 |
View Old Forum Thread
Old Forum Index » Baseball Beta Testing » Beta News » Adding Logic for baserunners and defense
tysonlowery
The problem is that Abe sometimes does unrealistic things with runners on 3rd, such as letting the winning run score. We would need a whole set of logic on top of what we currently have to govern these types of situations. Here's what I was thinking of doing.
The Defensive team would be able to setup scenarios under which they play the infield in or the corners in. These scenarios would include where the runners are on base, how many outs there are, what inning or later it is, and whether the runners are winning or tying runs for the other team. These would all include having a runner at 3rd base.
If a team is playing the infield in, there would be a higher chance of a single, and a slightly higher chance of doubles, for the offense. But any ground balls would force the runners to stay put. Or in the case of the bases loaded, would result in a play at the plate.
If a team is playing the corners in, only balls hit to 1B or 3B would force the runenrs to stay put. The offense would have a higher chance of singles and doubles, but not at the same rate if the whole infield is playing in. Any ball hit to the 2B or SS would allow the runner on third to score.
Thoughts on this? Do I have the general logic right?
Adding Logic for baserunners and defense
August 09, 2006 at 03:08PM View BBCode
I'm trying to figure out how to set this up and would appreciate your feedback.The problem is that Abe sometimes does unrealistic things with runners on 3rd, such as letting the winning run score. We would need a whole set of logic on top of what we currently have to govern these types of situations. Here's what I was thinking of doing.
The Defensive team would be able to setup scenarios under which they play the infield in or the corners in. These scenarios would include where the runners are on base, how many outs there are, what inning or later it is, and whether the runners are winning or tying runs for the other team. These would all include having a runner at 3rd base.
If a team is playing the infield in, there would be a higher chance of a single, and a slightly higher chance of doubles, for the offense. But any ground balls would force the runners to stay put. Or in the case of the bases loaded, would result in a play at the plate.
If a team is playing the corners in, only balls hit to 1B or 3B would force the runenrs to stay put. The offense would have a higher chance of singles and doubles, but not at the same rate if the whole infield is playing in. Any ball hit to the 2B or SS would allow the runner on third to score.
Thoughts on this? Do I have the general logic right?
barterer2002
August 09, 2006 at 10:55PM View BBCode
I think in general you've got it. Do you, however, need to slightly bump the triples possibility as well though?
ScooterPie
I'm not entirely sure about this. Mightn't a ground ball hit sharply enough give the fielder a chance to hold the runner at 3rd? I just don't know. And that may be more detailed than you're looking to debate right now.
To respond to barterer, when I think triple, I think fly ball or line drive to the outfield, not down the line. I'm probably wrong, though. Yeah, I'm wrong. There should be slightly more triples as well.
Oh! one more thing: Tyson, you mention that the logic will consider whether runners are the tying or winning run(s). Don't teams also sometimes bring in their infielders when they're down, say, 2-0, to keep the game in reach? (If you're just choosing not to get this detailed, I think that's a fine decision. Just wanted it to be considered -- assuming I'm right that it's common enough that we should consider it.)
Oh yeah, another one more thing: The decision to bring in the infield is, I think, often based in part on factors like the batter's and runners' speed, and the batter's propensity to hit ground balls. I'm not sure how to phrase this next question, but here goes: Is that information that a manager needs in order to make the right decision? I guess my point is, if it's something that will affect the outcome of the play, then we should be able to tell ABE to consider it, right? And what weight to assign it?
(When I think about this problem, it gets so complex in my head, with all the variables that will be unique to each specific at-bat, that I think we should cede a lot of control to ABE, who has perfect information, rather than the human players trying to micro-manage every input into the decision. But then, I have yet to write a baseball simulation good enough to be played and beloved by thousands, so what the heck do I know?)
scooter
August 10, 2006 at 05:43AM View BBCode
Originally posted by tysonlowery
If a team is playing the corners in, [...] Any ball hit to the 2B or SS would allow the runner on third to score.
I'm not entirely sure about this. Mightn't a ground ball hit sharply enough give the fielder a chance to hold the runner at 3rd? I just don't know. And that may be more detailed than you're looking to debate right now.
