tell us what year represents a modern neutral MLB environment. It is quite amazing though how often those resulting stats are very close to historical stats regardless of the year selected.

I'm going to state that resulting stats are reasonably close to a player's real life stats for that year and are valid for this discussion. If you disagree then present your case.

[Brad K]

Originally Posted by uruguru

Less randomness exposes AI deficiencies because the game becomes more deterministic and predictable to the human player. Writing a competent AI is extremely difficult and even moreso when you have so many optional features as in OOTP. This means that devs of games like this are encouraged to add more randomness to their games in order to hide those deficiencies.

What AI deficiency is being hidden by this miscarriage of randomness? See 82 and 83.

[fredbeene]

Originally Posted by uruguru

I suspect it's the concept that Maris abnormally peaked at 61 HRs, it was not a typical season for him. This is a common idea for pretty much any season in which a player sets a single-season record.

So if you wanted to account for it being atypical, you might set Maris' ratings so that repeated replays of 1961 might see Maris average 50 homers, not 61, but he would still occasionally peak around 61 in some of those sims.

The counterpoint is that if you set his ratings to average 61 homers, then he would literally hit over 61 homers in half of the 1961 simulations you ran. That would break verisimilitude for a lot of fans and make the simulation's accuracy seem suspect.

Doesn't POTENTIAL handle all that? He has the potential to hit 61 HR
Unless you import him AFTER he turns 30 then POTENTIAL means nothing.

[fredbeene]

Originally Posted by Brad K

That's what they say but under close questioning they can't tell us what year represents a modern neutral MLB environment. It is quite amazing though how often those resulting stats are very close to historical stats regardless of the year selected.

I'm going to state that resulting stats are reasonably close to a player's real life stats for that year and are valid for this discussion. If you disagree then present your case.

When you adjust the game to achieve xxx output...it will give you that output.
In strat if you 'happen' to get super lucky and roll 10 HRs in a row it gets logged as such.
In otp if you start to hit too many homers a modifier is applied to ensure you get what you want as you set up the game.

Not sure what happens when you play hx and don't adjust managers since every time you play the settings are the same and there are no randomizers