theoneandonly
Footballguy
If D.Will can come back and start beasting the rest of the way i may have a chance. D.Will Bradshaw Spiller L.Washington and Foster at rb i really striked out there.
Fantasy Football - Footballguys Forums
This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!
Subscriber Contest (2 Viewers)
- Thread starter eaglezzz
- Start date
Ignoratio Elenchi
Footballguy
Not bad... I didn't save the work I did last week, not sure what I would've had for week 10 predictions. This will be fun, though. Here are my early week projection for number of surviving entries by roster size:Modog814 said:
Code:
Size Survive18 18919 11120 9321 7522 7723 8724 6925 7426 5727 5128 3329 2930 66If you're only considering the size of the roster, then I'd say a team with a 30-player roster definitely has the best chance of surviving and making it to the final 250.Given that it survives to the final 250 teams, I'd somewhat less certain that a 30-player roster team has the best chance of winning the contest. But I "think" it would have the best chance of winning.From right now (at the end of week 10), I'd say that a 30-man team definitely has the best chance of surviving and winning the contest.A different question -- What is the over/under on roster size for the winning team? I think it's about 26.5. If so, I'd take the over.
Wide Right
Footballguy
With less that 4 points encompassing the whole spread for average points - it looks like there is no statistical difference between the average score and the roster size. I read this as any size roster has an equal chance to win.Ignoratio Elenchi said:
I did the same. Down to 3 unused players:Robiskie - He is such a non factor I don't see him contributing, but I guess anything is possible. I felt the same way about Chambers until this week.Gano - He has had decent weeks, but I have 4 kickers. Maybe 4 was 1 too many.Bernard Scott - I have Benson too, so this was a handcuff.Wide Right said:
Saint
Footballguy
Thanks Iggy! I appreciate the information!
Now what's interesting to me is the survival rate that goes with them...... as I recall the survival rate is higher amongst the 25+ rosters. With those 2 attributes combined, I would say that statistically, you have a greater chance of making the cuts with a larger roster... at which point the rosters normalize.So for example, would you want to be the team that has a 6% chance of making the playoffs and putting up ~170 pts... or the team that has a 9% chance of making the playoffs and putting up ~170 pts. Statistically speaking, I'd take 9% over 6%, which is the 25+ man roster.
Marauder
Footballguy
I'm not sure average score is a good indicator though. The average teams, even this late, are not likely to make the money. I would think it's better to look at the outliers in a contest like this, or maybe figure out the standard deviation by roster size. The higher standard deviation should be better, no?
Ignoratio Elenchi
Footballguy
We looked at this a week or two ago IIRC. The highest score on any given week tended to be a smaller roster; however, the highest cumulative score over a rolling three week period tended to be larger. This is my gripe with the "smaller rosters are more likely to win due to variance" argument - we don't hand out the money after a one week playoff, we hand it out after a three week playoff, which mitigates whatever variance-related advantage small rosters might have imo.
Maik Jeaunz
Footballguy
One thing I think that's being ignored when looking at average scores by roster size and survival rates so far by roster size, is how byes and injuries affect the smaller roster size teams. One of the key advantages of larger roster sizes is the protection it affords against byes and injuries. However, after this week, byes are no longer a factor, and injuries become less a concern with each passing week. So, while larger rosters seem to have some statistical advantage in terms of survival through bye weeks and early season injuries, that advantage won't be as significant going forward. I'm not certain how much less impact, but certainly some, and the statistical advantage wasn't that significant to begin with.
So, larger roster sizes may have helped to get you through week 10, but it won't necessarily help as much to move forward from here, if that makes sense.
So, larger roster sizes may have helped to get you through week 10, but it won't necessarily help as much to move forward from here, if that makes sense.
Marauder
Footballguy
I get that (BTW, I'm firmly a large roster guy) but average score of all the remaining entries still just doesn't feel right as a predictor of success in the final 250.
Ignoratio Elenchi
Footballguy
It's likely that small rosters are more adversely affected by bye weeks than larger rosters, but to say the statistical advantage wasn't that significant is silly. The advantage of larger rosters during the bye weeks is unbelievably and undeniably enormous. The numbers are astounding.In any case, the larger rosters did better last year after the byes, and they did better in weeks 1-3 this season, when there were no byes and few, if any, meaningful injuries. It's more likely that small rosters are worse than large rosters during the bye weeks, because small rosters are worse than large rosters in general. It's not like larger rosters don't have byes, too - in fact, they have more of them to deal with. The same reason that large rosters do better during weeks 4-10 should also apply during weeks 1-3 and 11-16 - they have more available players.