To respond to barterer, when I think triple, I think fly ball or line drive to the outfield, not down the line. I'm probably wrong, though. Yeah, I'm wrong. There should be slightly more triples as well.
Oh! one more thing: Tyson, you mention that the logic will consider whether runners are the tying or winning run(s). Don't teams also sometimes bring in their infielders when they're down, say, 2-0, to keep the game in reach? (If you're just choosing not to get this detailed, I think that's a fine decision. Just wanted it to be considered -- assuming I'm right that it's common enough that we should consider it.)
Oh yeah, another one more thing: The decision to bring in the infield is, I think, often based in part on factors like the batter's and runners' speed, and the batter's propensity to hit ground balls. I'm not sure how to phrase this next question, but here goes: Is that information that a manager needs in order to make the right decision? I guess my point is, if it's something that will affect the outcome of the play, then we should be able to tell ABE to consider it, right? And what weight to assign it?
(When I think about this problem, it gets so complex in my head, with all the variables that will be unique to each specific at-bat, that I think we should cede a lot of control to ABE, who has perfect information, rather than the human players trying to micro-manage every input into the decision. But then, I have yet to write a baseball simulation good enough to be played and beloved by thousands, so what the heck do I know?)
scooter
ScooterPie
Okay. Now I'm done. Seriously.
scooter
August 10, 2006 at 05:50AM View BBCode
Sorry, I just saw one more thing.Originally posted by tysonloweryAnd said play at the plate would have a higher chance of being an out than if the infield were not in, yes?
If a team is playing the infield in, [...] any ground balls would force the runners to stay put. Or in the case of the bases loaded, would result in a play at the plate.
Okay. Now I'm done. Seriously.
scooter
tysonlowery
Ok, here was my thought on this. I wanted to keep it somewhat simple and keep the Offense out of the equation. Meaning, we would somewhat simplify things and not have plays at the plate (unless the bases were loaded). Otherwise, the Offense would have to set a similiar set of preferences about when to run home and it gets 10 times more complex. So with less than 2 outs, and the bases not loaded, the offense would always stay put at 3rd with the infield in, and score with the infield back. I see your point about the sharp ground ball to a player though - the runner might stay at 3rd in those cases. How often do you think that actually happens though? 1 in 20?
August 10, 2006 at 04:03PM View BBCode
Perhaps there should be a tiny higher chance at a triple. I think to hit a triple on the ground, its going to have to be a scorcher that gets by the outfielder, or the outfield will have to be playing the other way. I think these types of batted balls won't be caught by infielders playing in or back - what do you think?Don't teams also sometimes bring in their infielders when they're down, say, 2-0, to keep the game in reach?Yes, you'd be able to setup any combination of score, inning, situation you wanted to for playing the infield in.
Ok, here was my thought on this. I wanted to keep it somewhat simple and keep the Offense out of the equation. Meaning, we would somewhat simplify things and not have plays at the plate (unless the bases were loaded). Otherwise, the Offense would have to set a similiar set of preferences about when to run home and it gets 10 times more complex. So with less than 2 outs, and the bases not loaded, the offense would always stay put at 3rd with the infield in, and score with the infield back. I see your point about the sharp ground ball to a player though - the runner might stay at 3rd in those cases. How often do you think that actually happens though? 1 in 20?
ScooterPie
1. Triples. Like you, I'm thinking hot smash down the line. I do think the 1b or 3b might have cut if off if they were playing back, but I don't know whether it would have been a lineout, groundout, or infield single. Maybe bart has something more intelligent to say -- he started it, after all.
2. Singles/Doubles. Bringing in the corners increases the chance of a single or double, as you say. Obviously, bringing in the 2b and SS should further increase the chances of a single. But it seems to me that playing the 2b and SS in won't further increase the chances of a double very much. Sure, I can see a liner to the gap that the middle IF might have stopped, but the main double risk, I'd think, is a ball hit down the line.
3. Offense. Including offensive preferences is where things really got too complex in my head. You're probably wise to keep this to the defensive side, and let them drive the decision-making.
