Subscriber Contest (2 Viewers)

[The_Hoser]

After going to the final cut last season, I have crashed and burned already this year. When Forte puts up 6.3 and is your second-best scoring back, you deserve to go home.

 

[QuizGuy66]

More than 40% of all entries were the minimum roster size while just over 20% have 23 or more players. I would theorize that people that took the time to find ways to get a larger roster size are more likely to have taken the competition "seriously". There's likely a higher percentage of people in the 18 to 20 player range that filled out the contest form quickly without really looking over the rules or contemplating what it would take to survive or win. Thus, you have a much higher % of dropouts in that range based on lack of effort than at the larger roster size. I have no way of quantifying this but I think it is a fairly "safe" theory.Taking that into account, it is hard to say what is better to "win" the game. It seems obvious that to just survive, bigger is better, to a point. Even just looking at the 20 to 30 roster sizes and not trying to determine how many 18 or 19 team rosters are "legit", there is a definite positive correlation between more players and better chance of survival. I find it interesting that there is a significant dip at 28 players but then another rise at 29 and 30. So, theoretically looking at just survival, it's would seem that 27 is the optimal number of players at this point. I also have pulled queries for each position by count. that will be the next "chapter".

This is the crux of the 18-player argument. I don't subscribe to it, but you're spot on here. I'm sure that the thoughtfulness that went into the entries with 20+ players does explain part (but certainly not all) of their survival advantage. It'd be kinda fun if the stats on how many times someone changed their entry were available and how many players they ended up with. Surely it would seem that the bigger the roster, the more adjustments that were made (and probably more thought and over-thought).Just for the heck of it, I looked at this week's top 50 to see how many players they each have. One would expect, based on the 18-player guy argument, that there would be more such entries at the top of the list.Stats include the 23 staff entries, of which 19 are alive, btw.18 players. 16 of the top 50 entries (32%) including #1. 5416 entries out of 13061 (41.5%). 3166 out of 8521 live entries (37.2%).19 players. 5 of the top 50 entries (10%). 1905 entries out of 13,061 (14.6%). 1230 out of 8521 live entries (14.4%).20 players. 6 of the top 50 entries (12%). 1338 entries out of 13,061 (10.2%). 900 out of 8521 live entries (10.6%).21 players. 5 of the top 50 entries (10%). 961 entries out of 13,061 (7.4%). 660 out of 8521 live entries (7.7%)22 players. 2 of the top 50 entries (4%). 799 entries out of 13,061 (6.1%). 553 out of 8521 live entries (6.5%).23 players. 2 of the top 50 entries (4%). 615 entries out of 13,061 (4.7%). 449 out of 8521 live entries (5.3%).24 players. 3 of the top 50 entries (6%). 491 entries out of 13,061 (3.8%). 368 out of 8521 live entries (4.3%).25 players. 3 of the top 50 entries (6%). 402 entries out of 13,061 (3.1%). 307 out of 8521 live entries (3.6%).26 players. 3 of the top 50 entries (6%). 308 entries out of 13,061 (2.4%). 240 out of 8521 live entries (2.8%).27 players. 0 of the top 50 entries (0%). 233 entries out of 13,061 (1.8%). 189 out of 8521 live entries (2.2%).28 players. 1 of the top 50 entries (2%). 180 entries out of 13,061 (1.4%). 130 out of 8521 live entries (1.5%).29 players. 3 of the top 50 entries (6%). 142 entries out of 13,061 (1.1%). 112 out of 8521 live entries (1.3%).30 players. 1 of the top 50 entries (2%). 271 entries out of 13,061 (2.1%). 217 out of 8521 live entries (2.5%).Most of the balance of the 18-player entries were in the 26-50 range FWIW.Looking at the top 10 from the last 2 weeks:Week 1: 8 of the top 10 are still alive (1st place and 8th place have been knocked out)Week 2: 6 of the top 10 are still alive (6th, 8th, 9th, and 10th place have been knocked out)(FWIW, 5 of the 6 entries that have been eliminated had 18 players. The other had 21 players.)So there you go.-QG

 

[Modog814]

Actually many of the small roster guys argue quality over quantity will get them through to the finals. They stand by this for a few weeks until the decimation begins and they have to abandoned that postion. Then the argument solely focuses on the playoff weeks. Once those come, the cycle repeats and they disappeared until next year when the same argument is renewed with the qualifiers that the rules and/or prices changed or the sample size of two seasons isn't enough to make a conclusion.That said, they should have a fighting chance this year based on the sheer number of 18 man rosters entered.

If this is true (that they've been arguing this) then either I'm totally missing it, or we're interprating the same arguments differently. Every argument I've seen (or at least the ones I'm taking seriously) is that they realize that it hurts their chances to get to the 250, but they believe it will give them an advantage if they do get there.

 

[Modog814]

More is better to a point. Someone could do a point scored per position per dollar spent analysis. Once you rank the position where the dollars were best spent from top to bottom, you'd optimized the stats above to fit that model.

yes, and if we had that data we could do that, however, I'm not sure that would be so helpful as all it would prove is that that was the optimal lineup for a single sample, and not necessarily every sample, or even the most likely occurrence.

 

[BassNBrew]

ALL - 13061 / 8521

18 - 5416 / 3166

19 - 1905 / 1230

20 - 1338 / 900

21 - 961 / 660

22 - 799 / 553

23 - 615 / 449

24 - 491 / 368

25 - 402 / 307

26 - 308 / 240

27 - 233 / 189

28 - 180 / 130

29 - 142 / 112

30 - 271 / 217

18/19 man rosters started off as 56.1% of the entries, they are down to 51.6% of the entries. Assuming the same drop off continues, they'll compromise 36.6% of the entires in the finals. That's still a very large percentage.

