Good Mets Years vs. Bad Mets Years

[G-Fafif]

Any thought given, when calculating all-time rankings, to weighting seasons based on overall Met success? I get that this is based on who was ranked highest in a given season, but it strikes me that being the No. 3 Met in a lousy year isn't spiritually equal to being the No. 3 Met in a great year (or the No. 6 Met in a great year, for that matter). This occurs to me after recently having ascended through the ranks from Lenny Dykstra (five Met seasons, all of them winning, one of them the most magnificent of Met seasons) to -- with Hubie Brooks between them -- Bobby Bonilla (five Met seasons, four of them losing, the other a great one he had no positive impact on).

[Guest John Cougar Lunchbucket]

I don't know the formula but it's already baked in.

[Frayed Knot]

Yes, they are already baked in.
The player's score he gets from finishing high or low on the ranks of his fellow Mets for a particular season is then multiplied by the number of team wins* that year.
So a top of the list player in a 90 win season is rewarded 1.5x as well as his counterpart in a 60 win year.





* Plus bonus points for any post-season advancement.

[Edgy MD]

The root of the formula is (R^2*W)/10

R is Ranking with a number one ranking equal 30 and a number 30 ranking equaling 1.

W is wins by the team (pro-rated to 162 for strike- or weather-shortened years).

There's also some adjustments made for playoff success.

So, while Bob Ojeda's 1986 was not the best season by any Met ever (probably not in the top ten), that season got him the most points of any player in a single season. But it works a lot better in the macro than in the micro.

[batmagadanleadoff]

I don't like that the best player always scores a 30 and the second best player always scores a 29. The ranking system wrongly assumes that the gap between any two players within a season ranked next to each other is always the same. The ranking system also wrongly assumes that, season to season, the performance of the best player (and second best player, and third best player, all down the line) is always the same. The best player on the 1980 squad shouldn't receive the same "30" that the best player on the 1986 squad received.

[Edgy MD]

We've had some leniancy in that system.

[Frayed Knot]

Specifically to the Dykstra/Bonilla thing;

- Dykstra's problem was that in his five seasons with the Mets two of them were partials and in the others he platooned much of the time. So, while he may have been on some great teams, he never wound up as being one of the top guys in any one of them ended up ranking as the 19th, 10th, 6th, 5th, & 19th best Mets on those squads.

- Bonilla on the other hand had much higher ranks - he was voted 3rd, 1st, 3rd & 4th place in 1992-95, and then 27th best for his 'contributions' for 1999 - even though they were all sub-80 win teams.

[Edgy MD]

Relative to the league, Bobby Bonilla was a two time All-Star with the Mets. Dykstra none.

On the other hand, bb-r.com says Bonilla had 8.7 WARs for the Mets. Dykstra had 16.4.

[Frayed Knot]

The best player on the 1980 squad shouldn't receive the same "30" that the best player on the 1986 squad received.

He might receive the same '30' but a very different score.

[batmagadanleadoff]

Frayed Knot wrote: The best player on the 1980 squad shouldn't receive the same "30" that the best player on the 1986 squad received. He might receive the same '30' but a very different score.

Yes, but the different "final" score is based entirely on the performance of their teammates. Before adjusting for team W-L record, the best '86er should have a higher score than the best '80er. The ranking system overemphasizes team performance at the expense of the player's performance

[Edgy MD]

make up your own system.

[Frayed Knot]

batmagadanleadoff wrote: Before adjusting for team W-L record, the best '86er should have a higher score than the best '80er. The ranking system overemphasizes team performance at the expense of the player's performance

Aren't those two sentences at odds with each other?
- In the first you say that the team outcome should trump what the individuals did - and do so twice, and apparently do so automatically regardless of the accomplishments in question.
- while in the next you claim that team performance is overemphasized at the expense of the player.

[batmagadanleadoff]

Before adjusting for team W-L record, the best '86er should have a higher score than the best '80er. The ranking system overemphasizes team performance at the expense of the player's performance Aren't those two sentences at odds with each other? - In the first you say that the team outcome should trump what the individuals did - and do so twice, and apparently do so automatically regardless of the accomplishments in question. - while in the next you claim that team performance is overemphasized at the expense of the player.