4. Infield back, runner holds anyway.
Oops, I'll have to finish later.
scooter
August 11, 2006 at 09:10AM View BBCode
A few thoughts; I hope I remember them all:1. Triples. Like you, I'm thinking hot smash down the line. I do think the 1b or 3b might have cut if off if they were playing back, but I don't know whether it would have been a lineout, groundout, or infield single. Maybe bart has something more intelligent to say -- he started it, after all.
2. Singles/Doubles. Bringing in the corners increases the chance of a single or double, as you say. Obviously, bringing in the 2b and SS should further increase the chances of a single. But it seems to me that playing the 2b and SS in won't further increase the chances of a double very much. Sure, I can see a liner to the gap that the middle IF might have stopped, but the main double risk, I'd think, is a ball hit down the line.
3. Offense. Including offensive preferences is where things really got too complex in my head. You're probably wise to keep this to the defensive side, and let them drive the decision-making.
4. Infield back, runner holds anyway.
I see your point about the sharp ground ball to a player though - the runner might stay at 3rd in those cases. How often do you think that actually happens though? 1 in 20?
Oops, I'll have to finish later.
scooter
tysonlowery
August 11, 2006 at 03:15PM View BBCode
I'm going to start on this on Monday. Let me know if anyone else has ideas before then.
ScooterPie
This is the last thing I wanted to spout off about. First, I'm totally not the guy to give you a "1 in X" quote. The answer is, I simply don't know.
Here's what I'm wondering. The question you pose is: with a man on third, less than two out, and the infield playing at its normal depth, how often does that runner score on a ground ball? Unless I'm thinking about it wrong, what we're talking about here is a plain ol' RBI groundout -- a play that doesn't exist in SD, but that I really really hope exists some day. The best thing to do, it seems to me, is figure out how to add in RBI groundouts and then just use that logic in these situations when the infield is back.
Hmm ... now I'm not sure how well this project would work without figuring out RBI groundouts -- otherwise, will we see a situation where a guy scores on a grounder only if it's late in a close game (because that's where the new logic is)?
That's all I have time for now. Happy to keep talking about this, if you want to keep going on this track. But maybe you want to see what other comments you get from the group, in case someone else brings up a more important point.
scooter
August 11, 2006 at 06:34PM View BBCode
Originally posted by tysonlowery
I see your point about the sharp ground ball to a player though - the runner might stay at 3rd in those cases. How often do you think that actually happens though? 1 in 20?
This is the last thing I wanted to spout off about. First, I'm totally not the guy to give you a "1 in X" quote. The answer is, I simply don't know.
Here's what I'm wondering. The question you pose is: with a man on third, less than two out, and the infield playing at its normal depth, how often does that runner score on a ground ball? Unless I'm thinking about it wrong, what we're talking about here is a plain ol' RBI groundout -- a play that doesn't exist in SD, but that I really really hope exists some day. The best thing to do, it seems to me, is figure out how to add in RBI groundouts and then just use that logic in these situations when the infield is back.
Hmm ... now I'm not sure how well this project would work without figuring out RBI groundouts -- otherwise, will we see a situation where a guy scores on a grounder only if it's late in a close game (because that's where the new logic is)?
That's all I have time for now. Happy to keep talking about this, if you want to keep going on this track. But maybe you want to see what other comments you get from the group, in case someone else brings up a more important point.
scooter
Closer
August 11, 2006 at 06:47PM View BBCode
I'm not sure I agree that the run should score automaically on a ball to 2nd or SS with the corners in. Shouldn't it revert to the same logic that is used now with a man on 3rd? What's the difference? I don't see any. It's only the corners that make a difference. If a grounder to 2nd or SS is automatic here, it should be automatic with the infield at normal depth. No??