 

[Modog814]

My problem with this analysis , specifically for QB and RB's is that there are multiple strategies within each grouping. Take 2 QB's for example, this could mean 2 high priced studs, 2 middle tier player, 2 lower (less likely), it could mean 1 high priced and a middle tier, 1 high priced and a lower tier, 1 middle tier and a lower tier. So this one alone has 6 possible strategies (3 QB's would be even more). So let's just say, the majority of the top 250 teams (let's say 90%) had 2 QB's. So we draw the conclusion that the optimal roster has 2 QB's. Ok, great...now what? Which of these 2 QB strategies is the way to go?

That's true just looking at 18 vs 30 too, right? Right now, the stats show that having 2, 3 or 4 QBs isn't that big of a difference when bye weeks aren't involved. I think it will be intesting to compare QB stats as the bye weeks hit. My hypothesis is that 3 and 4 QB teams will have an advantage over 2 QB teams when the bye Weeks hit. Time will tell.At RB I think it shows that, at least so far, there isn't a huge advantage to having more than 4 RBs. I think as bye weeks roll around that teams with 5, 6 and 7 will fair significantly better than teams with 4. Right now, if you just pull the top stats it says you should have 3 QBs, 6 RBs, 10 WRs, 4 TEs, 3 Ks and 3Ds for a total of 29 players. Means nothing but just interesting to me.

Absolutely true. My contention is that what makes this contest so great is the infinite strategic paths possible, but it's these paths that will prevent us from ever truly learning what the optimal (if there even is one) strategy is. I mean we have basic data, but it's too basic and doesn't, imo, capture enough of these paths to draw an adequate conclusion. And even if we could break down the data to incorporate the different strategies, it'll lead to samples being too small, so any conclusion we may draw from that wouldn't be very conclusive. Yes I agree that 3 or 4 QB teams will have an advantage over 2 QB teams during the bye weeks overall. However, on a team by team basis, and quite frankly isn't that what we really care about, our teams, the advantage is really only for 2 weeks (the weeks in which our top 2 qb's have a bye). So as a whole, yes 3 QB teams may survive at a rate slightly higher than 2 QB teams, but the question becomes, for my individual team, am I better spending that money elsewhere (on an extra WR) that on a QB? I don't know the answer to that.

 

[BassNBrew]

One thing I find interesting that no one really discusses is the distribution of the number of teams entered at each roster size:

There was some discussion about this a few weeks ago, starting here.I still think it is safe to assume that quite a few of the 18-player rosters were "phoned in" - just filled out very quickly without much thought. It's just a fact that if you ask 10,000 people to perform a task, there will be a certain percentage of those people who will put in the absolute minimum amount of effort required, even if there is a reward involved. I'm sure there are plenty of 18-player rosters that were put together with a lot of thought too, but the others are the reason why there is such a large number of them.

While it would seem that way, their elimination rate increased significantly from wk 2 to wk 3 relative to the field. One would expect the opposite if quite a few were phoned in. Of course to be fair, the significantly better survival rate of the 19 man rosters supports your argument.

 

[LittlePhatty]

18/19 man rosters started off as 56.1% of the entries, they are down to 51.6% of the entries. Assuming the same drop off continues, they'll compromise 36.6% of the entries in the finals. That's still a very large percentage.

Their survival rate will almost certainly go down during the next 7 weeks worth of byes.

 

[Modog814]

One thing I find interesting that no one really discusses is the distribution of the number of teams entered at each roster size:

There was some discussion about this a few weeks ago, starting here.I still think it is safe to assume that quite a few of the 18-player rosters were "phoned in" - just filled out very quickly without much thought. It's just a fact that if you ask 10,000 people to perform a task, there will be a certain percentage of those people who will put in the absolute minimum amount of effort required, even if there is a reward involved. I'm sure there are plenty of 18-player rosters that were put together with a lot of thought too, but the others are the reason why there is such a large number of them.

While it would seem that way, their elimination rate increased significantly from wk 2 to wk 3 relative to the field. One would expect the opposite if quite a few were phoned in. Of course to be fair, the significantly better survival rate of the 19 man rosters supports your argument.

You'd expect it, but expectations (even if are correct) doesn't mean that it will occur as such. Look at the however many K guy still being alive.

 

[BassNBrew]

My problem with this analysis , specifically for QB and RB's is that there are multiple strategies within each grouping. Take 2 QB's for example, this could mean 2 high priced studs, 2 middle tier player, 2 lower (less likely), it could mean 1 high priced and a middle tier, 1 high priced and a lower tier, 1 middle tier and a lower tier. So this one alone has 6 possible strategies (3 QB's would be even more). So let's just say, the majority of the top 250 teams (let's say 90%) had 2 QB's. So we draw the conclusion that the optimal roster has 2 QB's. Ok, great...now what? Which of these 2 QB strategies is the way to go?

That's true just looking at 18 vs 30 too, right? Right now, the stats show that having 2, 3 or 4 QBs isn't that big of a difference when bye weeks aren't involved. I think it will be intesting to compare QB stats as the bye weeks hit. My hypothesis is that 3 and 4 QB teams will have an advantage over 2 QB teams when the bye Weeks hit. Time will tell.At RB I think it shows that, at least so far, there isn't a huge advantage to having more than 4 RBs. I think as bye weeks roll around that teams with 5, 6 and 7 will fair significantly better than teams with 4.

Right now, if you just pull the top stats it says you should have 3 QBs, 6 RBs, 10 WRs, 4 TEs, 3 Ks and 3Ds for a total of 29 players. Means nothing but just interesting to me.

Absolutely true. My contention is that what makes this contest so great is the infinite strategic paths possible, but it's these paths that will prevent us from ever truly learning what the optimal (if there even is one) strategy is. I mean we have basic data, but it's too basic and doesn't, imo, capture enough of these paths to draw an adequate conclusion. And even if we could break down the data to incorporate the different strategies, it'll lead to samples being too small, so any conclusion we may draw from that wouldn't be very conclusive.

