## How to use Marginal SGP's properly

Moderator: Baseball Moderators

### How to use Marginal SGP's properly

I've done a bit of reading on some player value strategies to use for my 22 team 5x5 roto auction next year. I'm intrigued with Art McGee's Marginal Standings Gain Points. I want to make sure I have some things right before I start using it however.

1) We would theoretically, draft 22 second basemen, say (would be affected by multi-position guys, util and bench spots, but let's go with it). Now I read that the 22nd best is our "replacement level" 2B. Is this simply the 22nd ranked 2B in terms of projected SGP for the 2010 season, or the 22nd best projected total in each category? I would have to think the latter. Suppose, for instance, that Willy Tavares was our "replacement level" OF. That would throw off any SGP projections for SB. So this seems obvious to me but based on what I've read it seems to suggest one player is our across the board replacement level guy. Can anyone confirm or correct my interpretation of this?

2) How exactly do we calculate SGP for counting stats? I've used the average separation of standings points in my league for this past season, but I'm not sure if I can trust it. For example, HR points were separated by an average of 5.9 HR. This is skewed by a 44 HR difference between second last and last place (leave that out and it's 4). What I've read indicates something closer to 9. How do I go about getting this right?

3) How are rate stat SGP's calculated?

4) Once we've got that, let me see if I have the rest straight. The 22nd best projected HR total for 1B is, let's say, 14. Prince Fielder is projected for, let's say, 36. Each SGP in the HR category is 5.9. 22/5.9=Prince earns 3.7 marginal SGP for HR. Then we move on to RBI. The 22nd best projected 1B total is 66, Prince is projected for 126. Each SGP in my league is worth 15, Prince is worth 4 MSGP in RBI, etc. etc. and then we total it up at the end and we see that Prince is worth 14.4 or whatever more SGP than the composite 22nd best 1B. Each SGP is assigned like \$3 on the basis of a calculation I don't even want to think about right now, and Prince's total worth in this league comes in at \$44 (14.4 x \$3 plus the \$1 I'd spend on a replacement level 1B). Again, all the exact numbers are hypothetical I'm just going through the process.

I'm sure there will be more but those are the basics for now. Does anyone else use this draft strategy?
GabbyJay
College Coach

Posts: 230
(Past Year: -3)
Joined: 1 Aug 2007
Home Cafe: Baseball

### Re: How to use Marginal SGP's properly

I haven't read anything by McGee, but use a similar approach. There's a couple things I do differently than him/your interpretation.

1.Yes, 1 palayer across the board not 22nd best in each category. It's better to take the avg of say spots 21-23 to mitigate having a 1 cat guy like Tavares as your replacement level player. Don't use 22 as replacement level. You need to factor in bench players. Take the total roster spots in the league and do a rough allocation by position.

2.Don't use the top or bottom few teams in each category to avoid the team 44 HR above or below. 22 teams makes this even more valuable. You can compare the 5th place team vs. the 18th place team and get a very meaningful number. I go one step further and use a 3 team avg again. So avg HR's by the 4-6 place team, than avg. 17-19 and calculate the difference and divide.
mblax10
Minor League Mentor

Posts: 422
Joined: 20 Jul 2008
Home Cafe: Baseball

### Re: How to use Marginal SGP's properly

Okay, I did some reading on this, and I'm thinking I don't care for it too much.

Art McGee wrote:for example, in the home run category you'll find that in a standard league it typically takes about 9 and a half homers to move up one spot in home runs --- to gain one marginal point in that category. a guy who hits 30 home runs is probably worth about 3 points in home runs in a typical standard roto league. if you don't play in a standard league, you can still apply the same kind of thinking, but instead of using that 9.5-homer figure, you're going to have to look at leagues similar to yours and estimate how many homers, or steals, or saves it takes to gain 1 SGP in your format.

First of all, I think it's too dependent on good projections. Because of that, it takes away from your focus, which should be on coming up with solid projections for the players in your player pool. Second, SGP calculation (based on Art's comments above) seem to generate standard (linear) gaps between point levels. I don't play categorical leagues, but I would imagine the actual, end-season ranking difference in each category is probably larger on the ends and smaller in the middle, leading to a bell curve-like distribution. If that's accurate, using a linear ranking formula would create a disparity in the projected-vs-actual values. Someone let me know if I'm wrong on this one...

