Scoring Categories

[Colby]

Why isn't the K/BB ratio weighted by innings pitched?  That would be the proper way to use this statistic... For instance the Dodgers team K/BB ratio would be much more heavily weighted by Billingsley's ratio over Hong-Chih Kuo.  This would be a more accurate depiction of quality of team pitching.  Why include RA when we have FIP?  FIP is a more accurate statistic measuring essentially the same thing.

It is the constant battle between cumulative and ratio statistics (which I argue we need 50/50 of).  Isn't FIP a ratio statistic?

[Canada8999]

Why isn't the K/BB ratio weighted by innings pitched?  That would be the proper way to use this statistic... For instance the Dodgers team K/BB ratio would be much more heavily weighted by Billingsley's ratio over Hong-Chih Kuo.  This would be a more accurate depiction of quality of team pitching.  Why include RA when we have FIP?  FIP is a more accurate statistic measuring essentially the same thing.

It will inherently be more heavily weighted by a player that has more total K's and BB's (which are going to correlate relatively close to IP), but as Colby mentions the idea is to offer up an additional positive counting stat.

[mjmezzetti]

It will inherently overvalue the counting statistics that are variables in multiple categories.
Any time you use ratios, the most balanced team will win while the "best" team may not.
For instance, if a team's ops is .400 points better but their fip is .02 worse that's an even matchup.
Why not establish point values for the statistics that most closely are cleanly affiliated with pitchers'/hitters'?  We're never going to get this perfect but we'll get closest with a points system.
If we must use a ratio system our best bets are probably Runs Created, FIP and RF only but then we have two big issues... One, team performance is roughly, 48% hitting, 36% pitching and 17% fielding, they'd each be weighted 33%.  Second, and this one is big, this does not account for the performance differential in each category (you either win the category or you don't).
Perhaps we could have a conference call/ brainstorming?  Does anyone agree with what I'm stating here.  I think it's important we draft a scoring system that most accurately reflects team performance.  It's vital to making this league work.

[Colby]

M.J.,

Your 48% hitting, 36% pitching and 17% fielding recommendation has been an integral part of drafting the 2010 stats.  I'll give my reason for the ratio stats and then we can move on from there.  We try to simulate the real as much as we can, but we depend on the real transactions in the real game itself.  Over time, most likely, larger market teams will have more starters where small market teams scramble to pick up the extra AB and IP needed to get their countable stats up.  The ratio stats are used almost as a pro-ration factor to show how the players on a team would do if they had the average number of AB or IP.  Now, this brings up a good question for Fantrax.  I will ask them if they would consider prorating the stats per AB/IP (or another user inputted field) on a matchup (usually weekly) or seasonal basis.

Colby

[mjmezzetti]

Only way this league will work is a points system based on the following formulas...

Hitting 48%   
Runs Created ((H+BB-CS+HBP-GIDP)*((S+2D+3D+4D)*(.26(BB-IBB+HBP))+(.52*(SH+SF+SB))/(AB+BB+HBP+SH+SF)

Pitching 36%   
H.Times FIP ((13HR+3(BB+HBP)-2K)/IP)+3.20

Fielding 17%   
RF (Putouts+Assists)/Innings Played at a Position

We have to pull out the ratio portion and establish properly weighting point values for relevant statistics. Otherwise, this won't work.  Anyone else follow what I'm saying here?

1. Fantrax doesn't properly weight pitchers and hitters ratios based on each sample size.  It simply averages each hitters' and pitchers' ratio (regardless of at bats and innings pitched) causing a huge margin of error.

2. I can simply plug a minor league player into a weak position to skew my numbers (simply adding one counting category doesn't remedy this problem)

3. The margin of difference from one ratio statistic to another is simply thrown out, this is a huge discrepancy.

Without coming up with a solid system that'll properly evaluate each team it's difficult to imagine this being worthwhile, time-wise.  I love the idea but it's all a waste of time if the scoring doesn't reflect team performance.

[Colby]

Only way this league will work is a points system based on the following formulas...

Hitting 48%   
Runs Created ((H+BB-CS+HBP-GIDP)*((S+2D+3D+4D)*(.26(BB-IBB+HBP))+(.52*(SH+SF+SB))/(AB+BB+HBP+SH+SF)

Pitching 36%   
H.Times FIP ((13HR+3(BB+HBP)-2K)/IP)+3.20

Fielding 17%   
RF (Putouts+Assists)/Innings Played at a Position

We have to pull out the ratio portion and establish properly weighting point values for relevant statistics. Otherwise, this won't work.  Anyone else follow what I'm saying here?

1. Fantrax doesn't properly weight pitchers and hitters ratios based on each sample size.  It simply averages each hitters' and pitchers' ratio (regardless of at bats and innings pitched) causing a huge margin of error.

2. I can simply plug a minor league player into a weak position to skew my numbers (simply adding one counting category doesn't remedy this problem)

3. The margin of difference from one ratio statistic to another is simply thrown out, this is a huge discrepancy.

Without coming up with a solid system that'll properly evaluate each team it's difficult to imagine this being worthwhile, time-wise.  I love the idea but it's all a waste of time if the scoring doesn't reflect team performance.

I like your proposal with my pro-ration scheme that I mentioned in the last post.

[Canada8999]

Only way this league will work is a points system based on the following formulas...

There are countless numbers of successful and entertaining leagues out there, all using their own scoring systems.  I don't think there is any one correct way to do it.  Some people prefer Roto, some H2H.  Some prefer points based leagues, some prefer breaking it into categories.  As long as all owners know the system, they adjust their strategies accordingly.

Hitting 48%   
Runs Created ((H+BB-CS+HBP-GIDP)*((S+2D+3D+4D)*(.26(BB-IBB+HBP))+(.52*(SH+SF+SB))/(AB+BB+HBP+SH+SF)

Pitching 36%   
H.Times FIP ((13HR+3(BB+HBP)-2K)/IP)+3.20

Fielding 17%   
RF (Putouts+Assists)/Innings Played at a Position

These are good categories, but having a number of categories (more than 3) adds to the strategic value of the game... you can try to be balanced or target specific categories you want to win everytime.

1. Fantrax doesn't properly weight pitchers and hitters ratios based on each sample size.  It simply averages each hitters' and pitchers' ratio (regardless of at bats and innings pitched) causing a huge margin of error.

Ratios should be calculated based on team totals, no weighting is necessary.  If Fantrax is incorrectly calculating team ratios by simply averaging the individual player ratios then that's a bug and should be corrected, although I'm not sure that is the case.

2. I can simply plug a minor league player into a weak position to skew my numbers (simply adding one counting category doesn't remedy this problem)

I agree that we need a method to reward/enforce realistic levels of playing time... this is my primary concern with whatever system we select.

3. The margin of difference from one ratio statistic to another is simply thrown out, this is a huge discrepancy.

This is realistic.  In MLB a 10-0 win counts the same in the standings as a 1-0 win.

[Canada8999]

We have to pull out the ratio portion and establish properly weighting point values for relevant statistics. Otherwise, this won't work.  Anyone else follow what I'm saying here?

Ratios are important to reward quality over quantity... we need to find a balance between the two.  A scoring system that leads to teams chasing playing time and streaming in as many starters as possible would lead the league away from reality.

[mjmezzetti]

We play weekly lineup changes and we have a fixed number of starters.  Also it's a 30 team league.  Streaming players isn't an option.  I think a points league remedies all the issues I outlined.

[mjmezzetti]

"3. The margin of difference from one ratio statistic to another is simply thrown out, this is a huge discrepancy.

This is realistic.  In MLB a 10-0 win counts the same in the standings as a 1-0 win."

I meant this as an argument against roto where you might hold a huge lead in half categories and a slight loss in the other half or slight majority.  Your team would have performed better but would tie or lose.  In a points league you have a value attributed to each statistic given a more accurate total team score.

I'm sorry I'm harping on this, if you guys disagree let me know.  I'm just trying to explain my view, given I'm probably doing a shitty job, so we can agree on the best way.  I just want the league to be as accurate as possible, however I understand if you guys see things differently.  I've been in both roto, head to head and points leagues... My strong preference is head to head, points leagues.   I've been in a couple very competitive league for 5-6 years now and dropped the others.