For those who think Trout over miggy

[jalopy]

Not saying Defense is not important, but assuming that if you are in the MLB you must be at least decent defensively(except for DH), i dont see defense to be as big of a factor... everyone should make the easy plays defensively!!!

We don't see eye to eye. Ask any Tiger pitcher (especially Fister and Porcello) how much defense matters. If Porcello was on a good fielding team his ERA (yes, earned runs) would likely be a full point lower. Yes, he will take the added run support that the offense gives him but come contract time, it matters.

I'm not proposing that defense make up 50% of the metrics because the pitcher factors in a disproportionate amount on that side of the ball but to say that 80-85% of baseball is offense is silly.

Do you create your own formulas?? or do you just look at statistics that are already there and make your analysis there?? I value offense more than defense, so sue me!!!

I haven't to this point. I rely greatly on my eyeballs and my experience. I lived in SoCal and saw Trout play live about 5-6 times and Miggy for 5 games in the regular season and two games against Oakland in the playoffs. I probably saw the Tigers on TV another 20 games or so. I found myself apologizing for Miggy's play on almost a daily basis. I give him credit for playing out of position for the benefit of the team but he must be analyzed for for the position he attempts to play. He is horrible.

A little defensive with the "sue me" comment. I think I mentioned a few times that your assumptions would be the key. You left that up to the reader to infer.

 

[jalopy]

I don't like OPS all that much. It's alright, but there are a couple problems, and you just listed one of them.

The idea behind OPS is to include a "not making outs" component with OBP and a "power" component with slugging, but it assumes they're worth the same amount, and it assumes a doubles is twice as valuable as a single, etc., which simply isn't the case.

The nice thing about it is that it often does track well with the best hitters, and that it's simple and obvious to understand what it's telling you.

I prefer to look at wOBA, though. It's like on on-base percentage which also factors in extra-base hits, eliminating the double-counting, but it also uses season-by-season offensive environments to determine the coefficients with which walks, singles, doubles, triples, and home runs are weighted, ultimately giving a pretty darned telling number.

I believe that the wOBA no longer uses defensive or baserunning metrics but it is great for a pure production number for hitters. It can also be easily converted to offensive runs above average. I do have a question as to why the league wOPA varies so greatly from year to year. Does the formula change or is it just attributed to variance?

 

[StanMarsh51]

Now did you take into consideration that NL catchers throw out a better % of base runners than AL catchers?

We're talking 28% to 26% here...that doesn't make up for a 10% gap in stolen base %, especially when it's not as if Molina's responsible for most of his CS

 

[MilkSpiller22]

We don't see eye to eye. Ask any Tiger pitcher (especially Fister and Porcello) how much defense matters. If Porcello was on a good fielding team his ERA (yes, earned runs) would likely be a full point lower. Yes, he will take the added run support that the offense gives him but come contract time, it matters.

I'm not proposing that defense make up 50% of the metrics because the pitcher factors in a disproportionate amount on that side of the ball but to say that 80-85% of baseball is offense is silly.



I haven't to this point. I rely greatly on my eyeballs and my experience. I lived in SoCal and saw Trout play live about 5-6 times and Miggy for 5 games in the regular season and two games against Oakland in the playoffs. I probably saw the Tigers on TV another 20 games or so. I found myself apologizing for Miggy's play on almost a daily basis. I give him credit for playing out of position for the benefit of the team but he must be analyzed for for the position he attempts to play. He is horrible.

A little defensive with the "sue me" comment. I think I mentioned a few times that your assumptions would be the key. You left that up to the reader to infer.

I will probably add more value to defense next year when i do my analysis... To be honest, i have changed the formula slightly everyday since the baseball season ended(more when we got closer to the awards), have even changed it a little since discussing my formula... But of course now it is irrelevant since the awards were given so i am not accomplishing much with it... But i dont see myself changing the categories(only the percentages) unless someone convinces me that i can improve it... I like the categories i have in place...