Closer
August 11, 2006 at 06:50PM View BBCode
Oh, and, in all cases, rather than anything being automatic, could there be a coin-flip in each case, just weighted different for each situation, therefore adding more drama (and reality) to the matter. i.e for example only 3rd in, 70/30 a hit, 3 rd back 30/70 a hit, the figures are just for illustration.
tysonlowery
In order to add a coin toss in there, you have to way all the factors. One of those would be, "does the runner attempt to score"? In order to add that in, I would need to add in a whole series of prefs on the Offensive side as well and it gets hairy really fast. It goes from a 2 day project to a 2 week project.
I think by doing what I propose here, it would make it more realistic. It might leave some room for improvement, but I think we'd see less complaints.
August 11, 2006 at 07:20PM View BBCode
Shouldn't it revert to the same logic that is used now with a man on 3rd? What's the difference?I'm not sure any logic is being used there. At the moment, runners either never score from 3rd, or rarely score from 3rd when a ball is hit to an infielder. I'd have to dig through the code to figure out which it is, but either way I think this logic is very flawed and unrealistic. It gets quite a few complaints.
In order to add a coin toss in there, you have to way all the factors. One of those would be, "does the runner attempt to score"? In order to add that in, I would need to add in a whole series of prefs on the Offensive side as well and it gets hairy really fast. It goes from a 2 day project to a 2 week project.
I think by doing what I propose here, it would make it more realistic. It might leave some room for improvement, but I think we'd see less complaints.
Closer
August 11, 2006 at 10:42PM View BBCode
I was thinking more on the line of what we were chatting about last weeik. When you were giving a quick rundown on how ABE decides a single, double, etc. And I thought there was something you said about ABE using percentages as to when a guy scored from 2nd on a single or 1st on a double.. something to the effect he uses the percentage that is in MLB....or something like that. Thats where my head was on the post. Use a weighted or percentage type coin-flip to see does he or doesn't he score on the grouner to 2nd or SS. I guess you could have a certain set percentage that is then mitigated by speed (+4%A, +3%B, etc.), arm(-4%A, -3%B, etc.) or something like that. Just so an owner is never sure anyone will always score or not.
Closer
Have seen the ocassional fly out, throw out at the plate from 3rd probably more than the runner scoring from 3rd on a ground out.
Oh yeah, and you can just remove the steal home option from pref.s It has not been attempted once in the whole time I have been here and I've put out posts for anyone to respond to from time to time.
[Edited on 8-11-2006 by Closer]
August 11, 2006 at 10:47PM View BBCode
Oh, and my experience has been, that "rarely" is the right adjective, rather than "never". But, lately haven't been seeing it much at all and it is slipping towards never.Have seen the ocassional fly out, throw out at the plate from 3rd probably more than the runner scoring from 3rd on a ground out.
Oh yeah, and you can just remove the steal home option from pref.s It has not been attempted once in the whole time I have been here and I've put out posts for anyone to respond to from time to time.
[Edited on 8-11-2006 by Closer]
ScooterPie
I'd like to just get clear on what you're proposing to do. Am I right in understanding that:
1. For the most part, if a man is on third and the batter grounds out, the runner will continue to not score, just as it is now; except
2. When the situation is such that a manager might bring in the infield, but chooses not to, that runner will score automatically.
Or, I guess it could be
3. There will be no restrictions on when we can bring in the infield, so any time they're playing back, runners will score every time there's a grounder to the infield.
If it's 1 and 2, I can't get past the logical disconnect (yet). If it's 3, then I wonder whether it's going too far in the other direction. I throw these out for consideration.
Finally, I was thinking exactly what Closer asked: You apparently have data, based on 50+ years of MLB play, revealing how often runners score from first on a double etc. Are there no similar data on runners scoring on ground ball outs? Just asking -- me, I wouldn't have a clue where to even begin looking for this stuff. But I was hoping maybe you did.
scooter
August 12, 2006 at 08:54PM View BBCode
Originally posted by tysonloweryIt's "rarely." As far as I can tell, runners score from 3rd in the following two situations: a) when the bases are loaded, and it's a force; and b) if there are men on second and third and a grounder is hit to the right side -- since the man on second usually advances, I think the runner on third "thinks" he's forced. Usually in the (b) situation, nobody advances, but I have seen it happen. Wait a second ... okay, I just checked my collection of hen's teeth, and I do in fact have more of them than I've seen runners scoring on groundouts when the bases aren't loaded. Now that that's out of the way ...
