Too many two out inning rallies by AI!! (Comeback AI Discussion)

[kehlis]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

Clutch doesn't exist in baseball so I don't think it should be in the game.

Ahh, a fellow moneyball believer.

I agree.

[swaldo]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

Swaldo,

I looked at the data you posted and agreed, the best way to conclude if there even is a "comeback code" is to take about 10-15 games cpu vs. cpu. If in 15 games, the trailing team in the 7-9th innings ends up winning 10/15 games then I would start to believe in it. Otherwise, if it's only like maybe 4-5 games then there's nothing to worry about b/c that's realistic, just as you stated.

The human instinct is to rest which sometimes causes a loss in even a smidget of focus when winning which can cause the CPU to mount a comeback. It's a lot harder to pitch from the stretch than the windup in my mind and in some higher difficulties (legend especially) if I'm not really close on the white/yellow marker when trying to aim the pitch it's pretty much a meatball coming to the plate.

I do love the pitching in this game. I thought it was going to be simplistic but is very deep and well done. In any case I think an in depth tutorial is needed for beginners or this thread may live a long time. I'm going to continue with the CPU vs CPU testing - if for no other reason I'm a stats buff and just have to do it.

[phillyfan23]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

another game in my season.....

6-2 win phillies vs nats

got out to a 5-0 lead after 2 innings....

mowed them down with Myers but then in the 7th they get a 2 r HR with one out. I made a mound visit, K looking and ground out to end the 7th.

I tack on another run in top of 8th and went smooth sailing with Myers in 8th and Durbin in the 9th.

I also took a look at marino's dynasty scores in the dynasty forum...

here are HIS scores on default all-star ( thanks danmarino hehe)

4/6/2009 @ San Diego W 3-2
4/8/2009 @ San Diego L 4-2
4/9/2009 @ San Diego W 8-2
4/10/2009 @ Arizona W 5-3
4/11/2009 @ Arizona W 13-6
4/12/2009 @ Arizona L 6-1
4/13/2009 San Francisco W 5-2
4/15/2009 San Francisco L 5-4
4/16/2009 San Francisco L 7-5
4/17/2009 Colorado W 3-2
4/18/2009 Colorado W 5-4
4/19/2009 ColoradoW 7-3
4/21/2009 @ Houston W 10-7
4/22/2009 @ Houston L 2-1
4/23/2009 @ Houston L 4-3
4/24/2009 @ Colorado W 5-1
4/25/2009 @ Colorado W 6-5

he has won a good amount of games with blowouts consistently

[bears5122]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

lol...you are apparently thinking in extremes.

Your analogy for reflective stats is heavily flawed (or another exxageration). Please, re-read what I typed:

"The reason stats don't bear it out is because it is dependent on other variables (skill, situation, mental state, etc). Meaning a player can be inherently more or less reliable to perform in clutch situations than another. The averaging out you speak of is looking at raw numbers at the back end and not taking the individual variables into consideration and can't quantify "clutch" on it's own merits."


I said that it can't really be quantified because it's DEPENDENT on other variables. There's no stat for heartrate, stress level, emotional state...lol. These are FACTORS in a players performance...same as physical health, fatigue, intelligence, etc.

Bottom line stats can't definitively reflect to what degree of influence they had on performance. It's a visual indicator. Some players more than others (due to internal physiological responses to stress) will perform worse than others in these "clutch" scenarios.

The only way people can really quantify such performance is by recording various situational stats. These are obviously very broad and don't tell the whole story, but do indicate who has more propensity to perform or fall off in these situations.

I'd also agree that no player becomes particularly better in these situations. Some do however "lock in" (basically focus) and have a greater chance to succeed.

Other players don't handle the pressures as well and fall off (basically, lose focus) in similar situations.

Also, positive/negative performance on the field goes beyong baseline stats. I'm sure you've heard the expression "won't be seen in the boxscore" before. You can help your team by moving a runner over, working a long count, handling a pitching staff (in the case of a catcher), etc.

Not everything is quantified via formula as you seem to infer in your argument. I'm not saying you aren't aware of that, but you seem to have (at least in this case) grasped onto an exxagerated belief that, again, goes beyond the data you reference.



lol...another example of extreme. C'mon, where did I say it anything close to resembling anything close to your Cy Young comment? I never made an argument that players become some form of "superman" when the moment arises...lol.

I simply stated that there is enough info available to believe certain players do perform better than others in "clutch" situations. I even acknowledged it varies based on situation.