 

[Darkstone42]

[/B]I believe that the wOBA no longer uses defensive or baserunning metrics but it is great for a pure production number for hitters. It can also be easily converted to offensive runs above average. I do have a question as to why the league wOPA varies so greatly from year to year. Does the formula change or is it just attributed to variance?

wOBA is a pure hitting stat. I'm not sure how much the weights vary year to year, but I can't imagine it's a huge margin. Since there are more pitcher-friendly parks, and since pitchers are throwing harder, and since teams are starting to buy into defensive shifts a little more, I think hitting has been pretty down lately, and that could have some impact on the coefficients.

 

[MilkSpiller22]

i dont neccessarily believe double counting is a bad thing...

 

[Darkstone42]

I will probably add more value to defense next year when i do my analysis... To be honest, i have changed the formula slightly everyday since the baseball season ended(more when we got closer to the awards), have even changed it a little since discussing my formula... But of course now it is irrelevant since the awards were given so i am not accomplishing much with it... But i dont see myself changing the categories(only the percentages) unless someone convinces me that i can improve it... I like the categories i have in place...

I suggest you use something wOBA-based for what you call the "theoretical" portion, and then cut down the RBI/R portion's ratio or use clutch, leverage, or RISP stats for that part, or both, since I think RBIs and runs are results which are generated by how well a player hits overall. This way, you can reward players for taking advantage of opportunities without driving half of your calculation with context-related parameters.

Amplify defense and baserunning a bit relative to hitting (I think 55-60% hitting would be a little more reasonable, but there are probably numbers out there which could give you a good number), and I think you have a much better system.

 

[jalopy]

I will probably add more value to defense next year when i do my analysis... To be honest, i have changed the formula slightly everyday since the baseball season ended(more when we got closer to the awards), have even changed it a little since discussing my formula... But of course now it is irrelevant since the awards were given so i am not accomplishing much with it... But i dont see myself changing the categories(only the percentages) unless someone convinces me that i can improve it... I like the categories i have in place...

I think now would be the time to assess exactly what you are trying to achieve. Is the model supposed to be a predictor for the MVP or are you trying to assess who is the best all-around player in the game using metrics. I think your model is a worthwhile effort but, as it stands, it is useful only to you (which may be your intent).

Miggy is undoubtedly the best hitter in the game and , if healthy, will win your award almost every year until his bat speed starts to decline. I place more value on being an all-around player than most because that is the way I tried to approach the game. As I said, personal preference.

 

[MilkSpiller22]

I suggest you use something wOBA-based for what you call the "theoretical" portion, and then cut down the RBI/R portion's ratio or use clutch, leverage, or RISP stats for that part, or both, since I think RBIs and runs are results which are generated by how well a player hits overall. This way, you can reward players for taking advantage of opportunities without driving half of your calculation with context-related parameters.

Amplify defense and baserunning a bit relative to hitting (I think 55-60% hitting would be a little more reasonable, but there are probably numbers out there which could give you a good number), and I think you have a much better system.

I guess with that statement we can agree to disagree... i dont mind double counting- i see the benefits of it sometimes... as i said i believe in keeping it simple inside the complications... I actually think my (R+RBI-HR) formulas are the strength to my analysis... I just believe what a player has actually produced is very important... I dont like using RISP stats, because that just opens up a whole nother case of worms- the did the game matter, did the runs matter... Would rather just keep it simple there... I dont know since it is built into my formulas, but i would think that Baserunning is already 10% of my analysis, i dont believe it should be any bit higher...

I do believe RC27 is the best theoretical Production statistic(even though i have no clue the stats you mentioned)... I just like the fact that when i was originally creating my formulas i was making the RC27 without realizing- then when i found out i updated my formula where i thought the RC27 was better...
I do believe that a theoretical and actual production stat is all you need offensively- and i believe that it is very telling...

 

[MilkSpiller22]

I have looked at wOBA before, i like what it tries to do, but i dont like statistical formulas that you look at and you dont understand why the numbers were chosen... the coeffecients are something that i believe should always be simple...