At the moment, runners either never score from 3rd, or rarely score from 3rd when a ball is hit to an infielder.
I'd like to just get clear on what you're proposing to do. Am I right in understanding that:
1. For the most part, if a man is on third and the batter grounds out, the runner will continue to not score, just as it is now; except
2. When the situation is such that a manager might bring in the infield, but chooses not to, that runner will score automatically.
Or, I guess it could be
3. There will be no restrictions on when we can bring in the infield, so any time they're playing back, runners will score every time there's a grounder to the infield.
If it's 1 and 2, I can't get past the logical disconnect (yet). If it's 3, then I wonder whether it's going too far in the other direction. I throw these out for consideration.
Finally, I was thinking exactly what Closer asked: You apparently have data, based on 50+ years of MLB play, revealing how often runners score from first on a double etc. Are there no similar data on runners scoring on ground ball outs? Just asking -- me, I wouldn't have a clue where to even begin looking for this stuff. But I was hoping maybe you did.
scooter
tysonlowery
Closer - you're talking about the Grid System. For the purposes of this conversation, forget anything you know or think you know about the Grid system.
Here's the simple version I was thinking about.
Runners would automatically score from 3rd unless the defense was playing the infield in. In that case, the runner would stay put at 3rd.
This fixes these complaints:
"In the 9th inning, Abe threw the runner out at 1st while the winning run came in from 3rd".
"How come the runner on third NEVER seems to score?"
August 14, 2006 at 06:45PM View BBCode
Wow, I think I've really confused you guys. Sorry, I'm am horribly bad at explaining things once the code gets involved. Maybe I should start from the top.Closer - you're talking about the Grid System. For the purposes of this conversation, forget anything you know or think you know about the Grid system.
Here's the simple version I was thinking about.
Runners would automatically score from 3rd unless the defense was playing the infield in. In that case, the runner would stay put at 3rd.
This fixes these complaints:
"In the 9th inning, Abe threw the runner out at 1st while the winning run came in from 3rd".
"How come the runner on third NEVER seems to score?"
Finally, I was thinking exactly what Closer asked: You apparently have data, based on 50+ years of MLB play, revealing how often runners score from first on a double etc. Are there no similar data on runners scoring on ground ball outs? Just asking -- me, I wouldn't have a clue where to even begin looking for this stuff. But I was hoping maybe you did.There might be - I don't have the software anymore and don't really remember how it works anymore either. I had it 3 years ago, its called A.S.S. (Astros Statistical Software) and its really cool, but the learning curve was pretty steep. I think its moot anyhow because of the added layer of decisions on top - it just gets too damn complicated really fast. I'd rather just improve it somewhat - if you guys think that this would be an improvement.
ChillyWilly
August 14, 2006 at 06:50PM View BBCode
When you pull the infield in depending upon the situation you sometimes pull the outfielders in as well. This is basically when the winning run is on third. This should increase the chance of a hit of all types. However doubles should not go up if you bring in the middle infielders. The only type of hit that should go up would be a single, which in real life would a grounder that got through, or a lazy/broken bat fly that squeaks by. Triples can go up, but they should barely go up.
ScooterPie
Tyson, thanks for clarifying. I didn't realize how many birds this stone would kill; now I'm even more excited than I was before.
Okay, on the frequency of runners scoring from third. I did find one source that kind of backs into a number. I came across an article trying to quantify how much Ichiro's speed helps his team: http://www.diamond-mind.com/articles/ichiro.htm Here's the money quote:
"On the 25 ground balls with less than two out, Ichiro scored an impressive 20 times, though four came on infield errors, one on a second-to-first double play, and five more on force outs at second. Even though Ichiro may not have been the focal point on these plays, and even though other runners were put out, 20 of 25 is a high rate for scoring on infield grounders. Based on a quick review of the typical advance rates in these situations, we'll credit him with 7 extra bases." (Emphasis added)
So the author a) has access to expected frequency of scoring and b) is very impressed with an 80% rate. Me, I'd call the errors and the DP automatic, and we're at 15/20, or 75%. If the author is giving Ichiro 7 extra bases, then I think he's expecting the runner to score either 13/18 (72%) or 8/13 (62%) of the time. (Of course, there's no way for us to tell how often the infield was playing in.)
