 |
View Old Forum Thread
Old Forum Index » Baseball Beta Testing » Beta News » More Research into the Batter/Pitcher Matchup and Stretching
Admin
I've been looking at all sorts of things, such as how fast players 2B/3B ratio differ from that of slow players.
Here's a question I tried to answer today that may be of some interest. How often does a guy with little or no speed hit a triple? I looked at all MLB player seasons from 1995-2008 who had 400 at bats and 0 SB. That seemed like a good way to get a good list of slow players. I found 233 players. Lots of catchers and firstbasemen. I sorted by triples and found 24 players who hit 4 or more (4-6) triples in a season. I noticed a few guys on the list that might not be fast, but couldn't be described as glacier (Jose Offerman and Troy O'Leary for instance). I decided to add another constraint that said a guy had to be caught stealing less than 4 times. That removed most of the not-glacier guys, except for a few like Alex Gonzalez. I decided to leave them in the data set to account for any truly slow guys that might have actually stolen a base by accident.
These players got a triple every .24% of plate appearances. They batted .279. For all players from 1995-2008, they got a triple .52% of the time and batted .283. Based on these numbers, you can attribute more than half of triples to speed.
My goal is to break the at bat into a few more buckets than we have today and then use Speed as a modifier after the initial calculation to shift the buckets. This is where stretching will come in, as there will be buckets for stretching.
Lots more to come...
Tyson
More Research into the Batter/Pitcher Matchup and Stretching
February 08, 2011 at 07:29PM View BBCode
I've been knee-deep in spreadsheets the last week looking at other ways to improve the batter-pitcher matchup.I've been looking at all sorts of things, such as how fast players 2B/3B ratio differ from that of slow players.
Here's a question I tried to answer today that may be of some interest. How often does a guy with little or no speed hit a triple? I looked at all MLB player seasons from 1995-2008 who had 400 at bats and 0 SB. That seemed like a good way to get a good list of slow players. I found 233 players. Lots of catchers and firstbasemen. I sorted by triples and found 24 players who hit 4 or more (4-6) triples in a season. I noticed a few guys on the list that might not be fast, but couldn't be described as glacier (Jose Offerman and Troy O'Leary for instance). I decided to add another constraint that said a guy had to be caught stealing less than 4 times. That removed most of the not-glacier guys, except for a few like Alex Gonzalez. I decided to leave them in the data set to account for any truly slow guys that might have actually stolen a base by accident.
These players got a triple every .24% of plate appearances. They batted .279. For all players from 1995-2008, they got a triple .52% of the time and batted .283. Based on these numbers, you can attribute more than half of triples to speed.
My goal is to break the at bat into a few more buckets than we have today and then use Speed as a modifier after the initial calculation to shift the buckets. This is where stretching will come in, as there will be buckets for stretching.
Lots more to come...
Tyson
Admin
How many doubles can a speedy player turn into triples? The league averages between .49% and .55% triples as a percentage of plate appearances. I'm using .52% as a target. Since we know from above that players without speed will hit about .24% triples, that means .28% of triples, on average, are stretched from doubles. For a speedy player, I looked at the best 3B percentages of the last 15 years. 1.5% seemed to be about the upper limit when throwing out the outliers. So that says that the fastest players have about 1.26% of their at bats where it is a double that they stretch into a triple.
Looking at this another way. An average player will hit .24% triples, stretch a double into a triple .28% of the time, and 1% of the time hit a double where a fast player would have stretched it into a triple. The fastest player will hit .24% triples and hit a double that is stretched into a triple about 1.26% of the time.
Incidentally, runners are thrown out trying to stretch a double into a triple .77% of the time they hit a double. Or put on the same terms as above, .04% of plate appearances are doubles where the runner is out at 3B.
Tyson
[Edited on 2-9-2011 by Admin]
February 09, 2011 at 06:04PM View BBCode
Here's some more of my thought process. If you see any reasoning that looks flawed, please pipe in.How many doubles can a speedy player turn into triples? The league averages between .49% and .55% triples as a percentage of plate appearances. I'm using .52% as a target. Since we know from above that players without speed will hit about .24% triples, that means .28% of triples, on average, are stretched from doubles. For a speedy player, I looked at the best 3B percentages of the last 15 years. 1.5% seemed to be about the upper limit when throwing out the outliers. So that says that the fastest players have about 1.26% of their at bats where it is a double that they stretch into a triple.
