Admin
Rewriting the Pitcher/Batter matchup
February 01, 2011 at 07:42PM View BBCode
As you may know, in most cases batters and pitchers are weighted evenly when calculating the odds of a particular event happening in the sim. For instance, the pitcher and the batter contribute equally to the formula to determine the odds that the matchup will result in a HR.
Based on studies such as this: http://www.beyondtheboxscore.com/2006/8/25/83511/7939 and this: http://www.philbirnbaum.com/btn1990-03.pdf - I have come to realize that this shouldn't be set at 50%.
Cyril Morong, the author of the first article, was kind enough to share his spreadsheets with me and give me some advice. Using the methods he outlined, and data from the 2010 season, I've calculated the % of variance contributed by the pitcher or batter for a given event. Here are the numbers I came up with. This will likely be part of a SABR article that I write, so I might as well share this with everyone:
Event: Batters Influence
1B: 63%
2B: 66%
3B: 60%
HR: 69%
BB: 60%
SO: 63%
HBP: 59%
Something that I stumbled upon is how important the batter is for drawing a HBP. In the sim today, the batter doesn't influence this outcome at all. It is 100% based on pitcher control.
The other area that I've seen evidence for, but which is tougher to quantify, is that really good pitchers tend to have a larger influence on outcomes.
With this information in mind, I have written 2 new options for leagues into the sim.
The first option alters the batter/pitcher influence on these events based on the percentages above. For HBP, batters contact will play a significant role and leadership will play a small role in the formula. Looking over the HBP rates in 2010, there seems to be some players that get hit a lot and others who don't. I think leadership is a nice way to sprinkle in whether someone is willing to take one for the team.
The 2nd option modifies these percentages based on the skill level of the pitcher. So a dominant pitcher will continue to influence these events at a clip of around 50%, but a weaker pitcher might be closer to 25%.
How will these options impact leagues? I am hoping for a few results:
1) Bad pitchers in SD are too bad
If you look at the variance in ERAs in SD vs MLB, there are far too many pitchers with ERAs over 6. Because these types of pitchers will have much less influence on the outcome, I suspect this will cause their ERAs to go to a more realistic level.
2) Higher league leader stat totals for batters
Because we are giving more influence to batters, this should make it more likely that a superstar hitter will hit 40 HRs, for instance. Or even 50 HRs. These totals are practically unattainable in SD with the current structure. That could change.
3) Rod Carews?
Because batters have more influence, I could see the high contact, low power hitters put up stats like a Wade Boggs or Rod Carew. We shall see what happens.
These 2 options are turned on for all the beta leagues. They will be turned off by default in all leagues when this change comes over. In order to make this change, a private/speed league would need unanimous consent. Trial and Dynasty leagues will not use this option.
As a tangent, there are several systems (perhaps this one is included) that I feel are superior to the default systems being used in the Dynasty leagues. It might be time to explore how we can turn these options on in the Dynasty leagues without impacting competitive strategies.
Tyson
Hamilton2
February 01, 2011 at 08:44PM View BBCode
Whoah. That is some crazy awesome cool stuff right there. Nice work, Tyson. I look forward to seeing the impact in the beta leagues.
uteric7
February 01, 2011 at 10:07PM View BBCode
props to you man...thanks for taking the time to check through all of this...one of the leagues im in is already talking about it....i hope we can make the switch and give it a try.
Admin
February 01, 2011 at 10:17PM View BBCode
I'm not sure if the changes will achieve the results I listed above, or maybe have some different unintented consequences. But that is what beta testing is for.
Tyson
the_soup
February 01, 2011 at 10:30PM View BBCode
Very interesting posting Tyson. I have noticed a lot of B+ starting pitchers with lousy ERA's. And hitters with over 40 homers in a season is rare indeed. Having B+ SP lower than 6.00 and league leaders in homers over 40 will definitely improve the SimD leagues.
Look forward to hearing about the beta results!
