Sabermetrics are a bunch of malarkey . . .

[1998 Yankees]

. . . at least in how they are calculated and used in OOTP, that is. I don't mean to offend any Bill James fans and I certainly do not intend to engage in any arguments that I am bound to lose here. This is just my own personal opinion and sort of a release of frustration after spending the last few hours trying to figure out just two ratings and their effect on the game: VORP and RC/27.

Actually I did not try to figure out VORP in terms of how it is calculated in OOTP; I don't think anyone can except for the designers, and they have never been too specific in their explanations. See the statistics definitions appendix in the manual for what I mean.

Anyway, with VORP, I started to look at it intensively when I noticed that two players who had spectacular seasons in recent years had been denied the MVP award. See below. In each case, the winner had higher VORP, which is usually the governing factor of who wins the award. OK, fine; VORP is meant to capture other performance measures than RBI's of course.

But look closer. Griffin's stats edge Eldridge in 2013 and blow away Bruno in 2015 but he lost in both years due to lower VORP. Motsuzuki in 2016, same story; similar numbers (except OBP, which I suspect is over-weighted in VORP for other reasons, but that's another thread) until you see 34 more RBI's than Reynolds. 34 more RBI's and only 2/3 of Reynolds VORP? How can that be? How can that difference in run production be offset by any other factors by such a margin?

OK, leave aside VORP. Perhaps the MVP winners were selected because they played on superior teams (which I know is not correct because MVP is nearly always highest VORP). In 2013 Griffin's team was 2nd in division but had a better record than the MVP winner's team which won their division. Maybe that division championship was it? Uh, no. In 2015 Griffin and the MVP winner were teammates and in 2016 Motsuzuki's 85-77 team finished fourth, ahead of the winner's fifth-place team which was 80-82.

Once in a blue moon, I notice that the highest VORP guy is passed over for the highest in RC/27. Aha, here is a number that I can calculate, according to the formula in the manual! I actually did so, laying it out in Excel (see below for an English translation of the formula). OK, leave aside the seemingly arbitrary selection of factors to include and how they are arranged in the formula; I assume there is a reason for it all. What gets me really skeptical is the .26 and .52 multipliers. Where do they come from? Still, I had calculated a sabermetric statisic all by myself!

Oh, wait. The kicker, the final straw that brought this thread about, is after successfully calculating RC and comparing it to the game (close, I assume differences are in rounding), I happened to notice that RC/27 in the game does not equal RC divided by 27. What the heck? This is the simplest part of the James chapter and verse on runs created: "The same information provided by runs created can be expressed as a rate stat, rather than a raw number of runs contributed. This is usually expressed as runs created per some number of outs, e.g. (27 of course being the number of outs per team in a standard 9-inning baseball game)." RC/27 = RC divided by 27.

But no! See below. Just out of curiosity, does anybody have an explanation for this? The manual explains it as simple math. Why is it not so in the game?

No matter. I have forsworn using sabermetric statistics in OOTP, as I really have no faith in them in general, much less how they are used in the game. I hereby return to my former stance on the use of such statistics as VORP during my gameplay, as evidenced by these historic posts of mine (don't miss my temper tantrum in the last one ):

http://www.ootpdevelopments.com/boar...58&postcount=5
http://www.ootpdevelopments.com/boar...2&postcount=27
http://www.ootpdevelopments.com/boar...7&postcount=42

So, I will be choosing the awards from now on, and selecting HoF members, based on my own judgment. The game has changed since those posts, allowing me to do so, and I am going to take full advantage of that from now on. Remembering certain "How do I increase the immersion factor of my game?" threads, I anticipate that manual awards and inductions are going to greatly aid in that aspect as well.

[darkcloud4579]

I always put 10000 in the VORP category. HoF should not be picked based on a stat like that and it's annoying when it happens automatically.

[Tony M]

1998Y...

it's not just you.