This is just the barest of beginnings, I know. But does it help at all to establish a baseline? Something more than "never," but less than "automatic"? For what it's worth, I think you could do a lot worse than picking a base frequency for scoring, say 60-70%, then putting in adjustments based on where the ball's hit (much easier for the 3B to hold the runner), runner speed, and fielder arm.
Of course, you decide which way you want to jump. I've exhausted all my ideas to find good hard data for now, but maybe I'll come up with something. You want I should keep trying?
scooter
August 15, 2006 at 06:11AM View BBCode
Originally posted by ChillyWillyHeh - I wasn't even gonna go there!
When you pull the infield in depending upon the situation you sometimes pull the outfielders in as well.
Tyson, thanks for clarifying. I didn't realize how many birds this stone would kill; now I'm even more excited than I was before.
Okay, on the frequency of runners scoring from third. I did find one source that kind of backs into a number. I came across an article trying to quantify how much Ichiro's speed helps his team: http://www.diamond-mind.com/articles/ichiro.htm Here's the money quote:
"On the 25 ground balls with less than two out, Ichiro scored an impressive 20 times, though four came on infield errors, one on a second-to-first double play, and five more on force outs at second. Even though Ichiro may not have been the focal point on these plays, and even though other runners were put out, 20 of 25 is a high rate for scoring on infield grounders. Based on a quick review of the typical advance rates in these situations, we'll credit him with 7 extra bases." (Emphasis added)
So the author a) has access to expected frequency of scoring and b) is very impressed with an 80% rate. Me, I'd call the errors and the DP automatic, and we're at 15/20, or 75%. If the author is giving Ichiro 7 extra bases, then I think he's expecting the runner to score either 13/18 (72%) or 8/13 (62%) of the time. (Of course, there's no way for us to tell how often the infield was playing in.)
This is just the barest of beginnings, I know. But does it help at all to establish a baseline? Something more than "never," but less than "automatic"? For what it's worth, I think you could do a lot worse than picking a base frequency for scoring, say 60-70%, then putting in adjustments based on where the ball's hit (much easier for the 3B to hold the runner), runner speed, and fielder arm.
Of course, you decide which way you want to jump. I've exhausted all my ideas to find good hard data for now, but maybe I'll come up with something. You want I should keep trying?
scooter
Closer
August 15, 2006 at 12:29PM View BBCode
Tyson, there definitely should be more men scoring from third on ground outs. But, I wince at the idea of "always" scoring from third on a ground out. Would it be that involved to have a set percentage of times the runner will score. Then "flip the coin", but have the percentage mitigated by the speed of the man on third. A+ percentage stays the same. A = a small drop in percentage, A- a further drop, B+ further, etc. This way there still is the chance the guy won't score also. And this could make or break a game. Also it rewards a team speed.
tysonlowery
August 15, 2006 at 02:26PM View BBCode
I'm not sure Closer, let me start programming these changes and see what I can figure out. I guess if the options are just stay put or score, we might be able to get around having Offensive settings for these situations. But if we add "Thrown out trying to score", that's when I don't think its fair to NOT have Offensive settings. We'll get some guy complaining about how his runner was thrown out and how you should never go home in that situation, blah blah blah :)
Closer
August 15, 2006 at 02:39PM View BBCode
I hear ya. No matter what you do.....there will be objections, no doubt! I guess what I'm saying is a score or no score situation....but adding the element that no one can be sure that in every instance the man from 3rd will score on a gound out. So in my previous post I'm talking about the mitigating speed factor just giving the man on 3rd more likelyhood to go and score than the slower guy going and scoring, but neither is guaranteed going at all. That clarify at all? Seem reasonable and simple? Not including the defense factor at all, just speed and a base percentage that is based on MLB or whatever base decided on.Pages: 1