Scripted Innings!

[thaSLAB]

Re: Scripted Innings!

This topic though.

Spoiler

I really just commented to get my stats evaluated.

:-D

Sent from my htcONE (M8) | Tapatalk

[Will I Am]

Re: Scripted Innings!

So basically on higher levels, the CPU gets a boost in certain areas. How is this realistic? Why wouldn't they differentiate the difficulty levels by simply making the CPU pitch more to corners or making the pitching meter harder to hit? For example, if I can have a team ERA on Allstar of 2.45 because I can pitch to corners all day by hitting the pitching meter, then wouldn't it be smart to simply make the meter harder to hit on HoF, thus making pitches more hittable? Does the CPU need a hitting boost to fabricate a harder difficulty? And this doesn't take into account anything regarding azst's late-game statistics and why the trailing team seems to score a much higher percentage of their runs in the later innings. Again, something seems amiss.

Make pitching harder on higher difficulty levels, what a novel idea.
No boosts just better pitching.

http://www.spiritstranslation.com/pitch_difficulty2012.html

[jvalverde88]

Re: Scripted Innings!

Make pitching harder on higher difficulty levels, what a novel idea.
No boosts just better pitching.

http://www.spiritstranslation.com/pi...culty2012.html

We get you love PYS, is there really a need to bring it up in a MLB The Show forum?

[kehlis]

Re: Scripted Innings!

Not when it's used to troll and I'm getting pretty tired of it.

Want to bring it up for an actual constructive reason? Feel free.

But enough as using it to troll another game. No one takes videos from the Show into the PYS forums to troll that game.

[kehlis]

Re: Scripted Innings!

Not trolling at all, I play both games a lot. Simply showing how they do it is all.

Nah, this is trolling. Nothing productive at all about this:

Make pitching harder on higher difficulty levels, what a novel idea.
No boosts just better pitching.

If you were being constructive you could have added to why you were posting this instead of saying "What a novel concept."

[Will I Am]

Re: Scripted Innings!

Nah, this is trolling. Nothing productive at all about this:




If you were being constructive you could have added to why you were posting this instead of saying "What a novel concept."

Well if you look at the graphs from the tests that were done it is explained very well.

[kehlis]

Re: Scripted Innings!

Well if you look at the graphs from the tests that were done it is explained very well.

None of that has anything to do with The Show.


And I did look at the graphs which are all very thorough nor are they surprising for anyone who watches a lot baseball.

[nomo17k]

Re: Scripted Innings!

You are correct in how I defined Trailing (behind by 1 or more runs at the end of the 6th inning).

It may very well be true that splitting the data in that manner creates a bias towards games that one would expect to see runs in later innings. However, just because a team is trailing, doesn't mean it didn't score runs nor does it mean it under-performed (not sure if we can easily determine what their normal performance level is). Though at the same time only the trailing team could have scored 0 runs, so it is entirely possibly that it was under-performing in the 1st half of the game. So, here is a comparison for both segments with exactly 2 and exactly 3 runs scored in the first 6 innings.

Run distribution when team scores exactly 2 runs in innings 1-6 and is Leading/Trailing after the end of the 6th:
Inn. | Leading | Trailing
1st | 12% | 9%
2nd | 12% | 6%
3rd | 8% | 10%
4th | 12% | 9%
5th | 11% | 11%
6th | 12% | 13%
7th | 12% | 19%
8th | 14% | 9%
9th | 8% | 14%
-------------------------------
7-9 | 33% | 42%

(Thus this is all instances where a team was leading after 6 innings and that team had only scored 2 runs in the first 6 innings, so 2-0, and 2-1 games. And the same for trailing - the team trailing after 6 only scored 2 runs, so any 2-3, 2-4, 2-X games.)

