Fantasy Football 2015: Value Based Guessing

The Something Awful Forums > Discussion > Sports Argument Stadium > The Armchair Quarterback > Fantasy Football 2015: Value Based Guessing

«‹›3 »

Reik: Mar 8, 2004

I'm running a fun league at work, only my second year playing so I'm not exactly sure what I'm doing. I was planning this year on converting the TE position to a WR/TE flex. It feels like the TE position is more volatile than other positions once you get past the third or fourth pick, so I figure instead of mandating it you could just lump the passing TEs in with the WRs. This also allows WRs to become a bigger part of the overalls core, giving them a bigger impact without implementing a PPR. What do you guys think about this change?

# ¿ Aug 3, 2015 19:24

Adbot: ADBOT LOVES YOU

# ¿ May 16, 2024 18:18

Reik: Mar 8, 2004

I've been running a few draft simulations on fantasypros and I always end up with the number 1 overall team, but only because my bench is incredibly higher than the other teams. My starters range from 3rd to 10th in a ten person draft. There must be some flaw in my drafting choices because this has been consistent over like 8+ drafts. Can anyone think what would be causing this?

# ¿ Aug 4, 2015 16:43

Reik: Mar 8, 2004

Metapod posted:

Are you drafting Randle as a starting hb?

No, but I've been not putting as much focus on QBs, picking up like Ryan Tannehill or Phillip Rivers in a late round. After Luck and Rodgers, it doesn't seem worth it to pick up a QB when there's still first in the depth chart RBs and WRs to grab.

# ¿ Aug 4, 2015 16:47

Reik: Mar 8, 2004

Forever_Peace posted:

If you pick up late-round TEs and QBs (like you should), fantasypros will think you have a sub-par starting lineup. Don't be concerned. Having good depth and good upside risk at RB and WR is how you win leagues.

Also getting the #1 team in fantasypros simulations is trivially easy because they sort your draft list by consensus rank but make the bots draft by some kind of noisy sub-optimal ADP system.

fake edit: now I kind of want turn my draft simulator into something interactive so you can actually practice against reasonable competition.

Yeah, if Gronk and Graham are off the table I just pick up a TE in like round 14 right before my defense and kicker. Is there another drafting website with better bots to draft against?

# ¿ Aug 4, 2015 16:55

Reik: Mar 8, 2004

More people should hang out in the IRC channel, #fantasyfootball on synirc.net.

# ¿ Aug 4, 2015 18:51

Reik: Mar 8, 2004

TheZissou posted:

My league allows two keepers but they can't be from my first two picks of the previous season. Here's what I have to choose from:

Andrew Luck
Golden Tate
Knowshon Moreno
Riley Cooper
Stevan Ridley
Panthers D
Hakeem Nicks
Delanie Walker
Benjarvis Green-Ellis
Jordan Todman

Goes without saying that I'm keeping luck, but not sure who else. Any thoughts?

I'd probably keep Tate? What round pick do you have to use on him?

# ¿ Aug 10, 2015 17:14

Reik: Mar 8, 2004

Honore_De_Balzac posted:

I agree.

#1 too many colts
#2 RB depth is weak, don't need both those TE

Seeing as Graham and Kelce are 2 of the 4 TEs (Gronk & Olsen) that have any value above whatever you can get off the waivers, does that make them good trade bait?

# ¿ Aug 14, 2015 16:43

Reik: Mar 8, 2004

Beer4TheBeerGod posted:

Every August I remember that I need to pay Drop Box to keep the pipes open.

Hey, if you are ever looking for help with beersheets stuff I happen to be a statistician and I'm pretty good at excel, just let me know if there's anything I can do.

# ¿ Aug 15, 2015 17:02

Reik: Mar 8, 2004

Beer4TheBeerGod posted:

Help me figure out a not-poo poo way to establish tiers.