I have often looked at RC and RC/27 and thought that I was missing something very simple when dividing RC by 27 and not getting the figure the game gives.

I was in the throes of trying to create a slightly different csv export to test something out and I couldn't fathom out why I got a different RC/27 to what the game was giving.

[lynchjm24]

Originally Posted by redsoxford

1998Y...

it's not just you.

I have often looked at RC and RC/27 and thought that I was missing something very simple when dividing RC by 27 and not getting the figure the game gives.

I was in the throes of trying to create a slightly different csv export to test something out and I couldn't fathom out why I got a different RC/27 to what the game was giving.

It's runs created per 27 outs. Not Runs Created just divided by 27. Why would that even be a stat?

O = AB - H + CS + GDP + SH + SF
RC/27 = RC / O * 27

[lynchjm24]

Originally Posted by 1998 Yankees

. . . at least in how they are calculated and used in OOTP, that is. I don't mean to offend any Bill James fans and I certainly do not intend to engage in any arguments that I am bound to lose here. This is just my own personal opinion and sort of a release of frustration after spending the last few hours trying to figure out just two ratings and their effect on the game: VORP and RC/27.

Actually I did not try to figure out VORP in terms of how it is calculated in OOTP; I don't think anyone can except for the designers, and they have never been too specific in their explanations. See the statistics definitions appendix in the manual for what I mean.

Anyway, with VORP, I started to look at it intensively when I noticed that two players who had spectacular seasons in recent years had been denied the MVP award. See below. In each case, the winner had higher VORP, which is usually the governing factor of who wins the award. OK, fine; VORP is meant to capture other performance measures than RBI's of course.

But look closer. Griffin's stats edge Eldridge in 2013 and blow away Bruno in 2015 but he lost in both years due to lower VORP. Motsuzuki in 2016, same story; similar numbers (except OBP, which I suspect is over-weighted in VORP for other reasons, but that's another thread) until you see 34 more RBI's than Reynolds. 34 more RBI's and only 2/3 of Reynolds VORP? How can that be? How can that difference in run production be offset by any other factors by such a margin?

OK, leave aside VORP. Perhaps the MVP winners were selected because they played on superior teams (which I know is not correct because MVP is nearly always highest VORP). In 2013 Griffin's team was 2nd in division but had a better record than the MVP winner's team which won their division. Maybe that division championship was it? Uh, no. In 2015 Griffin and the MVP winner were teammates and in 2016 Motsuzuki's 85-77 team finished fourth, ahead of the winner's fifth-place team which was 80-82.

Once in a blue moon, I notice that the highest VORP guy is passed over for the highest in RC/27. Aha, here is a number that I can calculate, according to the formula in the manual! I actually did so, laying it out in Excel (see below for an English translation of the formula). OK, leave aside the seemingly arbitrary selection of factors to include and how they are arranged in the formula; I assume there is a reason for it all. What gets me really skeptical is the .26 and .52 multipliers. Where do they come from? Still, I had calculated a sabermetric statisic all by myself!

Oh, wait. The kicker, the final straw that brought this thread about, is after successfully calculating RC and comparing it to the game (close, I assume differences are in rounding), I happened to notice that RC/27 in the game does not equal RC divided by 27. What the heck? This is the simplest part of the James chapter and verse on runs created: "The same information provided by runs created can be expressed as a rate stat, rather than a raw number of runs contributed. This is usually expressed as runs created per some number of outs, e.g. (27 of course being the number of outs per team in a standard 9-inning baseball game)." RC/27 = RC divided by 27.

But no! See below. Just out of curiosity, does anybody have an explanation for this? The manual explains it as simple math. Why is it not so in the game?