Run distribution when team scores exactly 3 runs in innings 1-6 and is Leading/Trailing after the end of the 6th:
Inn. | Leading | Trailing
1st | 13% | 12%
2nd | 10% | 12%
3rd | 13% | 14%
4th | 11% | 10%
5th | 11% | 6%
6th | 14% | 7%
7th | 9% | 10%
8th | 11% | 14%
9th | 8% | 16%
----------------------------------
7-9 | 28% | 39%

So if this evens out the under-performing bias, shouldn't the data align a little better for innings 7-9? Or is the best explanation that when it's HUM vs HUM there are too many variables at play that we can't expect it to align with real MLB game data.

Again, I don't believe there is a comeback code (possibly unintended consequences of a complex code?; non-visible bug like the catcher freeze or bat boy issue?; leftover '09 code but Woodweaver says that was removed; just randomness of the data?) but was curious enough to run the data. And more curious when user data started to align with the theory. And found it interesting that the data aligns with the '09 game explanation of the 2P Comeback Handicap that saintjimbo1885 posted a few pages back.

I don't exactly know what the sample size is for those data especially when you slice them, but I think the way the tabulated numbers fluctuate a few % and the change you are talking about is within that variation, it is very likely you are working in a regime where you don't have big enough sample to make statistically significant conclusions.

Just to prove my earlier point about how your method of calculating the run-scoring by inning, I ran some simulation (500,000 games) assuming totally even run-scoring ability in every inning for both home and visitor (i.e., 4 runs per 9 innings), and tabulated the numbers similarly as you did.

If I make a sample of game in which the visitor was behind after the 6th inning was completed (I didn't use the home team because if the home team wins after comeback, 9th inning may not be played which skews data even further; not sure how you dealt with that), then this is the results:

Inning Fraction of Total Runs 
1      0.0819028720168
2      0.0818385051823
3      0.0821242311303
4      0.0817050617451
5      0.0814774229403
6      0.0830175171985
7      0.15930948516
8      0.159700395935
9      0.159274946858
10     0.0127697519836
11     0.00730485074744
12     0.00416186531946
...

Roughly, the team losing at the completion of 6th inning scores 8.1 - 8.2% of total run per inning through the first 6 innings, but that fraction goes up to about 16% in 7 - 9 innings. Again, this happens without varying the ability of team to score in late innings, and it is purely due to bias in sampling.

I would actually say the difference in % looks pretty similar to what you tabulated for all the online games for users (only that your numbers fluctuate more due to small sample size). So it does not really provide strong evidence for late-inning comeback and boosts in scoring ability when the team is behind. I'd say that the best can be said is that it is still inconclusive.

Please don't take all this wrong and personally though, I'm not here to diss your work but just providing a perspective.

[HypoLuxa13]

Re: Scripted Innings!

My cat's breath smells like cat food.

[azstdogg]

Re: Scripted Innings!

Please don't take all this wrong and personally though, I'm not here to diss your work but just providing a perspective.

haha, I didn't at all. Hopefully I didn't come across that way either. I think it's cool that you can do that and I'm glad someone can pull a large amount of data. I would like to see something similar for online DD/exhibition games. Someone who ran http://baseballsimulator.com/diamond...yrecapdd20.php use to pull all kinds of interesting stats for DD games (thought that might have been you).

If I recall, the sample size for mine was about 100-120 half innings so slight less than your 500K. Regarding the 9th inning not being played for the home team should it be leading, I decided to ignore that and assume that it might average out for both segments with comebacks.

[TheWarmWind]

Re: Scripted Innings!

I'm going to be honest, I only gave the stats that have been presented here a brief overview, but I just thought of something. A lot of the stats deal with runs scored by a trailing team. Wouldn't that favour later innings because there is much more likely to BE a trailing team in later innings?

In fact, wouldn't that specifically favour the 7th, because a) there is much more likely to be a lead at that point. b) starting pitching is tiring/left the game, but not yet into the best RPs c) set up then closer for the 8th and 9th, representing a probable increase in pitching quality.

