 |
View Old Forum Thread
Old Forum Index » Baseball Beta Testing » Beta News » New MVP Formula
Admin
I'm going to run it against some prior seasons in Beta to test it to make sure it works. So whoever the MVP was might not be it anymore.
Here's the article about the formula:
http://www.simdynasty.com/mvp.jsp
I didn't quite use barterer's formula exactly, but it is really close. But here is what I did do:
Hits * 1.5
2B
3B
HR * 2.5
RBI * 2.5
BB
Runs * 2
AVG: +/-2 points per pct above/below .300
OBP: +/-1.25 points per pct above/below .400
SLG: +/-1.5 points per pct above/below .475
SB
Playoffs or within 5 games of making the playoffs: +150
90 Wins: +60
100 Wins: +135
110 Wins: +235
115 Wins: +285
If in 1st place: games ahead * 2
Defensive position bonuses: SS(190), C(185), 2B(130), RF(140), 3B(30), CF(30)
Team improvement: +20 per spot in standings improved
Team declines: -20 per spot in standings the team fell
Previous MVP: -100
-2 per game back of 1st, if more than 10 games back
New to the team this season: +50
If HR >= 50, 1 point per HR
If RBI >= 140, .333 points per RBI
If SB > 50, 1 point per SB
League leader in HR, RBI, AVG: +10 per category
League leader in R or SB: +5 per category
A couple logistical notes. This is way more complex than the old formula, so I think it will run once per day for all leagues in the middle of the night. The wins bonuses are NOT prorated. So once you get 90 wins, you will earn that bonus. The playoff bonus is within 5 wins of the 2nd place team - so if there is a tiebreaker, you'll need to be within 5 wins of the team that won the tiebreaker.
You need 81 games at a position to earn the position bonus.
The formula only takes into consideration players who have at least 100 all star votes. So early in the season, there won't be any players on the list. I am considering only running it once a league has played 100 games - may lead to some additional suspense.
I'll let you know when I've run it and think it looks correct.
Tyson
New MVP Formula
April 03, 2009 at 06:17PM View BBCode
I've been working the last few days on setting up the MVP formula that barterer2002 came up with to work on Beta.I'm going to run it against some prior seasons in Beta to test it to make sure it works. So whoever the MVP was might not be it anymore.
Here's the article about the formula:
http://www.simdynasty.com/mvp.jsp
I didn't quite use barterer's formula exactly, but it is really close. But here is what I did do:
Hits * 1.5
2B
3B
HR * 2.5
RBI * 2.5
BB
Runs * 2
AVG: +/-2 points per pct above/below .300
OBP: +/-1.25 points per pct above/below .400
SLG: +/-1.5 points per pct above/below .475
SB
Playoffs or within 5 games of making the playoffs: +150
90 Wins: +60
100 Wins: +135
110 Wins: +235
115 Wins: +285
If in 1st place: games ahead * 2
Defensive position bonuses: SS(190), C(185), 2B(130), RF(140), 3B(30), CF(30)
Team improvement: +20 per spot in standings improved
Team declines: -20 per spot in standings the team fell
Previous MVP: -100
-2 per game back of 1st, if more than 10 games back
New to the team this season: +50
If HR >= 50, 1 point per HR
If RBI >= 140, .333 points per RBI
If SB > 50, 1 point per SB
League leader in HR, RBI, AVG: +10 per category
League leader in R or SB: +5 per category
A couple logistical notes. This is way more complex than the old formula, so I think it will run once per day for all leagues in the middle of the night. The wins bonuses are NOT prorated. So once you get 90 wins, you will earn that bonus. The playoff bonus is within 5 wins of the 2nd place team - so if there is a tiebreaker, you'll need to be within 5 wins of the team that won the tiebreaker.
You need 81 games at a position to earn the position bonus.
The formula only takes into consideration players who have at least 100 all star votes. So early in the season, there won't be any players on the list. I am considering only running it once a league has played 100 games - may lead to some additional suspense.