Looking at this another way. An average player will hit .24% triples, stretch a double into a triple .28% of the time, and 1% of the time hit a double where a fast player would have stretched it into a triple. The fastest player will hit .24% triples and hit a double that is stretched into a triple about 1.26% of the time.
Incidentally, runners are thrown out trying to stretch a double into a triple .77% of the time they hit a double. Or put on the same terms as above, .04% of plate appearances are doubles where the runner is out at 3B.
Tyson
[Edited on 2-9-2011 by Admin]
greco4170
Idea about baserunning and basestealing
February 09, 2011 at 10:37PM View BBCode
Some players are great basestealers but not fast in the major leagues and some are great baserunner but not great getting a jump off a pitchers could there be a different rating for baserunning and basestealing a guy like David Wright is a good basestealer but not a great baserunner so maybe two different ratings would give you more ways to break it up
Admin
Ok, the next step of this madness. I tried to figure out how many singles were "strechable". I couldn't figure out a great way to get at this statistically, but I made some assumptions and I think the number I got was pretty good.
My hypothesis is that the percentage of doubles that are stretchable or streched shouldn't be that much different than the percentage of singles that are stretchable or stretched. The reasoning being you are talking about getting an extra base and the decision process for deciding whether to go for it is similar. I thought it would be higher because players with lesser baserunning skills might be more apt to try this, and there is often a greater benefit at being at 2B vs. 1B, instead of 3B vs. 2B.
I looked at the percentage of doubles where the batter was thrown out trying to stretch it into a triple (from 1995 to 2008). I modified the numbers from the other day to eliminate ground rule doubles from the data set. 0.75% of doubles resulted in the batter being thrown out at third. I then did the same thing for singles. which worked out to 0.82%. This includes some clearly unstretchable singles, such as infield hits. In 2010, 13.06% of singles were infield hits. If you subtract out those numbers, it goes to 0.96%. So there are slightly more of those, 28% more. I then applied this 128% to the numbers I found for stretching 2B into 3B to figure out how many strechable singles there were, and how many were stretched.
At this point, I have the percentages for how many 1B are stretched into 2B and how many 2B could be stretched into 3B but go unstretched. I subtracted these from the total percentage of 2B to get how many plain vanilla doubles there are.
I then took a look at what percentage of plate appearances result in a 2B for fast players (more than 10 SB). The highest guys were 8-9%, with 7% seeming to be below the outliers. The fastest player will stretch all 1B into 2B and all 2B into 3B. Adding these numbers up in my matrix gives me 7.19%. That looked good so I think these numbers are close.
Finally, I looked at how many Infield singles there were. League average in 2010 is that 2.09% of plate appearances result in an infield single. I then looked at the league leaders for infield single % and found the leaders are around 4%. So that tells me that about 4% of ground balls are eligible to be stretched into a single, with the average player having half of those singles and half of those ground outs.
Next, I'm going to see if I can figure out how having an available force out affects the percentage of infield singles.
Tyson
February 10, 2011 at 08:02PM View BBCode
Hi Greco - that probably belongs in the Enhancements area.Ok, the next step of this madness. I tried to figure out how many singles were "strechable". I couldn't figure out a great way to get at this statistically, but I made some assumptions and I think the number I got was pretty good.
My hypothesis is that the percentage of doubles that are stretchable or streched shouldn't be that much different than the percentage of singles that are stretchable or stretched. The reasoning being you are talking about getting an extra base and the decision process for deciding whether to go for it is similar. I thought it would be higher because players with lesser baserunning skills might be more apt to try this, and there is often a greater benefit at being at 2B vs. 1B, instead of 3B vs. 2B.