Yes, they have more available players, but generally those players aren't as good. I can't deny the facts, and the facts suggest larger rosters, on average, are better than smaller rosters. But, I'm not trying to have an average team.My 19 man roster has made it this far, and I like my chances as much as any other team I've seen. I'd obviously change some of my players in retrospect, but I wouldn't necessarily change my roster size. I wasn't intentionally trying to have a small roster, but when I picked players I thought would outperform their price and who I thought had the potential to carry a team, I ended up with 19 players. My biggest regret is swapping out Rivers for Rodgers on my last submission. I could have had a better QB for $10 dollars less, and then used that money to upgrader someone like Berrian to Dez Bryant (which was one of my earlier options).It's likely that small rosters are more adversely affected by bye weeks than larger rosters, but to say the statistical advantage wasn't that significant is silly. The advantage of larger rosters during the bye weeks is unbelievably and undeniably enormous. The numbers are astounding.In any case, the larger rosters did better last year after the byes, and they did better in weeks 1-3 this season, when there were no byes and few, if any, meaningful injuries. It's more likely that small rosters are worse than large rosters during the bye weeks, because small rosters are worse than large rosters in general. It's not like larger rosters don't have byes, too - in fact, they have more of them to deal with. The same reason that large rosters do better during weeks 4-10 should also apply during weeks 1-3 and 11-16 - they have more available players.
The other team lost in week 1. 
del mar
Footballguy
I will start with the disclosure that my roster started with only 19 and is down to 18 as Clayton is on IR. I count myself as a convert to the larger roster reasoning, but with some qualification. Today I looked at how many survivors have each of my players. It occurred to me that Ray Rice is the most important man on my team and Foster is the least important player as virtually everyone owns him. This led me to thinking about what leads to success in this contest.
I looked at Iggy's earlier post regarding the team which would have won if the playoffs were weeks 8-10. It became clear that simply having 25 men on the roster was not the key to success. That team would have used only 16 men over the three week span (which included significant byes). What made that team successful was hitting home runs at the major posistions and some luck. He had Foster, Peterson, Dez Bryant and Tampa Mike who all counted in each week. That represents one-half of the 24 scores required from the major positions. Calvin Johnson was used twice. With the exception of Peterson and Johnson he had great success for moderate prices. He did not spend much at QB or TE (a total of $57 for 3 players at each position.) Remarkably he did not use two of those six players and they cost $14 and $15. So he got excellent results from some very targeted choices. Finally, he got quite lucky with his choices at K and DEF. Three of those six scores were the highest numbers posted by those players in the 10 week span.
What can we learn for this (he said somewhat presumptuously)?
First, more of this success is due to hitting the mark rather than to tossing stuff against the wall and waiting to see what sticks. A lot of people (myself included) found Tampa Mike and Foster. This guy also spent his money on Peterson (versus my Rice) and Bryant (versus my Gaffney). To win you can not have high priced disappointments. To win you alsos need to have moderately priced home runs. Sheer volume is not the key. Nine of his 25 players didn't count. I'll give him a pass on Ryan Grant being injured, but he did cover that with B Jackson who did count in one of the three weeks. Call it the purchase of the GB RB for $31.
Second, you need to take some low priced chances which have the potential to payoff in a large way. Lower priced TEs are probably the greatest investment. No one of them will equal Gates, but given the 1.5 PPR and the 6 points for TDs they can be huge. Along the sam lines Vick for $3 was, in retrospect, a home run waiting to happen. Past achievement may be a key to future potential. Some guys like Grant, Clayton and Floyd may be bargains next year based on stats which are depressed by injury time. I will personally be waiting for a $10 Wes Welker one year removed from his knee surgery.
Third, the question of diversity may be over analyzed. This winning team didn't win because he had one player who was totally unique to his team. His team probably was not very unique at all. He won because he had a unique combination of players. This involved some who were must haves, some who were studs who paid off, and some who were lucky guesses who simply paid off big at the right time.
Finally, at the risk of being redundant - Past achievement may be a key to future potential. Rookies have the potential to be full of potential. Tampa Mike is an exception, not the rule. Mathews, Best, and Spiller are dying primarily, not from lack of opportunity, but from lack of proven ability at this level and maybe also from lack of good blocking team mates.
I will revise these thoughts, but not until I have some real reason to do so. Thanks to Iggy for shining the light on the path.
I looked at Iggy's earlier post regarding the team which would have won if the playoffs were weeks 8-10. It became clear that simply having 25 men on the roster was not the key to success. That team would have used only 16 men over the three week span (which included significant byes). What made that team successful was hitting home runs at the major posistions and some luck. He had Foster, Peterson, Dez Bryant and Tampa Mike who all counted in each week. That represents one-half of the 24 scores required from the major positions. Calvin Johnson was used twice. With the exception of Peterson and Johnson he had great success for moderate prices. He did not spend much at QB or TE (a total of $57 for 3 players at each position.) Remarkably he did not use two of those six players and they cost $14 and $15. So he got excellent results from some very targeted choices. Finally, he got quite lucky with his choices at K and DEF. Three of those six scores were the highest numbers posted by those players in the 10 week span.
What can we learn for this (he said somewhat presumptuously)?
First, more of this success is due to hitting the mark rather than to tossing stuff against the wall and waiting to see what sticks. A lot of people (myself included) found Tampa Mike and Foster. This guy also spent his money on Peterson (versus my Rice) and Bryant (versus my Gaffney). To win you can not have high priced disappointments. To win you alsos need to have moderately priced home runs. Sheer volume is not the key. Nine of his 25 players didn't count. I'll give him a pass on Ryan Grant being injured, but he did cover that with B Jackson who did count in one of the three weeks. Call it the purchase of the GB RB for $31.