I'll let you know when I've run it and think it looks correct.
Tyson
Jughead
I like that a lot.
In the unlikely event a player with A+ health plays in all 162 games, to prevent him from receiving two position bonuses by playing 81 games at each, could you make the position bonus threshold 82 games?
April 03, 2009 at 06:28PM View BBCode
" I am considering only running it once a league has played 100 games - may lead to some additional suspense."I like that a lot.
In the unlikely event a player with A+ health plays in all 162 games, to prevent him from receiving two position bonuses by playing 81 games at each, could you make the position bonus threshold 82 games?
tworoosters
Historically these positions have not won a lot of MVP awards so I don't think the formula should be weighted as much as it is.
April 03, 2009 at 06:33PM View BBCode
We'll have to wait and see but I still have a problem with the large "up the middle" bonus for C, SS and 2B.Historically these positions have not won a lot of MVP awards so I don't think the formula should be weighted as much as it is.
Admin
Barterer's formula got many more right than my formula ever did. I don't think he liked the position bonuses either, but after looking at the data over and over again, there's no denying that the voters give credence to which position you play.
Tyson
April 03, 2009 at 06:55PM View BBCode
You'll only get credit for one position.Barterer's formula got many more right than my formula ever did. I don't think he liked the position bonuses either, but after looking at the data over and over again, there's no denying that the voters give credence to which position you play.
Tyson
Admin
Note that I didn't run the code to actually "set" the MVP, so it won't show up on the player card for last season. But you can see it from the All Star page.
I've set it up to run at 4:15AM for leagues that have passed 100 games. Right now, beta has been using the old MVP formula for the games played thus far - so those numbers will stay until we get to 100 games.
Tyson
April 03, 2009 at 07:44PM View BBCode
Ok, I ran this for last season on the site and checked the numbers. Seems to be working correctly.Note that I didn't run the code to actually "set" the MVP, so it won't show up on the player card for last season. But you can see it from the All Star page.
I've set it up to run at 4:15AM for leagues that have passed 100 games. Right now, beta has been using the old MVP formula for the games played thus far - so those numbers will stay until we get to 100 games.
Tyson
barterer2002
The truth of the matter, historically speaking is that hitters at C, 2B and SS can put up numbers that are less than those who play 1B or OF. A guy playing one of those positions (see Rodriguez, Ivan or Larkin, Barry or Kent, Jeff) who puts up numbers that are in the ballpark but not all the way up to what the raw statistics would say for a 1B is generally going to win the MVP. The exception that I've seen is Justin Morneau who somehow won over Carlos Guillen, Derek Jeter and Joe Mauer.
Also note, there is a significant bump for players who play RF which is drastically different than what SD currently has. Thus what we're seeing is that a catcher vs. a RF only gains about 45 points based on defense.
Its also important to note that typically the MVP goes to a player with more than 1500 points in this formula so even in a situation where we've got a firstbaseman (0 points) vs. a shortstop the bump is "only" just slightly more than 10% and that's the most extreme situations. The top vote getters (Bonds) comes in around 2300 points, the lowest MVP (I'll have to check the spreadsheet when I return home) came in around 1300.
April 03, 2009 at 10:51PM View BBCode
Roosters, I tried to have the position bonuses set much lower however it pulled a lot of wrong answers. I'm planning on some point at running through a larger pile of data but if you look at the numbers, the positional bonues work almost all the time. The only two MVP votes since 1994 that were way off based on this formula were the Morneau MVP and Mo Vaughn was significantly behind Albert Belle. I couldn't quite figure out how to add in a "jackass" part to the formula to take into account guys like Belle or Ted Williams. The flip side to that is that even if I were able to figure out such a factor-it would also have to account for guys like Bonds and Dick Allen (and for Williams when he did win).The truth of the matter, historically speaking is that hitters at C, 2B and SS can put up numbers that are less than those who play 1B or OF. A guy playing one of those positions (see Rodriguez, Ivan or Larkin, Barry or Kent, Jeff) who puts up numbers that are in the ballpark but not all the way up to what the raw statistics would say for a 1B is generally going to win the MVP. The exception that I've seen is Justin Morneau who somehow won over Carlos Guillen, Derek Jeter and Joe Mauer.