DW_Geoff
February 01, 2011 at 10:31PM View BBCode
Sounds good, though I think one of the reasons for ERAs over 6 is that many owners throw A or AA pitchers in with the attempts to tank
Admin
February 01, 2011 at 10:34PM View BBCode
Yeah, I'm not sure what the effects will be. I just know that this part isn't as realistic as it could be and perhaps correcting that will right some other areas that aren't realistic - but who knows.
Tyson
cdunn3
February 01, 2011 at 11:05PM View BBCode
Awesome possibility.
Based on 6 enthusiatic replies in a couple of hours,
The Greg Maddux League (which just had 100 percent
renewal) looks forward to site implementation.
tworoosters
February 01, 2011 at 11:29PM View BBCode
The BHL hasn't had a 40 homer season in 17 years, this season the league leader had 34 and only 5 players in the league hit 30 or more .
imalbundy2
February 02, 2011 at 07:24AM View BBCode
this is all fine and dandy, but how about giving bench pitchers the same ML improvement chances as bench hitters?
if you bring up a pitcher and don't use him, he gets ZERO improvment chances.
lvnwrth
February 02, 2011 at 03:39PM View BBCode
Is there any way that this might give us more realisitic K/IP totals as well?
Closer
February 02, 2011 at 04:23PM View BBCode
True there is no telling what the effects will be..........but I think it is definitely worth implementing to find out. I agree with the theory behind the changes and am pretty sure it can only do good.
Closer
February 02, 2011 at 04:23PM View BBCode
True there is no telling what the effects will be..........but I think it is definitely worth implementing to find out. I agree with the theory behind the changes and am pretty sure it can only do good.
Admin
February 02, 2011 at 05:00PM View BBCode
Is there any way that this might give us more realisitic K/IP totals as well?
I'm not sure if it will, because a dominant pitcher will only still have 50% control over strikeouts.
Let me look at the base percentage of strikeouts in the sim today, maybe that number needs updating.
The other thing I could look at is do an analysis of batter vs pitcher strikeouts. I can group the batters and pitchers by how much they strikeout, then compare how the groups do when facing one another. It is quite possible that someone else has already done this study too - I'll hunt for it as well.
Tyson
tworoosters
February 02, 2011 at 05:09PM View BBCode
Whatever is deflating the K/IP numbers it's doing an effective job, I've never seen a SP in SIM have a ratio approaching 1 K/1 IP, the best I can recall was 232 K in 259.2 by [url=http://www.simdynasty.com/player.jsp?id=7181191&statsorimps=stats]Stirnweiss[/url], who at the time was A+/A+ with the control no less than 98 .
My three current leagues have logged over 200 seasons without a pitcher exceeding 250 strikeouts in a single season .
[Edited on 2-2-2011 by tworoosters]
tm4559
February 02, 2011 at 05:46PM View BBCode
they are coming closer since the k's were turned up a little while back
http://simdynasty.com/player.jsp?id=7235031
252 on 270 ip (the control isn't really the place to look for the strikeouts, the velocity is i guess). it is worth noting i suppose that is the only 250+ K season in the GNL. one season of 249 and three of 248. one of which was a truly beautiful outlier all the way back in 1954, 248 K in 270 ip
http://simdynasty.com/player.jsp?id=1360521
i know there are a number of pitchers in real baseball that do the one to one. there really are not a lot of them, though are there? so, the K totals are not really way off, they just aren't high enough for a few truly great pitchers, they should be higher for those. right? but not more for the great high middling mediocre majority. the challenge is to keep those where they are, because they are in fact about right, aren't they?
Admin
February 02, 2011 at 06:26PM View BBCode
I looked at the dynasty leagues for last year, compared to MLB for 2010. In MLB, there was 34306 Ks in 43158 IP, or .795 Ks per IP. In SD, there was .767 Ks per IP. That's about 4% shy of where it should be, so I've attempted to increase strikeouts by 4%. We will see how that works out.
I am now looking at the standard deviation among pitchers to see how far we are off there.
Tyson
Admin
February 02, 2011 at 08:04PM View BBCode
Here's what I found on strikeouts. There is a huge disparity in the amount of variance in MLB pitchers vs SD pitchers, SD pitchers are too close to the mean. The same goes for batters, but the disparity isn't as high.
