Adjusting for position scarcity

Well folks, what I thought would be a weekly series has
turned out to be a little bit less frequent than that, but I’m back to pick up
the discussion on how to value and rank players on draft day. First, to shake
off the rust, here’s a quick review of what we’ve covered so far:

 

1. Player Value and Context

 

2. Player Valuation and SGP

 

3. Building Better Player Projections:

 

Or, to recap it all in a sentence, projections are important
to establish basic expectations for players, which we can then use to measure
them against other players in context to obtain relative values. This is done
by developing projections and the resultant Standings Gain Points (SGP) for
each player, then adjusting the raw SGP values based on position scarcity to determine
the relative value of a 30-steal shortstop compared to a 30-homer outfielder or
a 30-save closer.

 

To illustrate this, let’s use NFBC, since that’s the main
league I focus on in my draft prep; it’s a 15-team mixed with a 30-round draft,
although I only base my player valuations on the 23-man active rosters, or 345
total players… everyone beyond that is essentially replacement level by
definition.

 

To illustrate how player stats have different values at
different positions, let’s look at two positions with very different offensive
standards, first base and catcher. Here is a pair of players from each position
with the projected stats and raw SGP (that is, ranked against all players, not
by position) that I used to prepare for my NFBC draft last season:

NAME                     AVG     R          HR        RBI       SB        SGP (raw)

Pujols,
Albert           .337      119       38         125       6          11.80

Garko,
Ryan            .281      59         16         65         1          0.49

 

NAME                     AVG     R          HR        RBI       SB        SGP (raw)

Martin,
Russell        .293      89         16         73         14         4.60

Teagarden,
Taylor    .250      37         11         39         1          -3.10

 

Pujols.jpg

Hey,
it was a lousy projection on Martin, but a lot of people were wrong about him!
Anyway, Pujols (right) is clearly far ahead of Martin by any objective measure, as well
he should be. But a 15-team league only requires somewhere between 15-25 first
basemen to be drafted, depending on how many corner picks are first basemen,
whereas a minimum of 30 catchers must be selected in a two-catcher format such
as this. You need more catchers, and they generally provide less offense than first basemen, so you need to adjust for this scarcity.

To
adjust for position scarcity, use the replacement-level player at each position
as “0” and then adjust each higher-ranked player’s SGP accordingly. That’s where Garko and Teagarden come in; Garko was the lowest-ranked (by raw SGP) first baseman who made the
cut in the top 345, while Teagarden was my 30th-ranked catcher.

(Technically,
replacement level should be determined based on the highest-ranked undrafted
player, but as you get lower on each position list, players become more
interchangeable, so using these guys for this demonstration won’t significantly
adjust the math.)

 

After doing this math, Pujols’
weighted SGP becomes his raw value (11.80) minus the replacement value (0.49),
while the same holds true for the catchers. Let’s look at the same pair of
players with those positional SGP adjustments made:

 

NAME                     AVG     R          HR        RBI       SB        SGP (weighted)

Pujols,
Albert           .337      119       38         125       6          11.31

Garko,
Ryan            .281      59         16         65         1          0.00

 

NAME                     AVG     R          HR        RBI       SB        SGP (weighted)

Martin,
Russell        .293      89         16         73         14         7.70

Teagarden,
Taylor    .250      37         11         39         1          0.00

 

Martin.jpg

Pujols
is of course still ahead of Martin (right), but the gap has closed from 7.40 SGP to
3.61 SGP… in fact, Pujols actually lost a few fractions off his raw SGP because
the replacement level player at his position was still above the overall
replacement level for the entire 345 player universe. Pujols remains near the
top of the pack, but instead of Martin being a mid-round player based on these
projections, he moves up into the mid single-digit rounds.

 

Repeat
this exercise for all of the 345 players projected to be drafted, then re-rank
by these weighted SGP values, and you have a player rankings list that adjusts
for position scarcity.

 

One
further step you can consider is tweaking the SGP adjustments to try and tease
out how your league values saves, steals or other categories… for instance, I
weight my rankings against average draft position (ADP) data from other sources
to try and determine which players will be overvalued or undervalued. Certainly
the results of any real draft may vary greatly from ADP, but given that the
exercise of draft prep is less about figuring out the best players than trying
to figure out when is the right time to pick them, it’s another valuable step
of the process.