Also note, there is a significant bump for players who play RF which is drastically different than what SD currently has. Thus what we're seeing is that a catcher vs. a RF only gains about 45 points based on defense.
Its also important to note that typically the MVP goes to a player with more than 1500 points in this formula so even in a situation where we've got a firstbaseman (0 points) vs. a shortstop the bump is "only" just slightly more than 10% and that's the most extreme situations. The top vote getters (Bonds) comes in around 2300 points, the lowest MVP (I'll have to check the spreadsheet when I return home) came in around 1300.
shutout1277
Why should a previous MVP winner get a -100 in the calculation? I'm not saying that he should get any sort of bonus for being a former MVP, but I don't think he should be punished for it either.
April 04, 2009 at 03:32AM View BBCode
well, it's about time! This had needed fixing for quite some time now! Nice work Bryan and Tyson.... just one question/problem I have.Why should a previous MVP winner get a -100 in the calculation? I'm not saying that he should get any sort of bonus for being a former MVP, but I don't think he should be punished for it either.
barterer2002
April 04, 2009 at 03:36AM View BBCode
The reason is that given two equal players in all other regards, a voter is more likely-historically speaking-even recently historical-to vote for the player who had never won the award before than to give it to someone who had previously won.
tworoosters
Curry had a higher BA, OBP, Slugging %, a way better stolen base %, hit more homers and played "up the middle", something doesn't seem right to me.
Mind you it was a strange year as nobody exceeded 1300 points but that seems out of whack to me.
April 04, 2009 at 05:17AM View BBCode
So based on last year's numbers in Beta in the AL [url=http://beta.simdynasty.com/player.jsp?id=19506]Bret Ryan[/url], previously 5th, is the MVP and the MVP under the previous system [url=http://beta.simdynasty.com/player.jsp?id=19092]Tom Curry[/url] drops to 5th.Curry had a higher BA, OBP, Slugging %, a way better stolen base %, hit more homers and played "up the middle", something doesn't seem right to me.
Mind you it was a strange year as nobody exceeded 1300 points but that seems out of whack to me.
ScooterPie
Obscenely minor thought: you could make it for leagues that have passed 108 games. That's 2/3 of the season, and every league speed except 15 gpd ends a day with exactly 108 games played. Might cut down on posts saying "My leag just playd it's 100th gm and the MVPs' are'nt showing!!!1!!!!1"
scooter
April 04, 2009 at 12:07PM View BBCode
Originally posted by AdminThat's really gonna bum some people out. But as you say, maybe it'll turn the revelation of the voting into an event.
I've set it up to run at 4:15AM for leagues that have passed 100 games.
Obscenely minor thought: you could make it for leagues that have passed 108 games. That's 2/3 of the season, and every league speed except 15 gpd ends a day with exactly 108 games played. Might cut down on posts saying "My leag just playd it's 100th gm and the MVPs' are'nt showing!!!1!!!!1"
scooter
barterer2002
In seasons where nobody tops 1400 points, such as the AL last season, the formula is less accurate in terms of predicting a winner. Usually there are going to be 4-5 guys within 200 points of the winner in such a season and a case could be made for any of them.
Also note-the formula does not include pitchers in the MVP race although that is a modification that I may change in the future
April 04, 2009 at 02:20PM View BBCode
Originally posted by tworoosters
So based on last year's numbers in Beta in the AL [url=http://beta.simdynasty.com/player.jsp?id=19506]Bret Ryan[/url], previously 5th, is the MVP and the MVP under the previous system [url=http://beta.simdynasty.com/player.jsp?id=19092]Tom Curry[/url] drops to 5th.