Yes I agree that 3 or 4 QB teams will have an advantage over 2 QB teams during the bye weeks overall. However, on a team by team basis, and quite frankly isn't that what we really care about, our teams, the advantage is really only for 2 weeks (the weeks in which our top 2 qb's have a bye). So as a whole, yes 3 QB teams may survive at a rate slightly higher than 2 QB teams, but the question becomes, for my individual team, am I better spending that money elsewhere (on an extra WR) that on a QB? I don't know the answer to that.

Assuming you avoid injuries and get decent production from your QBs, you are definately better off with only 2 QBs in most instances. Top producing QBs are relatively consistant as a pair in best ball and expensive. When you start carrying more than two, you are tieing up alot of dollars that you usually won't be using. This coming from a guy who took three this year.

 

[BassNBrew]

18/19 man rosters started off as 56.1% of the entries, they are down to 51.6% of the entries. Assuming the same drop off continues, they'll compromise 36.6% of the entries in the finals. That's still a very large percentage.

Their survival rate will almost certainly go down during the next 7 weeks worth of byes.

Spot on.I wish we could sort by roster size and specific players.

 

[WisWolvrns]

Top 25 with Total Points Week 1-3

Rank	Entry	Total Pts	Roster1	105904	655.75	272	109740	652.65	223	111368	647.05	234	101434	643.65	185	106679	640.1	286	112144	632.4	237	107616	632.35	218	102676	629.85	189	107217	629.5	2910	104966	629	2611	109491	628.15	2012	109864	628	1913	100656	627.6	3014	106783	627.1	1915	109637	626.55	2516	108046	623.75	1917	103050	623.3	1818	104307	622.45	2019	110554	621.55	1820	106317	621.35	1821	106219	621.3	3022	105080	619.3	2023	101788	618.15	2024	104835	618.1	1825	110751	617.85	21

 

[BassNBrew]

Just for something to kick around, the 6 rosters eliminated by less than a yard this week.

Matt Schaub $23 10.05 42.05 16.05

Carson Palmer $16 29.15 8.25 14.55

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

Ray Rice $37 7.20 13.70 11.60

Rashard Mendenhall $30 20.50 7.70 20.30

Cadillac Williams $16 10.80 6.40 7.10

Arian Foster $13 42.30 15.30 14.60

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

Greg Jennings $26 19.20 6.60 9.80

Steve Smith $24 9.30 7.50 19.30

Malcom Floyd $17 7.80 18.50 21.70

Mike Williams $8 14.00 13.40 10.50

Laurent Robinson $7 10.80 1.40 0.00

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

Visanthe Shiancoe $12 19.60 17.60 2.20

Jermaine Gresham $6 17.50 6.00 11.90

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

Neil Rackers $4 13.00 17.00 8.00

Rob Bironas $3 10.00 3.00 11.00

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

Carolina Panthers $4 11.00 0.00 5.00

St. Louis Rams $2 10.00 6.00 5.00

Detroit Lions $2 12.00 6.00 8.00

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

TOTAL 198.05 158.75 144.95

CUTOFF 125.90 141.75 145.00

18 players. Balanced stud lineup w/ double digit production everywhere. QBs let them down and TE production was slightly off. Tough break for this team.

 

[QuizGuy66]

Top 50 with Total Points Week 1-31 105904 655.752 109740 652.653 111368 647.054 101434 643.655 106679 640.16 112144 632.47 107616 632.358 102676 629.859 107217 629.510 104966 62911 109491 628.1512 109864 62813 100656 627.614 106783 627.115 109637 626.5516 108046 623.7517 103050 623.318 104307 622.4519 110554 621.5520 106317 621.3521 106219 621.322 105080 619.323 101788 618.1524 104835 618.125 110751 617.8526 103659 617.2527 103054 616.0528 102417 61629 107697 61630 107533 615.731 107351 615.332 106384 615.2533 108464 614.734 110867 614.735 104400 613.9536 102667 612.737 109752 612.4538 108062 612.339 106843 612.1540 110984 611.9541 109520 611.142 101452 610.6543 104478 610.6544 101347 610.4545 107589 610.446 101476 610.147 100209 610.0548 100270 609.9549 108079 609.5550 101275 608.75I have to look up the team amounts for these next. By 18 roster theory, the majority of these teams should be 18 player rosters.

I'm at 529.05 -QG

 

[BassNBrew]

Philip Rivers $19 27.40 32.70 32.95

Carson Palmer $16 29.15 8.25 14.55

Chad Henne $14 9.00 11.70 29.25

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

Michael Turner $31 5.40 7.50 19.50

Ryan Mathews $29 8.20 6.50 0.00

Ronnie Brown $19 15.50 8.80 6.30

Leon Washington $8 1.20 1.70 0.00

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

Reggie Wayne $28 22.90 22.60 10.50

Terrell Owens $19 12.30 8.70 8.20

Early Doucet $6 5.70 0.00 0.00

Ted Ginn $3 2.90 0.00 0.00

Patrick Crayton $3 0.00 2.50 8.70

Troy Williamson $1 0.00 0.00 0.00

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

Jermichael Finley $21 10.70 16.30 25.00

Chris Cooley $13 17.00 16.90 12.80

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

Neil Rackers $4 13.00 17.00 8.00

Dan Carpenter $3 10.00 2.00 16.00

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

Miami Dolphins $5 5.00 17.00 0.00

New Orleans Saints $4 3.00 10.00 4.00

Carolina Panthers $4 11.00 0.00 5.00

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

TOTAL 145.45 150.00 144.95

CUTOFF 125.90 141.75 145.00

20 players. 3 QB and $14 on Henne hurt this team. Finley getting cramps was a killer. Went too cheap at WR and Mathews injury hurt.