Also there is the factor that the human player has to adjust to a change in pitching styles. Not really a factor in real baseball since you only control yourself. Probably the only real life stat that is comparable is % of pitches dropped by catchers before and after a pitching change.

[nomo17k]

Re: Scripted Innings!

haha, I didn't at all. Hopefully I didn't come across that way either. I think it's cool that you can do that and I'm glad someone can pull a large amount of data. I would like to see something similar for online DD/exhibition games. Someone who ran http://baseballsimulator.com/diamond...yrecapdd20.php use to pull all kinds of interesting stats for DD games (thought that might have been you).

...

That DD fan page was run by mtpoktz/baseballsim. It was very well done, indeed.

[kingdevin]

Re: Scripted Innings!

[QUOTE=ryanmc564;2047951859]someone call the people who made the "making a murderer" doc. and get them to make a doc about this "making a comebacker" (the lies and deceit of a video game company and the customers who go thru great lengths to expose the truth).


Dude... your post is HILARIOUS!!!!!!!!!! Great post!!!!! LOLOLOL

...with special appearances by an OP all start cast featuring Blzer, bcruise, nomo, azstdogg, nemesis, knight165, P.A.D. and Woodweaver!!!!!

Regardless of position, its been a great forum to read all of your points (even those not mentioned). These kind of intelligent debates are what I enjoy about these forums.

[azstdogg]

Re: Scripted Innings!

I'm going to be honest, I only gave the stats that have been presented here a brief overview, but I just thought of something. A lot of the stats deal with runs scored by a trailing team. Wouldn't that favour later innings because there is much more likely to BE a trailing team in later innings?

In fact, wouldn't that specifically favour the 7th, because a) there is much more likely to be a lead at that point. b) starting pitching is tiring/left the game, but not yet into the best RPs c) set up then closer for the 8th and 9th, representing a probable increase in pitching quality.

Also there is the factor that the human player has to adjust to a change in pitching styles. Not really a factor in real baseball since you only control yourself. Probably the only real life stat that is comparable is % of pitches dropped by catchers before and after a pitching change.

I was looking at run distribution comparison between the team trailing after 6 innings vs the team leading after 6 inning (choosing that cut off since others were mentioning the 7th or late inning comebacks) with the assumption that run distribution should be fairly similar. Nomo has provide additional info as to why this cutoff may have bias.

I also chose recent DD games since most teams are stacked and less likelihood of a user bringing in a subpar RP. For example, I have a rotation of Diamond pitchers and all my relieves are gold/diamond. In addition, I rarely let me SP bat twice so it's unlikely that he makes it past the 5th before I bring in Chapman, Eckersley, Kimbrel, etc.

Adjusting for the new pitcher's style may be a factor too, though I was targeting those with a lot of games played so they may be more used to those nuances.

[HypoLuxa13]

Re: Scripted Innings!

A topic for conversation:

How much of this alleged 7th inning phenomenon/boost is accounted for by:
1) leaving starting pitchers in too long, where their accuracy, break and velocity are beginning to suffer.

2) leaving starting pitchers in too long, where the user's pitch sequencing has become more predictable as the opponent's lineup turns over and the hitters are getting 3rd and 4th looks at the pitcher in the 7th inning.

3) most MLB teams do not have solid middle relievers that are consistent. So the 7th inning is left up to a mix-and-match group upper 60's to low 70's rated relievers trying to bridge the gap to the higher rated setup and closing pitchers in the 8th and 9th.

4) I mentioned user pitch sequencing, hasn't it been confirmed that the AI in The Show detects patterns in user pitch selection? I remember the old days of console baseball games when you could use the same pitch sequences batter after batter and the CPU would never register a pattern. For example, if I throw a first pitch fastball on the outer 1/3rd of the plate over and over and over again, batter after batter, how long does it take the CPU to start jumping on that first pitch fastball? 2nd time through the batting order? 3rd time?