I looked at the percentage of doubles where the batter was thrown out trying to stretch it into a triple (from 1995 to 2008). I modified the numbers from the other day to eliminate ground rule doubles from the data set. 0.75% of doubles resulted in the batter being thrown out at third. I then did the same thing for singles. which worked out to 0.82%. This includes some clearly unstretchable singles, such as infield hits. In 2010, 13.06% of singles were infield hits. If you subtract out those numbers, it goes to 0.96%. So there are slightly more of those, 28% more. I then applied this 128% to the numbers I found for stretching 2B into 3B to figure out how many strechable singles there were, and how many were stretched.
At this point, I have the percentages for how many 1B are stretched into 2B and how many 2B could be stretched into 3B but go unstretched. I subtracted these from the total percentage of 2B to get how many plain vanilla doubles there are.
I then took a look at what percentage of plate appearances result in a 2B for fast players (more than 10 SB). The highest guys were 8-9%, with 7% seeming to be below the outliers. The fastest player will stretch all 1B into 2B and all 2B into 3B. Adding these numbers up in my matrix gives me 7.19%. That looked good so I think these numbers are close.
Finally, I looked at how many Infield singles there were. League average in 2010 is that 2.09% of plate appearances result in an infield single. I then looked at the league leaders for infield single % and found the leaders are around 4%. So that tells me that about 4% of ground balls are eligible to be stretched into a single, with the average player having half of those singles and half of those ground outs.
Next, I'm going to see if I can figure out how having an available force out affects the percentage of infield singles.
Tyson
Admin
I looked at 5 different scenarios. Runner on 1st only, runner on 2nd or 3rd but not 1st, bases empty, runner on 1st and 2nd but not 3rd, and bases loaded.
I expected there to be more infield hits with the bases empty. But there are actually more when there is a runner on 2nd or 3rd but no force out. Maybe because the defense is reluctant to throw to first in some cases so the runner won't advance. Or when the defense goes for the play at the plate and is unsuccessful. Here are the numbers, as a percentage of non-bunt hits:
Runner on 2nd or 3rd but not 1st: 13.09%
Bases Empty: 12.71%
Runners on 1st and 2nd but not 3rd: 11.15%
Bases Loaded: 11.09%
Runner on 1st only: 9.69%
It is a bit counter intuitive that adding another force out actually increases the number of infield hits. Meaning, I'm surprised the Runner on 1st Only ranked behind 1st and 2nd, and bases loaded. With runners on 1st and 2nd, perhaps the defense is more likely to "put the ball in their pocket" on a grounder rather than chance an error.
I will use some form of these numbers when working out infield hits in the sim.
On a tangent, I spent quite a bit of time looking at bunt singles from non-pitchers with the bases empty. I'll likely publish this as an article of some kind, but it doesn't have much relevance to the sim. Here are a couple of figures I'll use though:
When a bunter attempting to bunt for a base hit actually puts the ball into play, he is safe at 1st about 43.8% of the time. The downside of attempting to bunt is that sometimes you'll use an a pitch in an at bat to try to bunt but take a strike or foul a pitch off, and abandon the bunt. In those cases, the batter ends up hitting about 0.229. I may use these numbers to modify how bunting for a base hit works in the sim.
The next thing I'm going to look at is pitcher sacrifice bunt success.
Tyson
March 02, 2011 at 07:24PM View BBCode
Ok, I took a closer look at infield singles. For 1995-2008 data, 11.74% of singles were infield hits (11.99% of non-bunt singles are infield hits). I'm going to use that figure instead of the 13.38% number mentioned above because that number was only for 2010 and may have included bunt hits.I looked at 5 different scenarios. Runner on 1st only, runner on 2nd or 3rd but not 1st, bases empty, runner on 1st and 2nd but not 3rd, and bases loaded.
I expected there to be more infield hits with the bases empty. But there are actually more when there is a runner on 2nd or 3rd but no force out. Maybe because the defense is reluctant to throw to first in some cases so the runner won't advance. Or when the defense goes for the play at the plate and is unsuccessful. Here are the numbers, as a percentage of non-bunt hits:
Runner on 2nd or 3rd but not 1st: 13.09%
Bases Empty: 12.71%
Runners on 1st and 2nd but not 3rd: 11.15%
Bases Loaded: 11.09%
Runner on 1st only: 9.69%
It is a bit counter intuitive that adding another force out actually increases the number of infield hits. Meaning, I'm surprised the Runner on 1st Only ranked behind 1st and 2nd, and bases loaded. With runners on 1st and 2nd, perhaps the defense is more likely to "put the ball in their pocket" on a grounder rather than chance an error.