 

[BassNBrew]

Drew Brees $28 17.25 25.40 34.25

Philip Rivers $19 27.40 32.70 32.95

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

Shonn Greene $27 3.20 6.50 3.60

LeSean McCoy $26 16.70 32.80 7.10

Clinton Portis $16 6.90 15.30 6.30

Arian Foster $13 42.30 15.30 14.60

Thomas Jones $7 3.90 9.30 16.10

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

Marques Colston $25 11.20 11.70 6.50

Chad Ochocinco $24 33.90 8.40 7.40

Nate Burleson $11 2.90 1.40 0.00

Kevin Walter $7 10.90 31.40 12.40

Anthony Gonzalez $7 2.20 0.00 0.00

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

Chris Cooley $13 17.00 16.90 12.80

Greg Olsen $11 9.70 11.40 19.90

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

Adam Vinatieri $4 6.00 9.00 11.00

Jeff Reed $3 14.00 15.00 8.00

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

Tennessee Titans $4 8.00 6.00 10.00

Washington Redskins $4 9.00 7.00 5.00

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

TOTAL 192.10 186.50 144.95

CUTOFF 125.90 141.75 145.00

18 players. $53 in stud RB production basically worthless this week, but the weak WR corps did them in.

 

[BassNBrew]

Drew Brees $28 17.25 25.40 34.25

Brett Favre $18 13.55 8.25 13.85

Alex Smith $15 9.25 20.55 16.60

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

Michael Turner $31 5.40 7.50 19.50

C.J. Spiller $17 3.40 4.60 11.40

Arian Foster $13 42.30 15.30 14.60

LaDainian Tomlinson $12 8.80 12.20 16.30

Tashard Choice $4 2.60 1.00 0.40

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

Marques Colston $25 11.20 11.70 6.50

Pierre Garcon $12 7.30 2.10 0.00

Mohamed Massaquoi $11 12.60 1.90 0.00

Jacoby Jones $10 4.90 17.30 10.10

Bernard Berrian $7 1.30 4.40 3.10

Darrius Heyward-Bey $7 2.10 14.00 7.90

Brandon Tate $6 7.60 2.70 6.80

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

Chris Cooley $13 17.00 16.90 12.80

Tony Scheffler $7 13.30 2.00 22.50

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

Jeff Reed $3 14.00 15.00 8.00

Phil Dawson $2 2.00 2.00 6.00

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

San Francisco 49ers $5 3.00 2.00 2.00

Washington Redskins $4 9.00 7.00 5.00

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

TOTAL 153.05 142.30 144.95

CUTOFF 125.90 141.75 145.00

21 players. Coin spent on 3rd QB for Favre/Smith spent at WR would have pushed this team through.

 

[BassNBrew]

Peyton Manning $28 39.65 30.45 34.15

Joe Flacco $17 11.60 10.60 30.70

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

Chris Johnson $40 29.00 7.80 24.50

Ray Rice $37 7.20 13.70 11.60

Arian Foster $13 42.30 15.30 14.60

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

Andre Johnson $32 6.30 33.80 10.40

Derrick Mason $17 4.10 10.10 7.00

Kenny Britt $13 0.00 9.10 10.60

Jerricho Cotchery $11 3.80 13.20 4.00

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

Jason Witten $19 7.20 12.60 16.10

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

Adam Vinatieri $4 6.00 9.00 11.00

Ryan Succop $2 3.00 11.00 8.00

Sebastian Janikowski $2 9.00 13.00 14.00

Olindo Mare $2 8.00 2.00 9.00

Phil Dawson $2 2.00 2.00 6.00

Graham Gano $2 9.00 11.00 10.00

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

Baltimore Ravens $6 4.00 1.00 2.00

Denver Broncos $3 1.00 9.00 0.00

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

TOTAL 152.55 160.25 144.95

CUTOFF 125.90 141.75 145.00

18 players and 6 kickers / 4 WRs. Hopefully more teams like this are hanging around.

 

[WisWolvrns]

No time to look at the rest of the 50, but the top 10 already showed only two 18 player rosters out of the 10. Much lower percentage, especially with the higher percentage of 18 player rosters built, than what I would have figured.

So if you want to win, it doesn't look like the theory of 18 player rosters being better holds weight. Just with using weeks 1-3 as weeks 14-16, the winner would have been a 27 roster team. With the best 18 roster team's being just 4th and 8th.

(barring any mistakes in my programming... )

 

[BassNBrew]

Matt Schaub $23 10.05 42.05 16.05

Philip Rivers $19 27.40 32.70 32.95

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

Ryan Mathews $29 8.20 6.50 0.00

Knowshon Moreno $24 12.90 19.80 0.00

Matt Forte $20 35.60 15.10 5.30

Arian Foster $13 42.30 15.30 14.60

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

Greg Jennings $26 19.20 6.60 9.80

Hakeem Nicks $24 29.50 11.80 12.60

Malcom Floyd $17 7.80 18.50 21.70

Kenny Britt $13 0.00 9.10 10.60

Jason Avant $4 8.10 6.30 0.00

Ted Ginn $3 2.90 0.00 0.00

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

Chris Cooley $13 17.00 16.90 12.80

Brandon Pettigrew $4 2.10 21.30 13.40

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

David Akers $5 10.00 5.00 4.00

Rob Bironas $3 10.00 3.00 11.00

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

Dallas Cowboys $5 1.00 7.00 10.00

San Francisco 49ers $5 3.00 2.00 2.00

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

TOTAL 205.00 166.75 144.95

CUTOFF 125.90 141.75 145.00

A solid 18 player team that got killed by injury at RB. This team could have done some damage.

 

[BassNBrew]

No time to look at the rest of the 50, but the top 10 already showed only two 18 player rosters out of the 10. Much lower percentage, especially with the higher percentage of 18 player rosters built, than what I would have figured.So if you want to win, it doesn't look like the theory of 18 player rosters being better holds weight. Just with using weeks 1-3 as weeks 14-16, the winner would have been a 27 roster team. With the best 18 roster team's being just 4th and 8th.(barring any mistakes in my programming... )

Let me thumb through my book of 18 man roster arguments and see what I can find.....there they are.Your sample size is too small.Studs are better condition athletes and produced better later in the year.

 

[WisWolvrns]

Studs are better condition athletes and produced better later in the year.

That was a good one!