Performance under stress is not an "all men are created equal" type attribute. Some handle it better than others. You could be the most skilled player and struggle under certain situations if you fail to manage stress. Yes, things balance out over time...that's usually due to situation and circumstance.

Look at A-Rod. He performed at or near his averages in the postseason. Then the collapse against Boston in 2004 happened where he didn't perform well. He hasn't since.

This is a confidence issue, but can obviously effect how he performs in the "clutch." Why then and not any other time?

Simply because, he's not being labeled a choke artist in April, so there is no pressure to perform. It ratchets up up bit when October hits and the media, fans, and teammates are all looking at him a little closer. This causes more stress, reduces confidence, and makes him less reliable in the clutch.

Does that mean he can't perform in the clutch? Of course not. It just means that he's not the guy you want up there (probabaly not the guy he wants up there either...lol).

That's my point, "clutch" isn't a "myth." Maybe perception has given it mythical status for some players, but that only cements it in an odd way.

It's like a "self-fulfilling prophesy." If a guy believes he's clutch because everyone says so, then he is. Why? Because he's more likely to not be worrying about coming through and instead will be more focused on the task.

A guy labeled "not-clutch" is more likely to fail because he's trying to shed a label. Happens all the time (and, of course, varies with personality).

If you were arguing that some of the concepts or perceptions surrounding "clutch" are false, I'd agree. I don't agree, however, that clutch is a myth.


I guess I'm saying I'm not really sure what your argument is exactly...lol. You seem to first argue "clutch" is a myth, but then you argue a point (and reference material) that "clutch" is misperceived and falsely attributed as a "hard" attribute. I can agree with you on the second point.

I don't believe there as many "clutch" players as people think. I do believe stats bear out (in most major sports) that there are extraordinary players that "rise to the occasion" under extraordinary circumstances. Not go far beyond their range of ability, but certainly don't wilt under the bright lights of such situations.

As I said a couple times, what you referenced isn't saying that there's no such thing as clutch, they're saying that no player is specifically a clutch player.

So which are you arguing? That clutch doesn't exist, or that clutch players don't exist? Or both?

Maybe we are talking about two different aspects or misunderstanding each other, because you seem to be pursuing two fronts...lol.

EDIT: and sorry for the drawn out responses...lol. I try and warn people when my long-windedness is about to strike, but I also think as I go and want to make sure I'm clear. So the warning got skipped...lol. Oops!

My argument is that clutch players don't exist. While many have bashed A-Rod for being a choke artist, people fail to mention that they are really basing it on about 40 at-bats. They don't mention how he tore it up during his first few years in the playoffs. 40 at-bats is a ridiculous sample size to determine someone is a choke artist.

I'm not arguing that there aren't psychological factors effecting players. I'm arguing that over time, a player will ultimately perform just as well in clutch situations as he does in non-clutch situations. That David Ortiz isn't clutch, he's just a great baseball player who did what great baseball players do. That if you put him up in those pressure situations 1000 times, his numbers would be similar to what his career averages are.

I'm not basing this analysis off what I see personally, but on decades of statistical analysis many experts have run. These reports really don't need much of an explanation. When someone can provide me with data that shows me certain players performing much better in pressure situations, I'll change my views.

[kehlis]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

My argument is that clutch players don't exist. While many have bashed A-Rod for being a choke artist, people fail to mention that they are really basing it on about 40 at-bats. They don't mention how he tore it up during his first few years in the playoffs. 40 at-bats is a ridiculous sample size to determine someone is a choke artist.

I'm not arguing that there aren't psychological factors effecting players. I'm arguing that over time, a player will ultimately perform just as well in clutch situations as he does in non-clutch situations. That David Ortiz isn't clutch, he's just a great baseball player who did what great baseball players do. That if you put him up in those pressure situations 1000 times, his numbers would be similar to what his career averages are.

I'm not basing this analysis off what I see personally, but on decades of statistical analysis many experts have run. These reports really don't need much of an explanation. When someone can provide me with data that shows me certain players performing much better in pressure situations, I'll change my views.

Bill James and Billy Beane agree!

[bears5122]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

Also, in reference to the video game, the clutch rating isn't really reflective of a player being clutch, it's an indicator of how a batter/pitcher performs in various "clutch" situations (which ARE statistically referenced) in-game.

You may disagree with the terminology in play (we can call it "situational" if you prefer), but the concept behind how it's reflected in-game is not unsupported.

Is there an aspect that concerns you as it relates to "clutch" (again bearing in mind that the rating is not necessarily reflective of the "mythical" properties you feel people attach to it)?

That's why I asked how we got on this topic. In what manner does your "myth" theory relate to the game?