 

Next
week we’ll focus a little more on closers, because getting a couple of the
top-shelf is a big part of the basic 411 strategy…

 

Thanks,

Cory

 

15 Comments

CS and the Crew,

Great work as always, I’m loving the 101 class on ranking players by Professor Schwartz. Since you’ll be discussing closers in the next section, I’ve a keeper question for you involving the first one on your list… alphabetically that is.

I have Aardsma in my AL-only at Round 15 this year, which is fantastic for any closer but I’m just not a believer in the DA (I am a believer in Dream Weaver, but I digress). Another owner offered me J. Shields at Round 9 this year, can keep him at Round 5 next year but probably won’t. DA can be kept at Round 13 in 2011, Round 10 in 2012, Round 6 in 2013 and Round 1 in 2014. This is essentially a swap of worst keepers, and all of my other keepers are hitters. Is this a good deal?

Thanks for the knowledge,

Scotty Mac
Seattle, WA

Cory,

Just wondering if you’re going to update the awesome player projections you provided last year for 2010 … any chance? I was following your SGP guidance and realize I don’t have reliable projections to build off of … anyway, let us know as I’m sure there are a bunch of us following this blog that’ll find it useful.

Thanks!
Perminator

Perminator, I’m still cobbling together the assorted projections I used as the basis for my own, so look for that some time in February.

Scotty Mac, I’ve kinda fallen out of love with Shields. His trends are flat/declining at a point when they should be peaking, and it seems like this is as good as he’s going to get. Not that that’s bad, but for a guy coming off a huge season like Aardsma is, I’d try to hold out for something more.

Thanks,
Cory

Cory,

As a fellow quant jock I want to get your advice on dealing with auction draft inflation. I’m in an NL ONly 4×4 keeper league. 10 teams. We can keep as many as 14 players so anybody coming off another fantasy team’s contract get big money. Pujols, Reyes, and Wright all went for$60+ last year!

Here’s my strategy. On average in my league the champ out scores the average team by 25% in all categories (obviously dominance varies by category but 25% average across the board). Which tells me I have to get at least 25% more value for my salary cap ($260) to win the league. So, somehow I need to get $325 worth of players for $260 on draft day.

Now, I have keepers many of whom are real bargains (Carpenter at $10, Ubaldo at $3, Ludwick at $1). I calculated an early value for these players of $241 at a salary of $79 (I’m still working on my $ values but this is an early guess). So the way I see it, I’m ahead of the game here … I have $181 ($260-79) to spend to get $92 ($325 – 241) in value.

This gives my “inflation cap”, basically the amount I can over bid on my average player and still meet the $325 target. I’m using my keeper bargains to subsidize overpaying for other players. I can afford to pay about 96% more than my estimated value to the average player on my team and still meet the $325 target value number.

What do you think of this approach and do you have another better way of making sense of draft inflation?

Thanks!
Perm

guys,

can only keep 5….5×5 12 team roto i have round drafted last year beside name..thanks

draft value

Jeter (7th)
Hill (waivers)
Haren (6th)
Fran Cordero (12th)
Choo (16th)
Sandoval (15th)
Nolasco (8th)
Scherz (14th)
Fuentes (11th)

Kyle,

My 5 would be Jeter, Hill, Cordero, Choo and Sandoval.

Zach

Perminator, your general concept makes sense, in that you use keeper bargains to subsidize paying premiums on draft day. However, I don’t agree with a straight across-the-board premium of “x” percent on each player, because of course you may still get some players at fair market or below value just through the nature of the draft itself.

Here’s an interesting article that’s been making the rounds the last couple of days, which might help you tackle the problem of figuring how much premium to pay, and for whom, as the draft is going on:

http://www.fantasyballjunkie.com/?p=2544

Enjoy!

CS

Who has the better career offensive numbers? Brett Wallace, Matt LaPorta, Travis Snider.

Thanks Cory. Great article. I’ve never felt more geeky than I did while reading it but I think it’ll help me win my league!