 

[WisWolvrns]

The top 3 spots were non-18 roster teams.

But in the top 25 there were:

6 - 18 rosters

3 - 19 rosters

4 - 20 rosters

2 - 21 rosters

1 - 22 roster

2 - 23 rosters

0 - 24 rosters

1 - 25 roster

1 - 26 roster

1 - 27 roster

1 - 28 roster

1 - 29 roster

2 - 30 rosters

With all the discussions of large or small rosters, this breakdown seems to follow more the distribution of rosters at the various roster sizes than any strategy. Guess it might just be luck after all. You're roster size doesn't matter so long as you pick the right players.

Albeit if a larger roster allows you to survive weeks 1-13 better, then that strategy still seems better.

 

[Jayrod]

Players who seem to have really made an impact on the 3 week score totals:

Austin Collie

Arian Foster

Jahvid Best

Phillip Rivers

Note that none of these players were high dollar, but in the mid to lower tiers. Rivers was somewhat expensive, but was $10 below the top QB.

 

[QuizGuy66]

The top 3 spots were non-18 roster teams.But in the top 25 there were:6 - 18 rosters - 24%3 - 19 rosters - 12%4 - 20 rosters - 16%2 - 21 rosters - 8%1 - 22 roster - 4%2 - 23 rosters - 8%0 - 24 rosters - 0%1 - 25 roster - 4%1 - 26 roster - 4%1 - 27 roster - 4%1 - 28 roster - 4%1 - 29 roster - 4%2 - 30 rosters - 8%With all the discussions of large or small rosters, this breakdown seems to follow more the distribution of rosters at the various roster sizes than any strategy. Guess it might just be luck after all. You're roster size doesn't matter so long as you pick the right players.Albeit if a larger roster allows you to survive weeks 1-13 better, then that strategy still seems better.

That is way disproportionate in the direction of the bigger rosters.40% of the top 25 have 22 or more guys. Only 26.5% of the entries at the start have 22 or more guys.28% of the top 25 have 25 or more guys. They made up only 11.9% of the entries at the start.-QG

 

[WisWolvrns]

I stand corrected on the stats analysis. sorry. was just going from what I recalled. But on your more exact comparision, you are right.

 

[Sister Havana]

Wowwww. I just squeaked by this week. If history is any guide, I will be out next week. (Out in week 3 last year, out in week 2 the year before.)

 

[jdoggydogg]

Is it worth spending more for defenses?? Noticed that the NYJ def was a popoular choice for $9 when I just get 3 low priced defenses. This week Seattle came up big with 26 pts and the NYj def has yet to break double figures.....

I don't think so. I spread around $13 on three defenses: San Francisco 49ers $5New York Giants $5Seattle Seahawks $3I don't regret using three defenses, but I do regret spending so much. The 49ers have one of the worst defenses in the league so far. I'm sure Kansas City was about $2, and I bet they're a top 5 D so far.

 

[jdoggydogg]

One thing I find interesting that no one really discusses is the distribution of the number of teams entered at each roster size:

There was some discussion about this a few weeks ago, starting here.I still think it is safe to assume that quite a few of the 18-player rosters were "phoned in" - just filled out very quickly without much thought. It's just a fact that if you ask 10,000 people to perform a task, there will be a certain percentage of those people who will put in the absolute minimum amount of effort required, even if there is a reward involved. I'm sure there are plenty of 18-player rosters that were put together with a lot of thought too, but the others are the reason why there is such a large number of them.

I think these are good points. I don't picture a casual participant to fill out a 30 man roster. It's a lot less work to spread your money around with 18 players. This isn't to say that all the small rosters were submitted by amateurs. I am certain that some very serious contestants believed in the small roster theory. But I think you guys are on to something.

 

[CalBear]

I think these are good points. I don't picture a casual participant to fill out a 30 man roster.

21-Kicker-Guy is clearly not a casual participant!

 

[The Iguana]

Assuming you avoid injuries and get decent production from your QBs, you are definately better off with only 2 QBs in most instances. Top producing QBs are relatively consistant as a pair in best ball and expensive. When you start carrying more than two, you are tieing up alot of dollars that you usually won't be using. This coming from a guy who took three this year.

This is one of the interesting debates. Personally I could never justify having 2 top QBs on my team. Just too much expense. i went with 3 middle to low priced guys and one flier. Too bad the flier wasn't Vick. I'd rather have Ryan and Sanchez (my main 2 QBs) for the same price as a Rodgers or Manning. And then I still have the same money you/they would spend on a 2nd QB to get a 3rd and/or 4th. Of course, I think Ryan will score about as many points as those guys will but only cost me $16 in stead of $28/29.

 

[The Iguana]

Top 25 with Total Points Week 1-3

Rank	Entry	Total Pts	Roster1	105904	655.75	272	109740	652.65	223	111368	647.05	234	101434	643.65	185	106679	640.1	286	112144	632.4	237	107616	632.35	218	102676	629.85	189	107217	629.5	2910	104966	629	2611	109491	628.15	2012	109864	628	1913	100656	627.6	3014	106783	627.1	1915	109637	626.55	2516	108046	623.75	1917	103050	623.3	1818	104307	622.45	2019	110554	621.55	1820	106317	621.35	1821	106219	621.3	30  <----- that's me. 22	105080	619.3	2023	101788	618.15	2024	104835	618.1	1825	110751	617.85	21

can we bottle this and save it for the last 3 weeks? pretty sure I'm peaking a little too soon.

 

[Maurile Tremblay]

I do think that one of these "stud" teams probably would be the favorite over many of the "value" teams once in the top 250. However, I'm not sure this advantage doesn't disappear when you consider that it's not a head to head thing.

If we accept the theory that smaller teams are higher in variance, the opposite should be true.If a small team and a large team are evenly matched in a head-to-head competition, but we suddenly enlarge the competition to 100 teams, the smaller team should be more likely to finish in the bottom ten, but also more likely to finish in the top ten.