The basic idea of your tiers as far as I can tell is you're trying to account for variance in the data to bucket players such that the top player in the tier's projection is not statistically significantly different than the bottom players. The best way to approach this would be to think of X as the higher players distribution and Y as the comparison players distribution, we want to find the probability that X > Y given the mean and variance of both distributions. This is equivalent to finding the probability that the joint distribution X - Y is equal to 0. I'm not sure how you create your projections, but assuming that X and Y are both roughly normal random variables then the joint distribution (X-Y) would also be a normal random variable where mean(X-Y) = Mean(X) - Mean(Y) and variance(X-Y) = Variance(X) + Variance(Y). Once you get your new mean and variance, you can figure out the probability that joint distribution is greater than 0, and compare that probability to whatever statistical significance you want. Using your current rule of one standard deviation would be about a 68% statistical significance. Based on how the tiers are defined, a 50% significance might be more appropriate. A 50% significance would be the equivalent of saying that given the projections, you might as well flip a coin to pick between the two players. Let me know if this helps!

Reik fucked around with this message at 18:47 on Aug 17, 2015

# ¿ Aug 17, 2015 18:42

Reik: Mar 8, 2004

Beer4TheBeerGod posted:

Projections are from 13 "expert" sources (FBG, CBS, 4For4, etc). So if I'm understanding you right, your suggestion is that I do the following:

1. Take the difference in means between two players to get the joint mean.
2. Combine the variance of the two players to get the joint variance.
3. Using the joint mean and joint variance, determine the probability that there is no difference (the joint mean is 0).
4. If the probability is less than 50%, put them in the same bucket.
5. If the probability is greater than 50%, create a new bucket and start over.

Is that correct?

What I was originally doing was taking the two means and two standard deviations. If two two means were within the greater of the two standard deviations, then I assumed they were in the same bucket.

Yeah, if the projections are close enough to normal, you can just use the NORM.DIST function in excel setting x to 0 and cumulative to true. Because the projections probably can't be negative, a gamma distribution might be more appropriate? If that's the case though the joint distributions are a little more complicated than just subtracting means and adding variances. Let me know how far down the rabbit hole you want to go with this. Lets just assume 13 is enough for it to converge to roughly normal.

Reik fucked around with this message at 20:12 on Aug 17, 2015

# ¿ Aug 17, 2015 19:47

Reik: Mar 8, 2004

Forever_Peace posted:

That could work, but the tiers may come out a bit strange sometimes because you're comparing every player to the first player in a tier. Like, consider if the first player happens to be Martavis or Randle. You'd get a bigger than normal tier because the initial variance happened to be high. You might also get some sudden cutoffs, where two players might be indistinguishable from one another, but one barely makes it into the higher tier while the other becomes the start of the lower tier.

Probably the "best" way to do 1D discretizing here is "kernel density estimation", followed by "Jenks natural breaks optimization". Both have solutions in R and Python etc but might be tougher in excel...

Yeah, instead of having a tier system you could have an extra column called like "individual tier" or something where for each player you go down until they're significantly different. These results might not always be intuitive, like the RB5 might have a very high variance so his tier goes down to 9 or 10, but the RB6 is a low variance so his tier only goes down to 8. I don't think that's a bad result, and it actually might help people consider the variance more.

# ¿ Aug 17, 2015 21:15

Reik: Mar 8, 2004

Beer4TheBeerGod posted:

Here's the raw data I'm working with: https://docs.google.com/spreadsheets/d/1Jz08tdIaz7WU9-9OfO8zpGPneSPdfGmNMPe8Z43f-rw/edit?usp=sharing

One of the guys on Reddit uses something called a Gaussian mixture model. Either of you two know anything about what that could be?

You haven't won a championship because FF is arbitrary and capricious.

Based on like 20 seconds on wikipedia, a gaussian mixture model sounds like a fancy way of saying a weighted average of normal (Gaussian) distributions. I haven't worked with one myself, but they probably use it to give more weight to the "better" experts when coming up with their averages?

# ¿ Aug 17, 2015 21:19

Reik: Mar 8, 2004

Beer4TheBeerGod posted:

Sounds like the latest rumor of the Ray Rice experience is that he could be a Bill.

But I liked the Bills

# ¿ Aug 19, 2015 15:33

Reik: Mar 8, 2004

cheese posted:

Gronk is the clear #1 TE, but TE's just dont score that much. His 178 standard points last year put him as RB#9 and WR#10. I want someone with the #1 pick who can score 230+ points in a season and the TE ceiling as we know it is just too low.

Isn't it not so much how many points he can score, but how many more points he can score compared to the next best option available?