But the rating is reflecting a mythical property. It is saying that certain players, regardless of their on-field ratings, perform better in "clutch" situations. Since I don't believe there is such a thing as a clutch player, I don't see why players should be rated with that attribute.

And even if you believe that players are clutch and perform better than normal in certain situations (which statistics show is not true), you should still not want this in the game. How are we to determine how clutch every player in baseball is? With regular ratings, you can determine what they are based on how closely they perform in the game to what they do in real life. But with clutch, you are just guessing on hundreds of players. What is the basis for Kerry Wood having a low clutch rating? Or who has figured that some unproven rookie deserves a high one?

[bears5122]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

Ahh, a fellow moneyball believer.

I agree.

Books that every baseball fan should read:

Moneyball
The Politics of Glory

And an honorable mention to Crazy 08.

[phillyfan23]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

won another game 6-1.....

man, i might NEED some comeback code if things keep going like this heheh just kidding....

playing the nats will be a good way to disprove the comeback AI for sure.

[kingdevin]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

won another game 6-1.....

man, i might NEED some comeback code if things keep going like this heheh just kidding....

playing the nats will be a good way to disprove the comeback AI for sure.

What sliders have u adjusted if any

[davewins]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

The thing is, it has been tested from a bunch of us, from the Red Sox and Yankees, to the Nationals and Royals. Same results.

More potent teams will come back against the Royals, and vise versa ... it's baseball.

I dont see how close games or a team trying to come back is a worrisome thing. It is baseball, relievers are not all that outstanding, and average baseball team's bullpen is suspect at best, thus as a gamer you need a much greater concentration, which the devs have captured remarkebly.

You hit the nail on the head man. People (including myself until now) don't even realize that starting pitcher's are the best in the game. You are a starter because you are nasty. You aren't a relief pitcher that comes in the 6th or 7th because you are nasty. Granted, of course there are plenty of great relief pitcher's but usually that's set up men and closers. When your starter is getting tired you have to take him out.

You CAN NOT make mistakes in this game just like in real life big league hitter's will hit these mistakes ESPECIALLY ON HITTER'S COUNTS. They are looking for fat pitches on hitter's counts unless they are completley fooled (nasty changeup that they haven't seen in any of their at bats previous). You need to listen to your catcher especially late in the game because this is when he's calling pitches that these batter's haven't seen before. Even if you have to throw a 2seamer up and away that's trying to catch the corner and it's your pitcher's worst pitch just go for it. Chances are the hitter isn't going to make great contact unless it's right down the middle. You can't keep your patterns. After the first 3 at bats of pitching the way you pitch (and you all have patterns it's inevitable) the cpu batter's pick up on it and even if you make a slight mistake you are going to get hit.

My advice is to listen to your catcher ESPECIALLY late in the game and FOCUS of course, last but not least MANAGE YOUR BULLPEN. If you do these things all of a sudden the comeback code will slowly start to vanish from your mind (and your game)

BTW, I think we might need to merge "Clutch discussion" into this already merged "too many 2 out inning rallies/a scientific approach" thread!! LOL

[Blzer]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

On a serious note, relievers most of the time are relievers because they are not good enough to be starters.

Sometimes I like to argue the other way... starters are starters because they're not good enough to be relievers.

[baa7]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

Sometimes I like to argue the other way... starters are starters because they're not good enough to be relievers.

The whole Joba Chamberlain as starter versus reliever thing shoots down your reverso logic I'm afraid.

[JT30]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

And we can shoot that down with Papelbon

[Bigtonyclark]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

Ahh, a fellow moneyball believer.

I agree.

Except that Moneyball has nothing to do with clutch. Chalk another one up for incorrect use of the word "Moneyball", a business strategy.

Bill James and Billy Beane agree!

Bill James himself has said he doesn't have a definitive answer for whether or not clutch exists:

http://sportsillustrated.cnn.com/2007/baseball/mlb/11/30/james.clutch/

[davewins]

Re: Too many two out inning rallies by AI!! (Comeback AI Discussion)

And we can shoot that down with Papelbon

I think closers are just as good if not better then starters (some anyway). I just think people are getting in trouble with their middle relief guys. Typically these pitcher's don't stack up with starter's because of their lack of consistency.

I think personally that starters are better overall because they have to pitch through at least 5 innings and face the same batters and continue to get them out. They have to have enough gas in the tank and enough weapons in the armory to continue to get these hitter's out 2 or 3 times a game. Relief pitchers typically have to just get a few guys out and some are specialists that only pitch to certain guys. Mostly never face the same batter twice in a game.