The Test (MUST READ THREAD FOR FANTASY PLAYERS)

[Mephistopheles]

This thread is about a particular strategy for fantasy. I've autodrafted four teams on Yahoo for no particular reason than to analyze the outcomes and use them four my four drafts on my regular ID. I don't play on keeping up with the teams and Im trying to make trades, just for no reason than its fun. Like I said I autodrafted the teams just so I could get a distribution of four Yahoo leagues based on the projections I have from James/ZiPS/PECOTA and how each team drafted. I am of course trying to build the best team. I have one hell of a team that I really reordered the crap out of the predraft rankings, one where I didn't change them at all (this team is soso), one where I changed them some (this teams alright).....and one team that is a draft where I reordered the rankings specifically for one fantasy strategy to test this out. And here's what this threads (sorta) about.

 

First before I get to that I have good news. The fantasy rankings spreadsheets you all want is almost done. I have completely finished the offense side of the work. Here's a preview:

http://i47.photobucket.com/albums/f189/kctigers23/Fantasy_1.gif

The first screenshot is replacement and average levels for a particular position. The bottom stuff is the distribution data needed in order to find out how valuable a certain player is by computing his impact on the average team. Then there is a page where you enter all the draft data for your league. You dont need positions or draft spot. Just type in the names where they go:

http://i47.photobucket.com/albums/f189/kctigers23/Fantasy_2.gif

On the fly another page calculates the team totals. It spits out rotoscores based on each stat. It calculates it for your league (!) and normalizes it for all leagues which is where the probability stuff comes in again. Here's that page for hitters on that league:

http://i47.photobucket.com/albums/f189/kctigers23/Fantasy_7.gif

The team that is in the data right now is the team that I said was epic or w/e, Electric Ladyland. As you can see the offense quite dynamic. I need to update the thing for data for pitchers. I am working now on using depth charts to predict playing time. I did this with the other stuff, so it corrects for say, Derrek Lee only having 480 PAs or whatever it was. I was relatively conservative on the PT projections for everyone. Another thing that you can do is enter your team. Delete the name of one player. Then you can order the spreadsheet and it will order all the players based on which one will help your team the most. Position is not included in this, but if you cant decide between a few players fire the thing up order it and see which one's strengths help you the most. Another thing this can do is quickly analyze trades and see if theyre good. You can also delete one of your key players, say for in example you have a high BA on your team but you need power and your SS is jeter. Perhaps the dropoff from Jeter to say Stephen Drew (the 15th or so rated SS) is not that large because Drews BA and SB problems arent as big of a deal for you. You can quickly see this and realize that trading Jeter for Drew and someone who is a fourth round or third round quality player elsewhere can lead to tremendous benefits for your team.

 

This strategy tends to help a few players. On my team listed here the most valuable player is probably Chris Young. With the exception of him and Reynolds everyone else has a decent chance of hitting over .300, so his one glaring hole is a non-factor. At the same time, if you have Adam Dunn on your team someone like Matt Kemp becomes ultra-valuable in the 10th round to offset the BA and add steals. A tool like this can help analyze your team.

 

Anyways, back to the test. Okay, the strategy is a decent one: It's a rotoleague and you're punting two categories and think you can win the rest. You punt HR and RBI. You want Reyes, but if not it's not THAT big of a deal. Draft pitching and leadoff hitters basically (guys good in BA, R and RBI). My team is this:

 

C - Russell Martin (3-33)

1B - Todd Helton (11-129)

2B - Placido Polanco (14-160)

SS - Ryan Theriot (15-177)

3B - Kevin Youkilis (16-184)

OF - Matt Kemp (10-112)

OF - Shane Victorino (9-105)

OF - Willy Taveras (13-153)

UT - Chone Figgins (4-40)

 

SP - Johan Santana (1-9)

SP - Jake Peavy (2-16)

SP - Cole Hamels (5-57)

SP - Felix Hernandez (7-81)

SP - Scott Kazmir (8-88)

 

RP - Francisco Rodriguez (6-64)

RP - Manny Corpas (12-126)

RP - CJ Wilson (21-249)

 

Remember 1250 innings limit. That's about 1200 right there. I can get a few saves here and there with the rest of the innings.

 

BN - Curt Schilling (17-201) Dropped

BN - Tom Gorzelanny (18-208) Trade Bait

BN - Bill Hall (20-232) Trade Bait if he bounces back

BN - Rocco Baldelli (19-225) trade Bait if he bounces back

 

I am clearly going to have a K/9 well over 9.00. I will win Ks, Ws, WHIP, ERA by HUGE margins. So that's 12*4=48 pts. I will probably finish top 3 in saves if I can get 30 saves off the waiver wire (which I usually can). So that's a 58 point pitching staff. Even if a guy goes down, I can pick up a decent guy and the rest of them will keep the WHIP and ERA so low any hit I take from a guy like Greg Maddux's no Ks won't affect me at all. I seriously will end up with about 1300-1350 Ks with this team it's unreal. So a 58-60 point pitching staff. Now the offense. It's not very good. Admittedly. It's average. Very very average. It'll finish dead last in HR and RBI by about 50 HRs and 300 RBI (lmao). But then again.....it's going to win batting average and it'll be second or first in runs.... and it's going to win steals by a ridiculous margin. It's a 200-250 steal team lol. That's with Ryan Theriot scoring 70 runs and hitting .260 at SS. In fact here's a substitution: Right now the team has 37.15 projected points on offense. We can drop Theriot for David Eckstein and go up to 37.5. So our team is about 5 points better than an average offense. We're predicted to get about all 12 points from R, RBI and BA...so that's 38 points couples with our 58 from pitching: poof 58+38=96 rotopoints which will win the league everytime.