# ¿ Aug 22, 2015 21:05

Reik: Mar 8, 2004

Is dropbox dead for anyone else? Can't get the site or app working right now.

# ¿ Aug 30, 2015 15:04

Reik: Mar 8, 2004

ShaneB posted:

I NEED MY SHEETS

Draft in 2 hours

Also need my sheets.

# ¿ Aug 30, 2015 16:01

Reik: Mar 8, 2004

Cpt. Mahatma Gandhi posted:

haha same. Had all my poo poo saved on my work computer (RIP my productivity), which I then so intelligently left at work on Friday.

The only thing I had memorized from the sheets was Charlie Sims' value of 1,000,000

EDIT: Dropbox started working. Charlie Sims must've seen my praise and rewarded me thusly.

# ¿ Aug 30, 2015 16:32

Reik: Mar 8, 2004

Michael Corleone posted:

If I take Andrew Luck 1.2 in my startup 4 keeper league, just how dumb am I? I really can't talk myself out of it. He won't be there at the end of the 2nd, and it isn't value, but I just want to do it, just like the guy that took Watt in the 1st round. Oh well. 6PTD 1PPR.

With 6PTD I think luck is like 4-5th overall, taking him second isn't that dumb.

# ¿ Aug 30, 2015 17:20

Reik: Mar 8, 2004

Who is the Shady McCoy handcuff now? Dixon or Brown?

# ¿ Aug 31, 2015 16:33

Reik: Mar 8, 2004

I'm the Demarco owner in the league, I missed the Ryan Mathews handcuff but I snagged Darren Sproles. Should I attempt to trade for Ryan Mathews? If I do should I trade away Darren Sproles?

# ¿ Aug 31, 2015 16:48

Reik: Mar 8, 2004

I've got an auction draft tomorrow night using the following scoring:

Trying to figure out the closest beersheet to this scoring system and what adjustments I should make, 10 teams $200 budget.. Any help would be appreciated.

# ¿ Sep 3, 2015 14:33

Reik: Mar 8, 2004

Cpt. Mahatma Gandhi posted:

Also, Tom Brady is a free man, so all the Pats WRs just got a major boost (Gronk too I guess, but he was already way up there).

# ¿ Sep 3, 2015 15:27

Reik: Mar 8, 2004

Dandy Kaiser posted:

Guess I didn't need to go Brady/Tannehill, Brady/Tannehill,, Brady/Rivers in three leagues

You've got great QBs during the pats bye week at least?

# ¿ Sep 3, 2015 15:34

Reik: Mar 8, 2004

Incoming headline: "Federal Judge took Tom Brady in the 9th round in his $10,000 fantasy league last week, suspension nullified."

# ¿ Sep 3, 2015 15:35

Reik: Mar 8, 2004

Reik posted:

I've got an auction draft tomorrow night using the following scoring:

Trying to figure out the closest beersheet to this scoring system and what adjustments I should make, 10 teams $200 budget.. Any help would be appreciated.

My bizarro scoring got buried in Tom Brady suspension news. Is there an easy way to see who will be returning punts and kicks for each team? I've never had to look this up before.

# ¿ Sep 3, 2015 16:35

Reik: Mar 8, 2004

Man, I feel so bad for Beer right now. He must be getting swamped with PMs asking for another update to the sheets he released yesterday but with the Tom Brady suspension accounted for.

# ¿ Sep 3, 2015 18:46

Reik: Mar 8, 2004

Leperflesh posted:

The OP is still pointing to sheets from over a week ago. Got a link to beersheets from yesterday?

Also re: Jarryd Hayne - he'll either make the team, or get sniped by another team. No way is he going to get practice squadded. Purely based on his special teams performance, there has to be a team in the league that will want him. I think he's likely to make the fifth spot on the niners roster, though. Purely based on his punt returns he's good enough - averaging 20 yards per return on six returns, and granted it's preaseason, but that's still fantastic. He's also shown the ability to make key tackles on kickoffs/punts. Being able to fill in on the occasional rushing play fills in the value.

http://bleacherreport.com/articles/2553967-san-francisco-49ers-who-jarryd-hayne-would-have-to-beat-out-to-make-the-roster