 

[Darkstone42]

I have looked at wOBA before, i like what it tries to do, but i dont like statistical formulas that you look at and you dont understand why the numbers were chosen... the coeffecients are something that i believe should always be simple...

The coefficients are there to mostly just say, for example, singles are a little more valuable than walks, doubles a little more valuable than singles, but not twice as valuable, etc. Since I'm of the mind that not making an out is more important than power, I agree with that assessment.

 

[jalopy]

I have looked at wOBA before, i like what it tries to do, but i dont like statistical formulas that you look at and you dont understand why the numbers were chosen... the coeffecients are something that i believe should always be simple...

.

Well, they shouldn't be needlessly complicate but they need to get the desired result. I think double counting complicates this needlessly.

Anyway, thank you for sharing and I think it was a good exercise. It certainly shows where our differences of opinion lie. I should have plenty of computer time this week so maybe I will give a shot at a model that incorporates my thinking.

 

[MilkSpiller22]

.

Well, they shouldn't be needlessly complicate but they need to get the desired result. I think double counting complicates this needlessly.

Anyway, thank you for sharing and I think it was a good exercise. It certainly shows where our differences of opinion lie. I should have plenty of computer time this week so maybe I will give a shot at a model that incorporates my thinking.

Even if you do, i guarantee you will be constantly editing it... Also you will probably see that personal preferences is a big factor when creating your own... Thats why even though under my calculation Votto should have won the NL MVP, but i would never complain about McCutchen winning, because there needs to be room for error...

 

[MilkSpiller22]

The coefficients are there to mostly just say, for example, singles are a little more valuable than walks, doubles a little more valuable than singles, but not twice as valuable, etc. Since I'm of the mind that not making an out is more important than power, I agree with that assessment.

I like patterns that i can see, so because of that i just dont trust it...

 

[Darkstone42]

I like patterns that i can see, so because of that i just dont trust it...

So what is your justification for weighting singles and walks the same (a single can score a runner from second, for example, and a walk cannot)? For weighting doubles 2.5 times singles? Just curious. You like patterns you can see, and I want to know what pattern you saw which lead to those coefficients.

 

[MilkSpiller22]

So what is your justification for weighting singles and walks the same (a single can score a runner from second, for example, and a walk cannot)? For weighting doubles 2.5 times singles? Just curious. You like patterns you can see, and I want to know what pattern you saw which lead to those coefficients.

Either u read it wrong or I wrote it wrong singles were worth 1.75.

So starting with walks as 1.00 the patern is .75

 

[jalopy]

Even if you do, i guarantee you will be constantly editing it... Also you will probably see that personal preferences is a big factor when creating your own... Thats why even though under my calculation Votto should have won the NL MVP, but i would never complain about McCutchen winning, because there needs to be room for error...

I doubt it. It might need some editing after a trial run for accuracy but if the parameters are set correctly from the onset, it should be set in stone. If Miggy comes out ahead, he comes out ahead.

Votto comes out ahead because your model overweights hitters that have traditional power production (RBI's, HR's)

 

[MilkSpiller22]

I doubt it. It might need some editing after a trial run for accuracy but if the parameters are set correctly from the onset, it should be set in stone. If Miggy comes out ahead, he comes out ahead.

Votto comes out ahead because your model overweights hitters that have traditional power production (RBI's, HR's)

Umm... Votto had a bad R+RBI- hr votto was so high because his obp was significantly better than the rest of the NL guys.

 

[jalopy]

Umm... Votto had a bad R+RBI- hr votto was so high because his obp was significantly better than the rest of the NL guys.

I'm curioius where Hanley Ramirez finished in your model. To my biased eyes, he was the best player on the best team in the NL. How did his limited PA's affect his score (if you extrapolate his totals on 600 PA's)?

 

[MilkSpiller22]

I did not even include Hanram in my calculations since he was not qualified... so it may take me a couple minutes to get his numbers in...