 

[Maurile Tremblay]

Not sure if these numbers are possible to get, but since we just completed 3 weeks, can a 3 week tally be done using weeks 1-3, just as if they were weeks 14-16? It would be a good test to compare who the top combined total teams were for weeks 1-3 with respect to amount of players on their teams.

Yes. Excellent idea.

The 18 player rosters should be a large majority if their theory is a small roster will have a better shot at winning if they survive the first 13 weeks, right? Also the first 3 weeks have no NFL teams on bye and the most players not injured yet, so less of those effects would be present, which should make the 18 team rosters even more powerful by their logic.

I disagree with that part. If smaller teams have more variance (and are thus more likely to finish at the top), the possibility of injuries should accentuate that fact. (There are no byes in weeks 14-16, so that part doesn't matter.) With respect to injuries, under the theory that you have to get lucky and avoid injuries in order to win, avoiding injuries should be worth more to a small team than to a big team. So getting lucky (which is a necessary component of winning) should give smaller teams a bigger boost than bigger teams.Nonetheless, I do think the weeks 1-3 experiment is rather instructive.

 

[Modog814]

No time to look at the rest of the 50, but the top 10 already showed only two 18 player rosters out of the 10. Much lower percentage, especially with the higher percentage of 18 player rosters built, than what I would have figured.So if you want to win, it doesn't look like the theory of 18 player rosters being better holds weight. Just with using weeks 1-3 as weeks 14-16, the winner would have been a 27 roster team. With the best 18 roster team's being just 4th and 8th.(barring any mistakes in my programming... )

Let me thumb through my book of 18 man roster arguments and see what I can find.....there they are.Your sample size is too small.Studs are better condition athletes and produced better later in the year.

Question for you guys who are bashing the "18 man" rosters....are you strictly against literally 18 man rosters? Are you okay with shorter rosters, but with more players say 20 or 21?I count 15 of the top 25 in the list below have a roster of 21 or less. Only 5 have 27 or more. That to me seems that if you're okay with using this as data (which I do think the sample size is too small, it's 1 three week period, we should keep this going with rolling 3 week periods) it shows that the theory of 18 players rosters, which I don't take literal I take them at meaning shorter rosters, does in fact hold weight.

 

[Maurile Tremblay]

Also, I was considering my 24-player team as being small, but it occurs to me that it may not be. It seems small to me (because it's on the small side of the roster sizes I was seriously considering), but within the 18-30 range, I guess it qualifies as medium.

 

[Modog814]

I do think that one of these "stud" teams probably would be the favorite over many of the "value" teams once in the top 250. However, I'm not sure this advantage doesn't disappear when you consider that it's not a head to head thing.

If we accept the theory that smaller teams are higher in variance, the opposite should be true.If a small team and a large team are evenly matched in a head-to-head competition, but we suddenly enlarge the competition to 100 teams, the smaller team should be more likely to finish in the bottom ten, but also more likely to finish in the top ten.

I'm not seeing where your example disagrees with what I said.

I'm saying that if a "stud" team has a head to head advantage against any single "value" team (which is the assumption we're assuming the smaller roster teams are making), they will be favored to win in a head to head matchup, this is obvious because it is the assumption that they have an advantage.

However, when a "stud" team is going against 100 "value" teams, the assumed advantage shrinks (or disappears or tilts in the other direction) due to the fact that the value teams have most players covered, so if a number of "non-stud" players perform well over a 3 week stretch, it's likely that one of these 100 values teams has that combination of players to outscore the "studs".

Then again, this is getting into the conversation of groups vs. individual teams. This analysis might suggest that it's more likely that any higher roster team is going to win, but the "stud" team might have the highest individual probability of winning. And in a contest like this, if I'm in the top 250, I don't care if it's more likely that a 30 player team is going to win, what I care about is that I have the highest probability of winning. The bolded words are the most important in that last statement.

 

[Modog814]

Also, I was considering my 24-player team as being small, but it occurs to me that it may not be. It seems small to me (because it's on the small side of the roster sizes I was seriously considering), but within the 18-30 range, I guess it qualifies as medium.

This I think is what the problem is with this discussion, we have loose terms that mean one thing to one person and something else to others. What is a stud? Is he a top 5 dollar cost player at each position? Is it any player where you spent more than x% of your total budget on? What's a small roster? What's a large roster? When people refer to 18 player rosters are they taking that literally? Or is the phrase "18 player roster" more general (meaning say 20 players or less)?

 

[CalBear]

Question for you guys who are bashing the "18 man" rosters....are you strictly against literally 18 man rosters? Are you okay with shorter rosters, but with more players say 20 or 21?I count 15 of the top 25 in the list below have a roster of 21 or less. Only 5 have 27 or more. That to me seems that if you're okay with using this as data (which I do think the sample size is too small, it's 1 three week period, we should keep this going with rolling 3 week periods) it shows that the theory of 18 players rosters, which I don't take literal I take them at meaning shorter rosters, does in fact hold weight.

74% of contest entries have 21 or fewer players; 60% of the top 25 has 21 or fewer players. So again, the smaller rosters are succeeding at a lower rate than the higher rosters.

 

[rzrback77]

Not that it means anything at all but I also find it interesting that after 3 weeks of scores, of my 30 man roster, I've used scores from exactly 18 of my players to survive to week 4. QBs - 3 of 4RBs - 4 of 7WRs - 5 of 10TEs - 2 of 3Ks - 2 of 3Ds - 2 of 3Flex - 2 RBs and 1 TE.

I have a 30 man roser as well - my scoring distribution for comparison, but it is also 18QBs - 2 of 3RBs - 4 of 6WRs - 5 of 11TEs - 3 of 3Ks - 1 of 3Ds - 3 of 4Flex - 2 TEs and 1 WR.Wanted to add that my non-counting RBs have three double digit scores, my WRs eight, and my TEs two. The only scoring position thus far in the single digits has been DST and that is something that will not happen on smaller roster teams as often.