Curry had a higher BA, OBP, Slugging %, a way better stolen base %, hit more homers and played "up the middle", something doesn't seem right to me.
Mind you it was a strange year as nobody exceeded 1300 points but that seems out of whack to me.
In seasons where nobody tops 1400 points, such as the AL last season, the formula is less accurate in terms of predicting a winner. Usually there are going to be 4-5 guys within 200 points of the winner in such a season and a case could be made for any of them.
Also note-the formula does not include pitchers in the MVP race although that is a modification that I may change in the future
tworoosters
Why not after 81 games, 1/2 way through the year ?
April 04, 2009 at 02:39PM View BBCode
Originally posted by Admin
I've set it up to run at 4:15AM for leagues that have passed 100 games.
Why not after 81 games, 1/2 way through the year ?
Admin
I set it to 100 games because at that point most guys would be eligible for the position bonus. I might try to change the logic on this.
Tyson
April 04, 2009 at 04:57PM View BBCode
For the AL race last year, the CF bonus is no longer as high as it used to be relative to the SS bonus.I set it to 100 games because at that point most guys would be eligible for the position bonus. I might try to change the logic on this.
Tyson
tworoosters
It also shows the backup All Star 3B leading the MVP race while the starting 3B is not in the top 5 despite both being on "playoff teams" so something appears to be amiss.
April 05, 2009 at 04:28PM View BBCode
The current beta league All Star page, after 44 games of 2032, shows MVP totals that are less than All Star votes.It also shows the backup All Star 3B leading the MVP race while the starting 3B is not in the top 5 despite both being on "playoff teams" so something appears to be amiss.
Admin
Tyson
April 06, 2009 at 12:45AM View BBCode
tworoosters:Right now, beta has been using the old MVP formula for the games played thus far - so those numbers will stay until we get to 100 games.
IS the playoff bonus cumulative to the wins bonusYes.
Tyson
tworoosters
April 06, 2009 at 12:56AM View BBCode
Originally posted by Admin
tworoosters:
Right now, beta has been using the old MVP formula for the games played thus far - so those numbers will stay until we get to 100 games.
If so how is it possible for players MVP vote totals to be lower than their All Star vote totals ?
The old system was simply a process of taking All Star votes and adding positional. and or, standings bonus points.
Currently the Beta MVP leader has 61 MVP votes as opposed to 133 All Star votes, impossible under the old system.
The current MVP leader also is third in All Star voting at his position yet is ahead of the positional leader in the MVP race despite the fact they both play on "playoff" teams, this is also impossible under the old MVP system.
Admin
tyson
April 06, 2009 at 01:00AM View BBCode
I turned off the old votes. So they are frozen in time from the time I turned them off. The new ones will kick in when we get to 100 games.tyson
Admin
It will run after game number 60 for pay leagues, game number 120 for trial leagues.
The trial league win part of the formula is halved, so you'll get bonuses for 45 wins, 50 wins, etc.
I've changed the position part of the formula to check if the guy has played more than half of the number of team games at that position. So if your team has played 70 games, and he has played 35 games at SS, he will get that position bonus.
Tyson
April 13, 2009 at 07:35PM View BBCode
I have made a few logistical changes to this.It will run after game number 60 for pay leagues, game number 120 for trial leagues.
The trial league win part of the formula is halved, so you'll get bonuses for 45 wins, 50 wins, etc.
I've changed the position part of the formula to check if the guy has played more than half of the number of team games at that position. So if your team has played 70 games, and he has played 35 games at SS, he will get that position bonus.
Tyson
rbertoldie
June 05, 2009 at 05:56PM View BBCode
I don't agree that the previous winner should get -100...maybe -50
Admin
Tyson
June 05, 2009 at 06:01PM View BBCode
I don't agree that the previous winner should get -100...maybe -50MLB voters disagree with you :) If you can come up with a formula that more accurately predicts MLB MVP winners, I will consider changing to your formula.
Tyson
Pages: 1