I've made another a couple of other adjustments for strikeouts, one of the changes will apply to all leagues, the other to leagues that use these new systems.
Tyson
[Edited on 2-2-2011 by Admin]
Hamilton2
from 6-11-10
February 02, 2011 at 08:44PM View BBCode
Originally Posted by Hamilton2
A note about strikeout numbers, since they have been referenced here. I've looked into the numbers behind this a little bit and have found two interesting statistical points.
A) The top-end SD pitchers perform at a far lower K/9 rate than the top-end MLB pitchers do.
B) The league average K/9 for SD is significantly higher than that of MLB.
This means, to me, that we are operating without sufficient variance in the strikeout category. Perhaps this is because a lot of the strikeouts come from the hitter's contact/power bundle rather than from the pitcher. Perhaps it is because there aren't enough degrees of separation from the best SD pitcher and the league scrubs. I don't really know why, but it is certainly interesting.
From [url=http://www.simdynasty.com/oldforum-viewthread.jsp?tid=267937&page=2]page 2 of this deadend thread[/url].
I was looking at data from 1950 through 2000 for my analysis, which indicates that the K/9 rates have gone up, on average, since 2000. The lack of any meaningful variance in the K/9 stats has been a problem for a long time. I'm glad that this is getting looked at and addressed. Thanks, Tyson!
Aparicio
February 03, 2011 at 01:26AM View BBCode
I can't pretend to understand all of the Math, but we DEFINITELY need a few "Babe Ruths".
And, yes, too many Pitchers are averaging over 6 ERA. They would be out of the league!
Hit By Pitch is a very minor refinement, though a welcome consideration.
My Main Input to this discussion would be:
How do we make FIELDING a more realistic aspect??! It doesn;t seem to influence games like it should. Very remarkable defensive players could be late inning substitutes or even preferred starters, if they really mattered. At this point, you throw out your "Big Hitters" onto the diamond.
We are ignoring a very important aspect of the game.
-AParicio
Aparicio
Why aren't you concentrating on Defense?
February 03, 2011 at 01:32AM View BBCode
Taking away hits needs to be applauded, guaranteeing a Louis Aparicio a spot in the lineup, as he makes unbelievable plays to prevent a big inning.
-AParicio
PS. Ozzie Smith, anyone??!
lvnwrth
February 03, 2011 at 01:35AM View BBCode
Originally posted by tm4559
i know there are a number of pitchers in real baseball that do the one to one. there really are not a lot of them, though are there? so, the K totals are not really way off, they just aren't high enough for a few truly great pitchers, they should be higher for those. right? but not more for the great high middling mediocre majority. the challenge is to keep those where they are, because they are in fact about right, aren't they?
Yes, there are a lot of them...or at least enough that you can't really consider them rare. In 2010 there were 47 pitchers with 60+ innings who averaged 9 K per 9 IP or better. 20 pitchers averaged better than 10/9; the top 6 guys were 12/9; and the MLB leader was 15/9.
In 2009 there were 46 pitchers with 60+ innings who averaged 9/9. 23 were 10/9, and the league leader was 13/9.
The NL league average was right at 7K/9 in those two years.
Without going back and doing league totals in SD, it appears you're right about overall totals. In Beta last season, the league average was 6.8K/9IP. So it doesn't appear to be a problem with total K...just with the guys at the high end of the spectrum.
Hamilton2
February 03, 2011 at 01:40AM View BBCode
Originally posted by Aparicio
Taking away hits needs to be applauded, guaranteeing a Louis Aparicio a spot in the lineup, as he makes unbelievable plays to prevent a big inning.
-AParicio
PS. Ozzie Smith, anyone??!
I'd like to see the statistical analysis of baseball that places fielding at greater than 10% significance in influencing outcomes. As far as I know fielding matters very little in the scheme of things for MLB. Since this is a statistical simulation it is dangerous to put an over emphasis on factors that don't really matter statistically.
Pages: 1 2