 

[BassNBrew]

No time to look at the rest of the 50, but the top 10 already showed only two 18 player rosters out of the 10. Much lower percentage, especially with the higher percentage of 18 player rosters built, than what I would have figured.So if you want to win, it doesn't look like the theory of 18 player rosters being better holds weight. Just with using weeks 1-3 as weeks 14-16, the winner would have been a 27 roster team. With the best 18 roster team's being just 4th and 8th.(barring any mistakes in my programming... )

Let me thumb through my book of 18 man roster arguments and see what I can find.....there they are.Your sample size is too small.Studs are better condition athletes and produced better later in the year.

Question for you guys who are bashing the "18 man" rosters....are you strictly against literally 18 man rosters? Are you okay with shorter rosters, but with more players say 20 or 21?I count 15 of the top 25 in the list below have a roster of 21 or less. Only 5 have 27 or more. That to me seems that if you're okay with using this as data (which I do think the sample size is too small, it's 1 three week period, we should keep this going with rolling 3 week periods) it shows that the theory of 18 players rosters, which I don't take literal I take them at meaning shorter rosters, does in fact hold weight.

My "bashing" is based on the fact that every statistic from last year and this year point towards small rosters being the inferior route to advance and win this contest. An 18 player roster might very well win, but it will be bucking every trend and stat posted thus far. 24 crushed 20 and 21 last year. With the change in values that might change. I personally don't think 30 is the best. I suspect we'll see a sweet spot as the contest progresses.56% of the entires in this contest started with 18/19 man rosters. They aren't holding their weight in the first rolling 3 wk period.

 

[LittlePhatty]

Question for you guys who are bashing the "18 man" rosters....are you strictly against literally 18 man rosters? Are you okay with shorter rosters, but with more players say 20 or 21?

I'd say yes, they have a better shot. But I still think the more the better. 19 is better than 18.22 is better than 21.This brings me back to an earlier point - if you went with the EXACT minimum of 18 players, and you chose precisely the "studs" that you REALLY wanted, what are the odds that those guys added up to $250? If it didn't, you would either add a guy or two (thus it isn't 18 anymore) or you would bump up/down an RB or WR to a different price (which means you are choosing based more on price as opposed to the players you really think are worth it).So yeah, I'm specifically calling out the 18 rosters. 19 to a much lesser extent.

 

[Modog814]

Question for you guys who are bashing the "18 man" rosters....are you strictly against literally 18 man rosters? Are you okay with shorter rosters, but with more players say 20 or 21?I count 15 of the top 25 in the list below have a roster of 21 or less. Only 5 have 27 or more. That to me seems that if you're okay with using this as data (which I do think the sample size is too small, it's 1 three week period, we should keep this going with rolling 3 week periods) it shows that the theory of 18 players rosters, which I don't take literal I take them at meaning shorter rosters, does in fact hold weight.

74% of contest entries have 21 or fewer players; 60% of the top 25 has 21 or fewer players. So again, the smaller rosters are succeeding at a lower rate than the higher rosters.

69% of LIVE entries have 21 or fewer, (i'm assuming that the top 3 weeks scores isn't counting the people that got eliminated in weeks 1 and 2, but I could be wrong). They make up 60% of top 25, and 64% of top 50. Since it's only a single week, I wouldn't consider that a significant difference. 7.6% of LIVE entries have 27 or greater, They make up 20% of top 25, but only 10% of top 50.

 

[BassNBrew]

I do think that one of these "stud" teams probably would be the favorite over many of the "value" teams once in the top 250. However, I'm not sure this advantage doesn't disappear when you consider that it's not a head to head thing.

If we accept the theory that smaller teams are higher in variance, the opposite should be true.If a small team and a large team are evenly matched in a head-to-head competition, but we suddenly enlarge the competition to 100 teams, the smaller team should be more likely to finish in the bottom ten, but also more likely to finish in the top ten.

I'm not seeing where your example disagrees with what I said.

I'm saying that if a "stud" team has a head to head advantage against any single "value" team (which is the assumption we're assuming the smaller roster teams are making), they will be favored to win in a head to head matchup, this is obvious because it is the assumption that they have an advantage.

However, when a "stud" team is going against 100 "value" teams, the assumed advantage shrinks (or disappears or tilts in the other direction) due to the fact that the value teams have most players covered, so if a number of "non-stud" players perform well over a 3 week stretch, it's likely that one of these 100 values teams has that combination of players to outscore the "studs".

Then again, this is getting into the conversation of groups vs. individual teams. This analysis might suggest that it's more likely that any higher roster team is going to win, but the "stud" team might have the highest individual probability of winning. And in a contest like this, if I'm in the top 250, I don't care if it's more likely that a 30 player team is going to win, what I care about is that I have the highest probability of winning. The bolded words are the most important in that last statement.

All teams will have some studs. An 18 man roster could have dropped a $30 stud (keeping it's other studs) and add 10 players averaging $3 leaving it with 27 players (hence a large roster). There's a good chance that as we move on, there will be a larger roster that has a good percentage of the smaller rosters stud players covered.Probably the best thing that could happen for the smaller rosters is for Foster to get injured after the bye weeks.

 

[Modog814]

Question for you guys who are bashing the "18 man" rosters....are you strictly against literally 18 man rosters? Are you okay with shorter rosters, but with more players say 20 or 21?

I'd say yes, they have a better shot. But I still think the more the better. 19 is better than 18.

22 is better than 21.

This brings me back to an earlier point - if you went with the EXACT minimum of 18 players, and you chose precisely the "studs" that you REALLY wanted, what are the odds that those guys added up to $250? If it didn't, you would either add a guy or two (thus it isn't 18 anymore) or you would bump up/down an RB or WR to a different price (which means you are choosing based more on price as opposed to the players you really think are worth it).

So yeah, I'm specifically calling out the 18 rosters. 19 to a much lesser extent.

This is what I figured, thanks for the clarification.