 

So it's all nice in theory, but this test is to see it in practice. I don't think I'll have to update it very much if at all. BUT I WILL KEEP YOU GUYS UPDATED!

[nilodnayr]

I had 96 points last year and came in second :cry: For the second year in a row the winner had over 100 points

 

Very cool tool.

 

No one has purposefully tried the punting category strategy in my league to the extreme, and given the dynamics of it, I don't think it would work. Our league has a few really good players a few really bad players and the rest are below average. Basically I dont need to punt everything to be in the top 3 in all categories (the only thing I wasn't top three was effen BA). I know, I should join a better league, but I kinda find if fun competing against the same few smart managers year in year out.

 

I prefer a well balanced team to "destroying the competition" in a few categories. The marginal return on an additional steal when you are already winning the category by 20 is nothing. Now obviously, then you can trade from a position of strength, BUT when you punt categories, the marginal return on an additional HR when you are already last in the category by 20 is nothing. I know I'm not telling you anything you don't already know. I like to plan to be above average in all categories. Then towards the end of the year I can decided to trade a closer and lose a point to gain 4 points in HR/RBIs or something like that. There always is some separation.

[Mephistopheles]

That's why this is a test. It works brilliant in theory, in practice I don't know. It's not going to work in theory in a H2H league though. If this is the same league as the other thread it won't work. This is based on relative scarcity working off of scratch each season. In a keeper league the distribution of quality players becomes more confined on one or two teams. It's very likely that one guy ends up with 3 of the top 10 players.....so in the end 100 pt teams are much more common. You've gotta get an average of 10 points in each category. So that means you have to be, on average, in the top 16 percent. If you assume that that they're correlated 100 percent (they are not) we'd expect at most a 16 percent chance of getting beat. It's probably more like 5 percent, but I am not sure. Usually the winners are between 85-95 Edited February 9, 2008 by Mephistopheles

[nilodnayr]

That's why this is a test. It works brilliant in theory, in practice I don't know. It's not going to work in theory in a H2H league though.

 

 

F H2H

[Mephistopheles]

I like the variability of h2h and the ability to monitor the league every day.

[nilodnayr]

I like the variability of h2h and the ability to monitor the league every day.

 

I don't like the variability of h2h and its more about gaminig the system than actually putting together a good team. I play roto and still monitor it everyday.

[Transmogrified Tiger]

I like the variability of h2h and the ability to monitor the league every day.

 

I don't like the variability of h2h and its more about gaminig the system than actually putting together a good team. I play roto and still monitor it everyday.

 

I really like that I have an opponent in H2H. Playing against the field makes it very easy for me to lose interest, I don't have the attention span to focus on the minutiae needed to gain subtle advantages in Roto.

[Mephistopheles]

To update.

 

The first waiver transactions have gone down. A couple notables have been dropped (idiots!). Jacoby Ellsbury and Carlos Marmol have been dropped. I have fourth waiver priority and have a claim on both of the. Ellsbury is the quintessential guy for my team when he plays. High BA, lots of steals, no power and lots of runs. Carlos Marmol would get me a handful of saves (maybe) and of course boost up that K rate (as if I needed it). I also have a claim on Troy Percival who was dropped for more saves. If he's healthy he ought to get a decent amount of save opps with the Rays solid offense and improving pitching staff as they march their way to about seventy-five runs.

[nilodnayr]

To update.

 

The first waiver transactions have gone down. A couple notables have been dropped (idiots!). Jacoby Ellsbury and Carlos Marmol have been dropped. I have fourth waiver priority and have a claim on both of the. Ellsbury is the quintessential guy for my team when he plays. High BA, lots of steals, no power and lots of runs. Carlos Marmol would get me a handful of saves (maybe) and of course boost up that K rate (as if I needed it). I also have a claim on Troy Percival who was dropped for more saves. If he's healthy he ought to get a decent amount of save opps with the Rays solid offense and improving pitching staff as they march their way to about seventy-five runs.

 

I think Percival is going to get a lot more save opps than people think. The DRays should have a much much much improved bullpen that will bridge the gap from their good young starters.

 

Getting back to the Rays' bullpen, the last three years have been especially gruesome—in that span the Rays relief has had aggregate ARPs of -40, -33.2, and -95.4, respectively. Last year's figure was the worst in the Baseball Prospectus database, which dates back to 1959. Using the sabermetrically-standard conversion of 10 runs equals a win, Devil Rays relievers cost the team nearly 10 wins compared to what an average pen would have produced pitching in the same situations.