No matter. I have forsworn using sabermetric statistics in OOTP, as I really have no faith in them in general, much less how they are used in the game. I hereby return to my former stance on the use of such statistics as VORP during my gameplay, as evidenced by these historic posts of mine (don't miss my temper tantrum in the last one ):

http://www.ootpdevelopments.com/boar...58&postcount=5
http://www.ootpdevelopments.com/boar...2&postcount=27
http://www.ootpdevelopments.com/boar...7&postcount=42

So, I will be choosing the awards from now on, and selecting HoF members, based on my own judgment. The game has changed since those posts, allowing me to do so, and I am going to take full advantage of that from now on. Remembering certain "How do I increase the immersion factor of my game?" threads, I anticipate that manual awards and inductions are going to greatly aid in that aspect as well.

I can't even believe I'm going to reply to this:

In 2013 Eldridge's season is vastly superior to Griffin assuming they play in similar offensive environments. His OBP and SLUG are notably higher. He drove in 15 fewer runs? Is that what you are talking about? Unless this post is supposed to be funny and I'm falling for it?

In 2015 it's the same issue. VORP is going to reward the high OBP/high slug. Bruno is clearly a better hitter the Griffin - he doesn't drive in as many runs because Griffin :
A. Hits in a different spot in the lineup
B. Has better teammates hitting in front of him
C. Over a short sample had an amazing season with runners in scoring position
D. Some combination of the A/B/C

The factors come from regression analysis done be extremely smart people. They aren't arbitrary and you are making a bit of a fool of yourself by pretending like they aren't accurate.

VORP has nothing to do with RBI totals. VORP is based on the components of what scores runs, not the actual runs that are scored. VORP will give you credit for walking with runners on second and third, and will penalize you when you force in a run on a 6-4 fielder's choice with the bases loaded.

OOTP seems to be trying to give the MVP to the actual MVP. You are right, in real life Griffin would win in 2015 - he might not in 2013, because the writers reward high RBI totals and have for a long time.

If you like high RBI totals you don't want to know what VORP thinks of low OBP guys like Joe Carter. He had some 100 RBI seasons that were probably just above zero because of his terrible OBPs.

Your argument is a very simple and legitimate one, but you are making it in a silly fashion. In real life the people who vote for MVP and Cy Young do so with simple counting stats. To model real life OOTP should do the same thing for it to be 'realistic'. Attacking the stats because you don't understand them is ridiculous, and someone with 4k posts on a board about a video game in 2 years probably has the time to spend understanding some of the more simple ones.

[Jestre]

Originally Posted by lynchjm24

I can't even believe I'm going to reply to this:

In 2013 Eldridge's season is vastly superior to Griffin assuming they play in similar offensive environments. His OBP and SLUG are notably higher. He drove in 15 fewer runs? Is that what you are talking about? Unless this post is supposed to be funny and I'm falling for it?

In 2015 it's the same issue. VORP is going to reward the high OBP/high slug. Bruno is clearly a better hitter the Griffin - he doesn't drive in as many runs because Griffin :
A. Hits in a different spot in the lineup
B. Has better teammates hitting in front of him
C. Over a short sample had an amazing season with runners in scoring position
D. Some combination of the A/B/C

The factors come from regression analysis done be extremely smart people. They aren't arbitrary and you are making a bit of a fool of yourself by pretending like they aren't accurate.

VORP has nothing to do with RBI totals. VORP is based on the components of what scores runs, not the actual runs that are scored. VORP will give you credit for walking with runners on second and third, and will penalize you when you force in a run on a 6-4 fielder's choice with the bases loaded.

OOTP seems to be trying to give the MVP to the actual MVP. You are right, in real life Griffin would win in 2015 - he might not in 2013, because the writers reward high RBI totals and have for a long time.

If you like high RBI totals you don't want to know what VORP thinks of low OBP guys like Joe Carter. He had some 100 RBI seasons that were probably just above zero because of his terrible OBPs.