In terms of fantasy value, you'd only take him as a longshot on a keeper team, ideally in a league that gives points for return yardage, and even then, you're probably just hoping for a 2016 backup RB spot behind someone injury prone.

e. In my dream scenario, the niners trade Kendall Hunter for some help on the OL, and roster Hyde, Bush, Davis, Miller, and Hayne.

https://www.reddit.com/r/fantasyfootball/comments/3jclcq/beersheets_20150902/

# ¿ Sep 3, 2015 19:13

Reik: Mar 8, 2004

IcePhoenix posted:

Someone just traded Jeremy Hill for Greg Olsen in a ppr keeper league (keep four, they are your first four picks) I'm in and I'm on tilt about it someone talk me down.

The only thing I can do to help me justify it is remember that we start 3 wr 2 rb with no flex and he's already keeping lynch and lacy but even then it only brought me down to where I'm at now.

When I used to start going full tilt in video games (LoL was the worst) I would just tell myself "these awful players might be wasting 30 minutes of my life, but I've wasted 30 minutes of my life in much worse ways without outside influence" and it seemed to make me not be so upset at those people because they're not doing something I haven't done to myself.

# ¿ Sep 4, 2015 15:33

Reik: Mar 8, 2004

I'm drafting in 3 hours and the league has .03 per kick return yard and .06 per punt return yard. It's 0.5 PPR and has a 2 point bonus for a 40 yard TD rush or reception. I pulled the list of punt and kick returners from ourlads.com depth charts and put the following together. Trying to figure out how much I should adjust the most relevant beersheet for these players. Very few of them actually show up on the beersheet, which is to be expected since they're returners. The ones I'm thinking to consider adjusting are Antonio Brown, Markus Wheaton, Knile Davis, CJ Spiller, Julian Edelman, Jarvis Landry.

code:

Ari- Brittan Golden
Atl - Devin Hester
Bal - Michael Campanaro, Asa Jackson
Buf - Percy Harvin, Marcus Thigpeh
Car - Ted Ginn
Chi - Marc Mariana
Cin - Adam Jones, Brandon Tate
Cle - Travis Benjamin, Marlon Moore
Dal - Lucky Whitehead
Den - Omar Bolden
Det - Jeremy Ross
GB - Micah Hyde, Ty Montgomery
Hou - Chandler Worthy
Ind - Phillip Dorsett, Dan Herron
Jag - Denard Robinson, Rashad Greene
KC - Knile Davis, De'Anthony Thomas
Mia - Jarvis Landry
Min - Marcus Sherels, Cordarrelle Patterson
NE - Julian Edelman, Danny Amendola
NO - CJ Spiller
NYG - Dwayne Harris
NYJ - Jeremy Kerley, Chris Owusu
OAK - Travis Carrie
PHI - Darren Sproles, Josh Huff
PIT - Antonio Brown, Markus Wheaton
SD - Jacoby Jones
SEA - Doug Baldwin
SF - Reggie Bush
StL - Tavon Austin, Benny Cunningham
TB - Kaelin Clay
TEN - Dexter McCluster
WAS - Andre Roberts

# ¿ Sep 5, 2015 01:20

Reik: Mar 8, 2004

Spoeank posted:

Reggie Bush isn't returning punts/kicks. Jarryd Hayne is probably the PR with Bruce Ellington as the KR.

Is this just from watching pre-season/practice stuff? I'd love a better source than just this one set of depth charts.

# ¿ Sep 5, 2015 01:36

Reik: Mar 8, 2004

This is the bench right now in one of my 10-man leagues. Trying to figure out if anyone should be considered droppable for some of the lottery ticket RBs out there.

It's 0.5 PPR, 2 point bonus for a 40 yd Rushing or Receiving TD, .03 per kick return yard, .06 per punt return yard

CJ Spiller
Allen Robinson
Steve Smith Sr.
Marques Colston
Devonta Freeman
Markus Wheaton
Knile Davis

I bolded the players I picked up because I feel they may see touches in addition to being listed on the special teams depth charts for kick and or punt returns. A Brown is listed as the punt returner, but Wheaton is listed 1 for kick returns and 2 for punt returns. I was thinking given how pivotal A Brown is they might shift all return duties to Wheaton to reduce his injury exposure. Are any of those players worth dropping for a high-upside high-risk RB?