That's my take on his system. Speaking specifically to a 22-team league, I think the best gains in the league will be made on the WW during the season. The first team to grab an "average" player - by 12-team standards - will get a HUGE benefit, and spending so much effort on the pre-draft might lead to the perception that the WW isn't as valuable as it should be.
JTWood
Hall of Fame Hero

Posts: 11253
(Past Year: -255)
Joined: 22 Jun 2004
Home Cafe: Baseball
Location: Unincorporated Heaven

### Re: How to use Marginal SGP's properly

JTWood wrote:First of all, I think it's too dependent on good projections. Because of that, it takes away from your focus, which should be on coming up with solid projections for the players in your player pool.

I'm not so sure about this. My feeling is that there is an endless number of projections put out by teams of people who do this for a living, and rather than spend my spare time trying to top them, I should find a set I trust and try and do something with them, like this.

Second, SGP calculation (based on Art's comments above) seem to generate standard (linear) gaps between point levels. I don't play categorical leagues, but I would imagine the actual, end-season ranking difference in each category is probably larger on the ends and smaller in the middle, leading to a bell curve-like distribution. If that's accurate, using a linear ranking formula would create a disparity in the projected-vs-actual values. Someone let me know if I'm wrong on this one...

I'm finding that the top and the bottom have large spreads and otherwise it's fairly even (of course, this is just looking at one season of one league), so dropping them off probably gives a decently reliable number.

That's my take on his system. Speaking specifically to a 22-team league, I think the best gains in the league will be made on the WW during the season. The first team to grab an "average" player - by 12-team standards - will get a HUGE benefit, and spending so much effort on the pre-draft might lead to the perception that the WW isn't as valuable as it should be.

Ordinarily I would be focusing a lot on the WW in a league like this but unfortunately we are limited to 8 pickups during the season in weekly mini-auctions that start in May. So you either have to nail your draft or nail your trades, or both, really.

I suppose what you could say is that finding those guys who will break out and cost \$20 in a mini-auction, and getting them for \$1 in the auction is a better use of time, but... that's why I'm thinking about this in November, haha.
GabbyJay
College Coach

Posts: 230
(Past Year: -3)
Joined: 1 Aug 2007
Home Cafe: Baseball

### Re: How to use Marginal SGP's properly

GabbyJay wrote:Ordinarily I would be focusing a lot on the WW in a league like this but unfortunately we are limited to 8 pickups during the season in weekly mini-auctions that start in May. So you either have to nail your draft or nail your trades, or both, really.

Ouch.
JTWood
Hall of Fame Hero

Posts: 11253
(Past Year: -255)
Joined: 22 Jun 2004
Home Cafe: Baseball
Location: Unincorporated Heaven

### Re: How to use Marginal SGP's properly

I use this approach in my keeper league and it almost always has me in the hunt. I track the team stats needed in each category, and then use those data, averaged over the prior 4 seasons, to figure out what it takes to finish 1st, 2nd, 3rd, etc in each category. As JTWood guesses, it becomes nonlinear at the top and bottom, but that doesn't matter because you don't need to win any individual category, you just need to earn enough points in each category to win the league. In my 16 team league, it's basically linear between 2nd place and 15th place.

The team totals give me targets for each stat (I usually identify what it takes to win each category as well as a repleacement level set at the midpoint of last and next to last place teams), as well as the marginal standing gains for each stat. I then use these to calculate my league specific player values. The most complicated things is doing this for pitchers, since you have to weight things for IP (technically, you have to do the same thing for hitters, but the differences in IP for pitchers is much greater).

Since my league is 6x6 with holds and OPS, it gives me better info that the usual rankings, and helps me identify under and over valued players in our league very well. It's taken me a few years to refine it, but it takes me just a day or so to get things set up now, and then I'm ready to roll. Highly recommend the approach for roto keeper leagues.
"I don't want to play golf. When I hit a ball, I want someone else to chase it."
GotowarMissAgnes
Hall of Fame Hero

Posts: 5469
(Past Year: -47)
Joined: 12 Dec 2003
Home Cafe: Baseball
Location: Happy Valley