Lee Mazzilli is the CPF highest ranking 1980 Met and therefore, is awarded 30 points.
viewtopic.php?f=4&t=600

Dwight Gooden, the highest ranking 1985 Met, gets the same "30" that Mazzilli80 did, even though Doc85 turned in the best pitching season in franchise history and Maz80 wouldn't have been the 4th best player on the '85 squad. Doc's '85 was one of the most dominant pitching seasons in modern history while Maz80 wasn't even the best at his position.

Yet they get the same "30". Only after adjusting for the performance of their teammates does Doc85 surpass Maz80. Doc85 scores higher than Maz80, but only because Doc had better teammates.

[batmagadanleadoff]

Before adjusting for team W-L record, the best '86er should have a higher score than the best '80er. The ranking system overemphasizes team performance at the expense of the player's performance Aren't those two sentences at odds with each other? - In the first you say that the team outcome should trump what the individuals did - and do so twice, and apparently do so automatically regardless of the accomplishments in question. - while in the next you claim that team performance is overemphasized at the expense of the player. Lee Mazzilli is the CPF highest ranking 1980 Met and therefore, is awarded 30 points. viewtopic.php?f=4&t=600 Dwight Gooden, the highest ranking 1985 Met, gets the same "30" that Mazzilli80 did, even though Doc85 turned in the best pitching season in franchise history and Maz80 wouldn't have been the 4th best player on the '85 squad. Doc's '85 was one of the most dominant pitching seasons in modern history while Maz80 wasn't even the best at his position. Yet they get the same "30". Only after adjusting for the performance of their teammates does Doc85 surpass Maz80. Doc85 scores higher than Maz80, but only because Doc had better teammates.

I'm not against accounting for the team's performance, but the system, as is, overdoes it.

[batmagadanleadoff]

Frayed Knot wrote: batmagadanleadoff wrote: Before adjusting for team W-L record, the best '86er should have a higher score than the best '80er.... Aren't those two sentences at odds with each other? - In the first you say that the team outcome should trump what the individuals did ....

I wrote that the best '86er should have a higher score than the best '80er before adjusting for team record --- because the best '86er, whoever that may be, had a better season than the best '80er. They shouldn't both get the same "30"

(see also above two posts for the rest of my response)

[Edgy MD]

We're going in circles here here. As you point out, Mazzilli's ranking of 30 would not have happened had he had he been on the 1985 team.

Better teammmates typically generate more wins, but lower rankings. So Gooden's better score is not the product of better teammates, but better Gooden, standing out in a better field, and getting more points for it.

Gooden scores higher, not because he had better teammates, but because he towered above better teammates.

Gooden received 8,820 points. Mazzilli received 6,030. That's not a particularly accurate representation of their relative worth (Gooden is probably worth at least two Mazzes, not 1.46), but it's not that innacurate either, and we're talking about an extreme case.

[batmagadanleadoff]

Edgy DC wrote: We're going in circles here here. As you point out, Mazzilli's ranking of 30 would not have happened had he had he been on the 1985 team.

Then Maz80 gets a "30" because he had lesser teammates than Doc85. You can't have it both ways.

Edgy DC wrote: Better teammmates typically generate more wins, but lower rankings. So Gooden's better score is not the product of better teammates, but better Gooden, standing out in a better field, and getting more points for it.

Because he had better teammates. Why didn't the '85 Mets lose 85 games? He didn't need Straw in RF or Keith at first to lead the league in strikeouts.

Had Doc85 pitched the same season for the '64 Mets, he'd have the same stats other than W-L record. Yet the system would punish Doc64 because his first baseman would've been Krane instead of Keith.

[batmagadanleadoff]

Edgy DC wrote: We're going in circles here here. As you point out, Mazzilli's ranking of 30 would not have happened had he had he been on the 1985 team. Better teammmates typically generate more wins, but lower rankings. So Gooden's better score is not the product of better teammates, but better Gooden, standing out in a better field, and getting more points for it. Gooden scores higher, not because he had better teammates, but because he towered above better teammates. Gooden received 8,820 points. Mazzilli received 6,030. That's not a particularly accurate representation of their relative worth (Gooden is probably worth at least two Mazzes, not 1.46), but it's not that innacurate either, and we're talking about an extreme case.

I'm not against adjusting for team performance. But I think that your argument is either flawed, or contrived just for contrary's sake. The system scores the players based on individual performance. (player score) And then a final adjustment is made to each player score for team performance. If you think that Maz80 should have the same player score as Doc85, then that's what you think.

[Edgy MD]

batmagadanleadoff wrote: If you think that Maz80 should have the same player score as Doc85, then that's what you think.

That's not what I think. What are you talking about?

[batmagadanleadoff]

Edgy DC wrote: If you think that Maz80 should have the same player score as Doc85, then that's what you think. That's not what I think. What are you talking about?

Before adjusting for team performance, Doesn't Maz80 score a "30"? Doesn't Doc85 score a "30"?

Edgy DC wrote: The root of the formula is (R^2*W)/10 R is Ranking with a number one ranking equal 30 and a number 30 ranking equaling 1. W is wins by the team (pro-rated to 162 for strike- or weather-shortened years). There's also some adjustments made for playoff success. So, while Bob Ojeda's 1986 was not the best season by any Met ever (probably not in the top ten), that season got him the most points of any player in a single season. But it works a lot better in the macro than in the micro.

[Edgy MD]

The only question is whether you deliberately obtuse are not, but I really dislike when you force an argument on somebody that they are not making.

The number 30 is a factor, not a product.

[batmagadanleadoff]

batmagadanleadoff wrote: batmagadanleadoff wrote: If you think that Maz80 should have the same player score as Doc85, then that's what you think. That's not what I think. What are you talking about? Before adjusting for team performance, Doesn't Maz80 score a "30"? Doesn't Doc85 score a "30"? The root of the formula is (R^2*W)/10 R is Ranking with a number one ranking equal 30 and a number 30 ranking equaling 1. W is wins by the team (pro-rated to 162 for strike- or weather-shortened years). There's also some adjustments made for playoff success. So, while Bob Ojeda's 1986 was not the best season by any Met ever (probably not in the top ten), that season got him the most points of any player in a single season. But it works a lot better in the macro than in the micro.

There are only two variables in the rankings formula you posted. Doc85 and Maz80 have the same player rank (30), and are thus, distinguished only by the performance of their teammates.

[Edgy MD]

I'm done.

[batmagadanleadoff]

Edgy DC wrote: The number 30 is a factor, not a product.

I don't follow. Can you explain?

[Benjamin Grimm]

The 30 is just part of the formula. Their actual scores, the products of the formula, is 8,820 points for Gooden and 6,030 for Mazzilli. It's those numbers that affect the all-time rankings.

[batmagadanleadoff]

Benjamin Grimm wrote: The 30 is just part of the formula. Their actual scores, the products of the formula, is 8,820 points for Gooden and 6,030 for Mazzilli. It's those numbers that affect the all-time rankings.

So then this isn't the entire formula (other than playoff adjustments)?

EdgyDC wrote: (R^2*W)/10 R is Ranking with a number one ranking equal 30 and a number 30 ranking equaling 1. W is wins by the team (pro-rated to 162 for strike- or weather-shortened years). There's also some adjustments made for playoff success.

[Frayed Knot]

Edgy DC wrote: I'm done.

I'm lost.

[batmagadanleadoff]

I'm done. I'm lost.

You guys must be pulling my legs.

I stand by my arguments. More than ever.

Here’s the formula, per Edgy’s earlier post:

The root of the formula is (R^2*W)/10

The formula has two variables:
R = the player’s score for a particular season based on his ranking relative to his teammates. The highest ranking player receives a score of 30; the next highest ranking player receives a score of 29, etc. R cannot exceed 30 (R <=30).

W = the number of wins by the Mets for the particular season under consideration.

My interpretation of the formula is consistent with the following point totals for Maz80 and Doc85, also provided by Edgy:

Gooden received 8,820 points. Mazzilli received 6,030.

Thus:

Maz80 = (30^2*67)/10 = 6,030
Doc85 = (30^2*98)/10 = 8,820

So ….