Your argument is a very simple and legitimate one, but you are making it in a silly fashion. In real life the people who vote for MVP and Cy Young do so with simple counting stats. To model real life OOTP should do the same thing for it to be 'realistic'. Attacking the stats because you don't understand them is ridiculous, and someone with 4k posts on a board about a video game in 2 years probably has the time to spend understanding some of the more simple ones.

Is a hit or walk that does not lead to a run of any value????VORP and RC think so, they give value to hits and walks that do not effect real runs. VORP and RC do not measure actual runs but hypothetical runs. A player who has a higher OBP and Slg % but fewer runs and rbi's is more valuable??? MVP voting should be applied to real runs created not hypothetical mathematical runs created. In 2013 Griffin deserved the MVP because he statistically produced more runs and rbi's than Eldridge and therefore was of more value to his team.

[bp_]

Originally Posted by lynchjm24

blah blah blah.

wow, you're a real jackass

[1998 Yankees]

Originally Posted by lynchjm24

It's runs created per 27 outs. Not Runs Created just divided by 27. Why would that even be a stat?

O = AB - H + CS + GDP + SH + SF
RC/27 = RC / O * 27

Sources:

[1998 Yankees]

Originally Posted by lynchjm24

In real life the people who vote for MVP and Cy Young do so with simple counting stats. To model real life OOTP should do the same thing for it to be 'realistic'.

This is a very perceptive comment. This is also why sabermetrics bother me, aside from not being able to put my finger on exactly how things are calculated and done by the game. This is another reason why I will follow the course that I have chosen from now on.

[lynchjm24]

Originally Posted by 1998 Yankees

Sources:

Looks as though the documentation is wrong. RC/27 is runs created per 27 outs. Not runs created divided by 27. That would be pointless.

[Left-handed Badger]

Well, I am one who completely hates VORP and RC. Completely useless stats IMO. I can look at other stats to tell me what I need to know.

Still I was a bit confounded by one MVP season. (I look later to see if I can find I made a thread on it.

(But the winner and the guy i thought should win were pretty similar in most offensive categories.) (HR-RBIS-AVG-OBP-SLG-OPS etc) The main difference: was the winner scored 103 Runs and the non-winner scored a whopping 154 runs (the guy batting behind probably should have got recognition for MVP in his own right, as he did have as many RBIs as previous 2, but his other stats fell a little short overall). The only other difference was the winner won a GG in RF, and the non-winner played 3B like a DH.

[lynchjm24]

Originally Posted by Jestre

Is a hit or walk that does not lead to a run of any value????VORP and RC think so, they give value to hits and walks that do not effect real runs. VORP and RC do not measure actual runs but hypothetical runs. A player who has a higher OBP and Slg % but fewer runs and rbi's is more valuable??? MVP voting should be applied to real runs created not hypothetical mathematical runs created. In 2013 Griffin deserved the MVP because he statistically produced more runs and rbi's than Eldridge and therefore was of more value to his team.

Um... ok. I knew I shouldn't have ever responded.

You have to be able to understand that Runs and RBI totals are a function of your teammates.

Let's just agree to disagree because if by 2008 you don't understand that much it's probably too late.

[Left-handed Badger]

I'd have given the 2013,2015,2016 MVPs to Griffin and Mostuzaki probably.


Course to rationalize, maybe it is like Ted Williams in 1950. Maybe the voters hate Griffin.

[Jestre]

Originally Posted by lynchjm24

Um... ok. I knew I shouldn't have ever responded.

You have to be able to understand that Runs and RBI totals are a function of your teammates.

Let's just agree to disagree because if by 2008 you don't understand that much it's probably too late.

Your right, cuz you think someone who went 4 for 5 with 2 doubles and no runs and no rbi's is more valuable to his team than someone who went 1 for 5 with a grand slam.

[Russ]

Originally Posted by Jestre

Your right, cuz you think someone who went 4 for 5 with 2 doubles and no runs and no rbi's is more valuable to his team than someone who went 1 for 5 with a grand slam.