# ¿ Sep 8, 2015 15:58

Reik: Mar 8, 2004

Dandy Kaiser posted:

I agree and disagree. I always try to leave a spot or two on my bench, depending on roster size, for fliers, but up until that point I'm taking the BPA for every single pick. Injuries happen, and if they don't, the more talent you have the better, never turn down a tradable asset because you never know when someone will be in need and you can turn bench fodder into a better roster. Depending on how good your starters are, that leaves you more open to taking on risk, but for most teams, I think that resigning a good portion of your bench to high-risk/reward guys instead of just grabbing a guy or two, will more often than not force you to spend the entire season playing WW roulette

I think how much you're willing to work on trades and how receptive your league is to trades should play a big factor in how you should roster your bench. If trading didn't exist, keeping 1-2 RB and WR to cover injuries then stashing a bunch of high-risk lottery tickets would be the way to go in my opinion. However, taking the BPA at all times definitely increases your ability to fill in other team's holes for the reward to improving the quality of your starters.

# ¿ Sep 8, 2015 16:12

Reik: Mar 8, 2004

winvirus posted:

I have Ertz and Delanie at TE and a WR corp of Emmanuel Sanders and DeSean Jackson. 14 man standard. I got a couple people interested in Ertz, should I try and trade him for an upgrade of WR or just hang tight? I feel kinda weak at WR.

Who's out there in the TE FA market?

# ¿ Sep 8, 2015 21:34

Reik: Mar 8, 2004

If all the top RBs perform according to their ADP this week, who would be a fair trade for Antonio Brown?

# ¿ Sep 11, 2015 15:23

Reik: Mar 8, 2004

Ty1990 posted:

PPR? Straight up?

In PPR I'd probably only give up Brown for Bell.

Sorry, 0.5 PPR

# ¿ Sep 11, 2015 15:49

Reik: Mar 8, 2004

Honore_De_Balzac posted:

Reed or Eifert

Seconding Eifert

# ¿ Sep 11, 2015 16:10

Reik: Mar 8, 2004

Prophecy120 posted:

Even in half ppr, Brown is still in the top 3 players in the game. Not sure why anyone would want to trade him right now.

There is a real chance A Brown catches a pass from a dog murderer this year. Would prefer to not have that bad mojo on my team.

# ¿ Sep 11, 2015 19:04

Reik: Mar 8, 2004

RVProfootballer posted:

lol

There's an even better chance he catches 120+ passes from a rapist this year

In my crazy screwed up head dog murderer is unacceptable.

# ¿ Sep 11, 2015 19:09

Reik: Mar 8, 2004

Spoeank posted:

let me tell you about the exploits of one Mister Benjamin Todd "Big Ben" "Ol' Gray Dick" Roethlisberger

RVProfootballer posted:

You can take solace in the fact that Vick was convicted and Ben only (1) settled out of court and (2) was never charged, you don't have to go through the kinds of mental gymnastics necessary to justify one over the other (Unless you were thinking my lol meant I think Vick is a cool dude, which I don't... the lol was solely because Ben is also probably a piece of poo poo too, haha)

Fortunately I'm dumping all Steelers players, A Brown just happened to be the only one I had in my leagues that wasn't waiver-fodder.

# ¿ Sep 11, 2015 19:13

Adbot: ADBOT LOVES YOU

# ¿ May 16, 2024 18:18

Reik: Mar 8, 2004

Dandy Kaiser posted:

you can take solace in the fact that vick's dog murdering ways were endemic of his upbringing and he's more than paid his dues and mainly seems like a more than respectable human being since

or you can just accept the fact that this game you're playing is basically the stock market with human lives and akin to gambling denarii on gladiators and just play to win and stop being a babby

It is playing to win if you're superstitious.

EDIT: 2009-2013 Eagles and 2014 Jets, furthest was first round in playoffs.

Reik fucked around with this message at 19:27 on Sep 11, 2015

# ¿ Sep 11, 2015 19:22

The Something Awful Forums > Discussion > Sports Argument Stadium > The Armchair Quarterback > Fantasy Football 2015: Value Based Guessing

«‹›3 »