1) the system is flawed because, by ranking players from 1 to 30, assumes that the gap between any two players ranked immediately next to each other (same season) is always the same, thus compromising the system’s ability to compare players same-season.

2) collapsing the player rankings portion of the formula (R variable) to <=30 compromises the system’s ability to perform cross-season comparisons.

In the Maz80 v Doc85 example,

Maz80 = (30^2*67)/10 = 6,030
Doc85 = (30^2*98)/10 = 8,820

Doc85 scores higher than Maz80, but only because Doc85 had better teammates. Both Doc85 and Maz80 received a score of 30 for variable R. Variable W is the only variable that distinguishes Doc85’s score from Maz80. The system doesn’t reflect that relative to each other, Doc’s ’85 was a season for the ages and Maz80 was, deservedly, not selected for any end of the year 1980 all star teams. The better player won’t always be surrounded by better teammates. This is especially important here because the Mets, historically, have had more bad teams than good. To me, it makes no sense to score the best player, every season with a "30" (variable R)

Placing a ceiling of 30 on variable R also compromises the system’s ability to gauge player performance, even if you''re not comparing one player to another. Gooden, in 1985, might’ve hypothetically been even better than he actually was. Suppose that Doc85 struck out an additional 50 batters over and beyond his actual total, and finished the season with an ERA a half a run lower than his actual ERA. The system wouldn’t recognize these hypothetical extra accomplishments even though the marginal value of the extra 50 K’s and -0.50 ERA might rank Gooden’s now make believe 1985 pitching season as the best all-time in the modern era. That's because the player rank portion of the formula (R) has a ceiling of 30. You might counter that Gooden’s additional success would yield the Mets additional wins, and that the W variable would increase Gooden’s ultimate score. But what if Doc’s hypothetical improvement was offset by Sid Fernandez, who hypothetically struck out 50 less batters and saw his ERA increase by half a run? Again, the best player won’t always be surrounded by better players, and the system shouldn’t need to rely on this assumption materializing in order to accurately gauge player performance.

The system should also be as accurate in the extreme as it is with the mundane. The best systems always are. Besides, the extreme is where most interests primarily lie. I’d be more interested in the spectacular (Doc85, Piazza2000) and the spectacularly inept (Flynnanyseason) than to see how the system might, for example, distinguish Staiger77 from Phillips75.

[G-Fafif]

The formula might not take into account that Dykstra ruled and Bonilla sucked. But as Janet Jackson mentioned when Bobby Bo was receiving that first All-Star selection, that's the way love goes.

[batmagadanleadoff]

I'm done. I'm lost. You guys must be pulling my legs.

I wrote the post referenced above to clarify what appeared to be some confusion:

If you think that Maz80 should have the same player score as Doc85, then that's what you think. That's not what I think. What are you talking about?

Before adjusting for team W-L record, the best '86er should have a higher score than the best '80er. The ranking system overemphasizes team performance at the expense of the player's performance Aren't those two sentences at odds with each other? - In the first you say that the team outcome should trump what the individuals did - and do so twice, and apparently do so automatically regardless of the accomplishments in question. - while in the next you claim that team performance is overemphasized at the expense of the player.

I was referring to variable R

The root of the formula is (R^2*W)/10

when I complained about Maz80 and Doc85 receiving the same "30". Having a disagreement is pointless when the opposite sides can't minimally understand the positions taken. And so I hope that my last post, if nothing else, at least clarified my position. My biggest gripe about the ranking system is that all best seasons collapse to a score of 30 (all second best seasons collapse to 29, etc). I think it's absurd to equate, as jujst one example, Maz80 and Doc85. Variable R is one of only two variables in the formula, and thus accounts for 50% of the formula, maybe even more because variable R is immediately squared. The system requires that players like Doc85 and Piazza2000 play for strong squads in order for it to generate the high rankings those player-seasons merit. A great player-season occurring within the context of a bad baseball team (Carlton72 is a classic example) gets the shaft, to a disproportionately higher degree than I think is merited.

You can disagree. But by now, I'll assume that you understand the basis of my arguments. Because if you disagree while simultaneously claiming confusion, then the entire dialogue is pointless.