I will use some form of these numbers when working out infield hits in the sim.
On a tangent, I spent quite a bit of time looking at bunt singles from non-pitchers with the bases empty. I'll likely publish this as an article of some kind, but it doesn't have much relevance to the sim. Here are a couple of figures I'll use though:
When a bunter attempting to bunt for a base hit actually puts the ball into play, he is safe at 1st about 43.8% of the time. The downside of attempting to bunt is that sometimes you'll use an a pitch in an at bat to try to bunt but take a strike or foul a pitch off, and abandon the bunt. In those cases, the batter ends up hitting about 0.229. I may use these numbers to modify how bunting for a base hit works in the sim.
The next thing I'm going to look at is pitcher sacrifice bunt success.
Tyson
redcped
I'm not personally huge on bunting, but I would like to see some settings, particularly with dropdowns for selecting 9th or extra innings, where when a guy reaches base with none or one out, you can move him into scoring position.
March 02, 2011 at 08:00PM View BBCode
Tyson, since you are looking into bunting, does this mean you might be implementing the sacrifice for non-pitchers sometime?I'm not personally huge on bunting, but I would like to see some settings, particularly with dropdowns for selecting 9th or extra innings, where when a guy reaches base with none or one out, you can move him into scoring position.
tm4559
March 11, 2011 at 03:49PM View BBCode
you know you have the strikeouts turned up too high, right? on my team there are three starters (rowe, conwell and foreman) and one reliever (stankweitz) where k is as much as ip, or more. there is one other starter that is just short (mee, 198 ip, 196 K) and one reliever (greene, 38 ip, 36 K). this clearly too many, no? i mean, they aren't even good, one of them is rated 84 (conwell) and one 75 (stankewitz). the rest of them are blue rated. just mentioning it, in case you turned your attention away to something else. this is far too many strikeouts, you can't be finished with it.
Admin
Tyson
March 11, 2011 at 04:38PM View BBCode
I haven't gone back yet and looked at the results from changing strikeouts. I will though over the next couple months.Tyson
tm4559
March 11, 2011 at 04:53PM View BBCode
oh cool, just wanted to make sure. it is pretty obvious from the ASL stats also they are just too high. this other stuff, the stretching and stuff, is actually more interesting, and its great you're doing some stuff with it!
tm4559
March 11, 2011 at 04:58PM View BBCode
(what i mean to say is, this stuff with the strikeouts is just about distributing the same outs a different way, more or less, if i understand it right, just a stats question. the stetching stuff [and the stuff on doubles and triples] is something that can i guess can actually change some outcomes, which is far, far more fun).
Admin
Tyson
March 11, 2011 at 05:36PM View BBCode
The thing I need to figure out is how this will impact existing leagues. I'm not sure yet whether speed will become more valuable or less valuable, that is something we need to take a close look at. Stretching is going to be grouped with the changes to how singles, doubles, and triples are calculated. So I won't be able to turn on stretching without turning on these other core changes.Tyson
tm4559
March 11, 2011 at 07:12PM View formatted
You are viewing the raw post code; this allows you to copy a message with BBCode formatting intact.
cool. let us know if you want to try some of out in beta (if you are not already doing so). we can put some different arm combos in the outfield, or speed in the lineup, whatever.
ballplayer3
http://beta.simdynasty.com/player.jsp?mode=player&playername=nobody&id=63073
March 14, 2011 at 11:55PM View BBCode
Check it out. Tony Heard finished the season with a 24-0 record!http://beta.simdynasty.com/player.jsp?mode=player&playername=nobody&id=63073
tm4559
March 15, 2011 at 09:55PM View BBCode
had one to do this in the wsl one time, i think it was 27-0. went out and lost the opening game of the playoffs, naturally.Pages: 1