What would you do with these gusy? Uggla at $20, Loney at $15, Brian Wilson at $15, Russel Martin at $20 (2 C league) …

Thanks!
Perm

Cory,
Good stuff as usual. I use SGP based on composite projections from 5 or 6 sources to get my dollar values. So far this year I only have chone and marcel. I have a question about scarcity. Do you lump SS and 2B into one MI pool and 1B, 3B into a CI pool or do you keep them seperate? Also how do you deal with DH only players?

Thanks,

Brandon

What’s up 411,
I’m glad baseball is starting to kick in. I have a question about switching from a draft to an auction. It’s a keeper head-to-head league in which up to 4 players could be kept at 1 round higher than the previous season. We all want to change to an auction, but are a little stuck on how to achieve this. One option that is being floated around is to get some projected dollar values based on our 10X10 rankings and then give a discount (2009 round X 2%) to those players. Does this seem like a fair way to do it, and does 2% seem like a good number?

Second question, we’ve been plagued by pre-deadline “dump” trades. I don’t necessarily like the word “dump” as it seems to have bad connotations, but they are generally two or three good to great players for a great player at a great discount. I don’t personally have a problem with these trades, but they’re causing a rift down the center of our league. Some of the more involved owners have been trying to hash out guidelines to curtail these trades, and have kind of agreed that a salary ceiling and floor would at least make it very cut-and-dry as to what trades would be allowed. And now to the question(s). What do you think of this idea? If we’re starting with $260 on draft day, what’s a reasonable ceiling/floor? What is generally done with players on the DL (can they be thrown in to help equal out salary)? Also, as far as FAAB bidding goes, how should their value be included in the cap (FAAB bid = cap hit, or a set value ($3-5) for each player added through FAAB process)? I know there is no right way to run a keeper league, but if you could lend some insight on your experiences with this sort of league it would be greatly appreciated. Thanks in advance and bring on the baseball.
Andy

Hey guys. I have a couple of quick questions. How does moving to NYY effect Vazquez? In a vacuum who do you like more between Nolan Reimold and Garrett Jones and why? Thanks!
Everett

Cory and Mike which players on the indians can have a good fantasy impact in 2010?

Quick answers, working backwards…

Everett, moving to NY is a net negative for Vazquez, because he’ll be pitching in a tougher pitchers’ park (where his occasional gopheritis will hurt more than in ATL), and in a tough offensive division in the tougher league. But, he still has excellent skills and will get tons of offensive and bullpen support, so figure a few more wins this year but a spike in ERA to the high 3.00’s… solid #2/3 in vanilla leagues. Also I like Reimold more than Jones, who looks more like a half-season wonder to me. I don’t expect greatness from either though.

Andy, why not just re-draft? Or if you’re going to maintain the keepers, assign the fair-market values in MLB.com’s draft preview? Re. dump trades, there has to be a penalty attached to discourage that… I prefer one where teams lower in the standings get penalized keepers!

Brandon, I treat all positions (including DH) distinctly, and if a player is eligible at multiple positions, I weight his SGP according to the weakest position where he is eligible (i.e. 2B instead of OF, as in Zobrist’s case).

Perm, keep Wilson, maybe keep Uggla, cut Loney and Martin unless you think they’d go for significantly more than that in the draft, although I might cut them anyway.

Thanks,
Cory

Cory:

I’ve been a big fan of your work and feel like the idea of SGP is the best I’ve seen in quite some time.

I’m trying to take the idea of the “weighted projection” to the next level in preparation for my 12-team mixed keeper league(been around for 12 years or so). Tell me if this follows your logic.

In the keeper league all the teams keep 5 players based upon a tiered system. I’ve gone through and subtracted out the keepers from the active pool and “re-weighted” the rankings to account for the additional keeper scarcity.

So for example, for SP our league will require 72 SP to be drafted. So to get the “weighted” SGP I went to the 72 ranked SP, set him at ?0?, and subtracted his SGP from the rest of the pool and sorted the list.

Next, to get “keeper weighted” I added up the amount of SP being kept, 9 of them, and subtracted it from the 72. So then I went to the 61st ranked AVAILABLE SP and set that player’s SGP at “0” and re-weighted everyone else (it took an almost additional 3 SGP’s from SP!).

Does this further step seem logical to you? I want to make sure I?m not missing anything.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 257 other followers

%d bloggers like this: