Scoring Categories

[Colby]

Yes, roughly.  Values should be assigned to each so that pitching is roughly, on average 36-37% of a teams performance.  Should I try to draft point values for discussion?  Should I include quality starts and ground balls?  Are there any other statistics, not mentioned in the formulas I listed that should be added?

Did you say the same thing over again?

Anyways, I think the rest of the R.C. trusts your guidance with the scoring categories.  I like what you have already proposed if it can be done in a points-based fashion rather than category based.  If we can get that going along with my requested enhancement for prorating stats (given they are above the minimum), then I think we will be set.  I can also set matchups on a basis that is not weekly such as X number of days.  A 3-day matchup against a team is a good representation of a series.  Of course, there is no 3-0, 2-1, 1-2, 0-3 result, just 1-0 and 0-1, so we will probably end up with 45-48 games/matches for the regular season.

[mjmezzetti]

I received an error message when I sent it so I sent it again... The three day series is an interesting idea.  I would probably favor a standard week, however, as sample size would be small and there's the potential one team might have a number of players with off-days, rain-outs, etc?  That's my opinion.  Also, weekly would coincide with lineup changes, if that matters.
I'll try to draft a points system today and post it for discussion.

[Colby]

I received an error message when I sent it so I sent it again... The three day series is an interesting idea.  I would probably favor a standard week, however, as sample size would be small and there's the potential one team might have a number of players with off-days, rain-outs, etc?  That's my opinion.  Also, weekly would coincide with lineup changes, if that matters.
I'll try to draft a points system today and post it for discussion.

We can do weekly... probably 23 weeks with:

15 weeks - Divisional Opponent - This would give at least 3 matchups with each one of your rivals... AL East would have 5 matchups, NL Central would have 3, and everyone else would be 3.75
2 weeks - Inter-league Play
6 weeks - League Opponent

[lp815]

 

[Canada8999]

I like the idea of a 3-day series, but I think it our pitching staffs might be an issue.  As teams diverge from their MLB counterparts, their pitchers starts will not be evenly distributed - they won't have one starter going every game.  In a 3 day series, you might have 4 starts in one and then only 1 or none in the next series (especially when injuries work their ways in).  I think we should use a 1-week period for now.

[Canada8999]

There seems to be good support for it, so I'll join in and agree with a points based system, with the category totals for Hitting, Fielding and Pitching weighted according to mj's suggestions.

Where do we stand on what those categories will be?  Would Fantrax be able to implement something for a Win Shares for each, or a Runs Created / Runs Prevented type system?  With quality based counting stats like those, I'm not sure we'd need a pro-rating system or playing time limits.

[Canada8999]

Fangraphs tracks each player's Win Values for Hitting, Fielding and Pitching... the also provide an explination for how each is calculated http://www.fangraphs.com/blogs/index.php/glossary/#winvalues.

The hitting values are essentially based on a players wOBA, adjusted based on league averaged and scaled by their PA.
wOBA = (0.72xNIBB + 0.75xHBP + 0.90x1B + 0.92xRBOE + 1.24x2B + 1.56x3B + 1.95xHR) / PA
http://www.insidethebook.com/woba.shtml

For pitching, it seems to be similar to hitting but with FIP as the metric (again, adjusted based on league average and scaled by IP).
Unfortunately for defense, they use UZR and Fantrax probably doesn't have the data for this.

-------------------------

Edit:

Rather than converting to this Win Value oriented approach, we could stick with the runs oriented one we already have.

For hitting, we use the existing Runs Created formula.
For pitching, we'd do (FIP - league_FIP) * IP, to give a measure of Runs Prevented.
still not sure on fielding, but RF isn't a bad approach. 

We get the team totals for these 3 categories, and then apply the weightings mj suggests.

[Canada8999]

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

... so exactly what you have here, but we adjust the FIP based on league average and the pitchers IP's.

[mjmezzetti]

I have put together hitting and pitching point values for review.  Once I figure out how to incorporate fielding I'll circulate the spreadsheet.

[mjmezzetti]

Two things, however, a team with two first basemen playing at CI will have an edge due to defense statistics.  We would have to pass a rule that only RP's can be placed at RP regardless of eligibility, otherwise an SP with RP eligibility would provide a big edge. Let me know what you guys think.