By the way, the third option of selecting 18 players, is the selection of say 15 players they really wanted that sum to less than $250 and then just fill in the last 3 with whatever $$$ they have available.

 

[BassNBrew]

Question for you guys who are bashing the "18 man" rosters....are you strictly against literally 18 man rosters? Are you okay with shorter rosters, but with more players say 20 or 21?I count 15 of the top 25 in the list below have a roster of 21 or less. Only 5 have 27 or more. That to me seems that if you're okay with using this as data (which I do think the sample size is too small, it's 1 three week period, we should keep this going with rolling 3 week periods) it shows that the theory of 18 players rosters, which I don't take literal I take them at meaning shorter rosters, does in fact hold weight.

74% of contest entries have 21 or fewer players; 60% of the top 25 has 21 or fewer players. So again, the smaller rosters are succeeding at a lower rate than the higher rosters.

69% of LIVE entries have 21 or fewer, (i'm assuming that the top 3 weeks scores isn't counting the people that got eliminated in weeks 1 and 2, but I could be wrong). They make up 60% of top 25, and 64% of top 50. Since it's only a single week, I wouldn't consider that a significant difference. 7.6% of LIVE entries have 27 or greater, They make up 20% of top 25, but only 10% of top 50.

I know it's early, but this small subset is an argument that the 27+ roster guys are the ones who configured their to win the big prize, not just compete. I suspect that will change.

 

[Ignoratio Elenchi]

I count 15 of the top 25 in the list below have a roster of 21 or less. Only 5 have 27 or more. That to me seems that if you're okay with using this as data (which I do think the sample size is too small, it's 1 three week period, we should keep this going with rolling 3 week periods) it shows that the theory of 18 players rosters, which I don't take literal I take them at meaning shorter rosters, does in fact hold weight.

You're missing something obvious here. There are waaay more small rosters than large rosters, so they should be represented more in the top 25, and in fact the small rosters are underrepresented. 18-man rosters made up 41.5% of all the entries yet only make up 24% of the early season top-25. OTOH, 30-man rosters made up just 2.1% of all the entries yet make up 8% of the early season top-25. More generally, 84.5% of the entries were "smaller" (less than 24 players) but these only represent 72% of the top 25.Any way you slice it, the smaller rosters are taking a beating so far. The bye weeks will only make this more evident. Of course, we still don't know if these small rosters will realize some kind of an advantage in weeks 14-16 (they didn't last year, IIRC - but, I know, I know, this year's pricing is different, etc.)

I think very few of the very small rosters will make the final 250, and I'm still not buying the idea that those who do will have a better chance of winning it all when they get there. If some 18-man roster correctly chose the "right" studs to keep them alive until the final 250, they'll probably find themselves competing against some 20-, 22-, 24-man rosters who have the most, if not all, of the same studs and better depth.

 

[domvin]

Joe Flacco $17 11.60 10.60 30.70

Ben Roethlisberger $14 0.00 0.00 0.00

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

Ahmad Bradshaw $18 16.30 8.90 20.30

C.J. Spiller $17 3.40 4.60 11.40

Tim Hightower $13 17.40 19.40 4.50

Arian Foster $13 42.30 15.30 14.60

Tashard Choice $4 2.60 1.00 0.40

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

Andre Johnson $32 6.30 33.80 10.40

Calvin Johnson $27 8.50 15.00 11.80

Dez Bryant $13 13.60 7.20 9.00

Pierre Garcon $12 7.30 2.10 0.00

Dexter McCluster $7 2.90 0.30 15.80

Josh Cribbs $7 4.65 16.50 12.80

Brian Hartline $6 0.00 11.80 13.40

Greg Camarillo $3 3.90 1.30 3.60

Patrick Crayton $3 0.00 2.50 8.70

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

Chris Cooley $13 17.00 16.90 12.80

Visanthe Shiancoe $12 19.60 17.60 2.20

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

Mason Crosby $6 14.00 12.00 6.00

Lawrence Tynes $4 8.00 2.00 7.00

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

San Francisco 49ers $5 3.00 2.00 2.00

New England Patriots $4 17.00 3.00 5.00

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

TOTAL 168.30 160.10 144.20

CUTOFF 125.90 141.75 145.00

I thought I had a shot with this team, but no dice

Probably too much tied up in my QBs, and not enough value found at the TE position. RBS stayed healthy, but that was probably going to catch up to me.

Kicking myself, bc I had Lance Moore on the list, and then took him off.

 

[Stinkin Ref]

big roster guy here (29)...have used 18-29....

One of the things I am interested in seeing is....the "don't get guys with late bye week" strategy.....with the increase in roster size, I went with the opposite.....I focused around a couple of guys I was high on with late bye weeks(Rivers/Orton...then added Stafford) thinking they could carry me through the first 8 weeks....I could see getting guys with late bye weeks as a good strategy to help get you through the first few weeks of bye weeks that may weed out the small roster guys......and then if you can somehow clear weeks 8-10 with your quantity over quality you might have enough to hang on.....the late bye weeks may help you get through, but I guess it could end up costing you if you lay an egg from a late bye week stud....

I would think most of the small roster guys spread out their bye weeks to help eliminate the risk.......so while a bunch may get knocked out, the ones that picked the right guys and make it through will be tough down the stretch....

 

[BassNBrew]

Week 1 trim to 12,000 teams (all advance if we have less entries) 91.9% advance

Week 2 trim to 10,000 teams 83.3%

Week 3 trim to 8,500 teams 85%

Week 4 trim to 7,250 teams 85.3% Very good week to have some bye players

Week 5 trim to 6,100 teams 84.1%

Week 6 trim to 5,000 teams 82%

Week 7 trim to 4,000 teams 80%

Week 8 trim to 3,100 teams 77.5%

Week 9 trim to 2,300 teams 74.2%

Week 10 trim to 1,600 teams 69.6% Not so good week to have some bye players

Week 11 trim to 1,000 teams 62.5%

Week 12 trim to 500 teams 50%

Week 13 trim to 250 teams 50%