I'm in no way trying to argue with you here. The point of a lot of these stats is that when he came up to bat, it was out of his hands whether the bases were loaded or if there was no one on base. The point is that the batter who comes up with the bases empty and hits a home run is as good as hitter (not as valuable, not as great a run producer) as the guy who through only the skills of his teammates or good fortune to have the bases loaded, hit a grand slam.

[Jestre]

Originally Posted by Russ

I'm in no way trying to argue with you here. The point of a lot of these stats is that when he came up to bat, it was out of his hands whether the bases were loaded or if there was no one on base. The point is that the batter who comes up with the bases empty and hits a home run is as good as hitter (not as valuable, not as great a run producer) as the guy who through only the skills of his teammates or good fortune to have the bases loaded, hit a grand slam.

Then VORP and RC should include production as a ratio. If those same two guys each had an opportunity with the bases loaded and the bases empty and the one guy hit a grandslam and struck out with the bases empty and the other guy hit a homer with the bases empty and struck out with the bases loaded then who is more valuable???? By VORP and RC they are equal.

[wlight1]

Originally Posted by Jestre

Your right, cuz you think someone who went 4 for 5 with 2 doubles and no runs and no rbi's is more valuable to his team than someone who went 1 for 5 with a grand slam.

Let's take an easier example. If I'm building a team and choosing players one by one to put on that team, it is obvious to me (and to most people who study baseball at a high level) that I'd rather have a player who goes 3 for 5 with 3 solo home runs than a player who goes 1 for 5 with one grand slam. You prefer the second player because he "produced more runs". I prefer the first player because he's a better hitter. So, yes, agree to disagree.

[gmo]

Originally Posted by Jestre

Your right, cuz you think someone who went 4 for 5 with 2 doubles and no runs and no rbi's is more valuable to his team than someone who went 1 for 5 with a grand slam.

The point of the fancier new stats is to give you a perspective with all else being equal and only considering what the particular player does.

Given the choice before the game (not knowing how anything else at all was going to turn out) of a guy in 5 PA getting either 2 singles and 2 doubles or getting a HR, which one would you choose?

In the above scenario, grand slam man is surely more valuable. If you knew that instead it was a choice between the 4 hits all driving in at least one run versus a solo HR, you would surely pick the multiple hit game.

The point is that you do not know which scenario will happen, so you want some way to pick out the one more likely to produce the most positive results.

[Jestre]

Originally Posted by wlight1

Let's take an easier example. If I'm building a team and choosing players one by one to put on that team, it is obvious to me (and to most people who study baseball at a high level) that I'd rather have a player who goes 3 for 5 with 3 solo home runs than a player who goes 1 for 5 with one grand slam. You prefer the second player because he "produced more runs". I prefer the first player because he's a better hitter. So, yes, agree to disagree.

No I don't, we are talking about MVP stats produced. We are not talking about future production or potential production we are talking about a given universe of stats and comparing them. Your example doesnt work, your guy is .660-3-3 and the other is .200-1-4....who is more valuable...duh. My example was .200-1-1 and .200-1-4 duh!!! According to VORP and RC my guys are equal even though one has 4 times the rbis. Your example is apples and oranges.

[Jestre]

Originally Posted by gmo

The point of the fancier new stats is to give you a perspective with all else being equal and only considering what the particular player does.

Given the choice before the game (not knowing how anything else at all was going to turn out) of a guy in 5 PA getting either 2 singles and 2 doubles or getting a HR, which one would you choose?

In the above scenario, grand slam man is surely more valuable. If you knew that instead it was a choice between the 4 hits all driving in at least one run versus a solo HR, you would surely pick the multiple hit game.

The point is that you do not know which scenario will happen, so you want some way to pick out the one more likely to produce the most positive results.

But this thread is a comparison of stats produced, not potential stats. Which is why the wrong guy won the MVP. VORP and RC do not measure actual production, just potential production. I will take real production over theoretical production.