Second, you need to take some low priced chances which have the potential to payoff in a large way. Lower priced TEs are probably the greatest investment. No one of them will equal Gates, but given the 1.5 PPR and the 6 points for TDs they can be huge. Along the sam lines Vick for $3 was, in retrospect, a home run waiting to happen. Past achievement may be a key to future potential. Some guys like Grant, Clayton and Floyd may be bargains next year based on stats which are depressed by injury time. I will personally be waiting for a $10 Wes Welker one year removed from his knee surgery.
Third, the question of diversity may be over analyzed. This winning team didn't win because he had one player who was totally unique to his team. His team probably was not very unique at all. He won because he had a unique combination of players. This involved some who were must haves, some who were studs who paid off, and some who were lucky guesses who simply paid off big at the right time.
Finally, at the risk of being redundant - Past achievement may be a key to future potential. Rookies have the potential to be full of potential. Tampa Mike is an exception, not the rule. Mathews, Best, and Spiller are dying primarily, not from lack of opportunity, but from lack of proven ability at this level and maybe also from lack of good blocking team mates.
I will revise these thoughts, but not until I have some real reason to do so. Thanks to Iggy for shining the light on the path.
LittlePhatty
Footballguy
We all know that end-of-season results never line up perfectly with pre-season rankings, right? I mean even weekly rankings are usually way off after the games have been played. Of course higher ranked players have a better chance of performing well. But every year there are late season studs that appear out of nowhere.
To be #1 you have to hit on as many of those studs as possible in the last 3 weeks. You have to take a chance on some cheaper guys too and hope you hit.
Thought experiment:
Let's say this contest was limited to only the first 250 entries. Those lucky enough to get in pick players using exactly the same method as this year's contest, during the pre-season. One big change though: weeks 1 - 13 don't matter at all. Everyone waits around and then totals are added up in weeks 14 - 16. Aside from not worrying about bye weeks, would that change your roster size strategy?
I'd still lean to the high side. 27 - 30 players probably. I'd want as many players as possible as long as I had a good feeling about every one of them. As soon as the season kicks off, those pre-season rankings don't mean squat. And you are picking these guys 14 weeks in advance. Who is currently leading the NFL in receiving yards? Brandon Lloyd. He wasn't even "on the radar" enough to be included in this contest.
To be #1 you have to hit on as many of those studs as possible in the last 3 weeks. You have to take a chance on some cheaper guys too and hope you hit.
Thought experiment:
Let's say this contest was limited to only the first 250 entries. Those lucky enough to get in pick players using exactly the same method as this year's contest, during the pre-season. One big change though: weeks 1 - 13 don't matter at all. Everyone waits around and then totals are added up in weeks 14 - 16. Aside from not worrying about bye weeks, would that change your roster size strategy?
I'd still lean to the high side. 27 - 30 players probably. I'd want as many players as possible as long as I had a good feeling about every one of them. As soon as the season kicks off, those pre-season rankings don't mean squat. And you are picking these guys 14 weeks in advance. Who is currently leading the NFL in receiving yards? Brandon Lloyd. He wasn't even "on the radar" enough to be included in this contest.
74 teams out of 1600
az_prof
Footballguy
This makes sense to me.
QuizGuy66
Footballguy
You should do this with entries 100001 through 100250 (of course I'm one of those
)-QG
apalmer
Footballguy
The link to the Querier is on the Weekly Content Page.
Come on Reg...My predicted breakdown for the top 1000 (1003 really since its based on simulations making it through)
Code:
Size Predicted Current Pred Rate18 194 333 58.26%19 106 186 56.99%20 96 152 63.16%21 78 122 63.93%22 81 122 66.39%23 77 131 58.78%24 68 106 64.15%25 81 107 75.70%26 55 84 65.48%27 40 74 54.05%28 34 49 69.39%29 23 42 54.76%30 70 93 75.27%
Last edited by a moderator:
massraider
Footballguy
You know where it'll help, and has helped probably every week for most large roster teams? At TE, K, and D. And come the final weeks, there will be teams that don't make the cut, and their K or their D will be the difference.az_prof said:
Last edited by a moderator:
These are the 28 teams that averaged more than 600 points over the last three weeks (Weeks 14-16, according to my sim), all 28 are in the 250 most likely to make the top 250.
Code:
108122 631.41 20
105076 630.98 22
100086 623.3 30
103802 623.24 30
100205 615.24 28
110146 612.79 22
108226 612.3 19
107775 611.63 30
102584 610.06 25
100189 608.94 23
107528 608.4 25
100492 608.11 25
105017 607.74 20
102003 606.4 18
101483 606.36 22
101754 605.75 21
108900 604.91 27
101170 604.1 26
100257 603.12 26
102206 602.98 18
105145 602.25 22
100634 602.23 24
111098 601.95 29
101771 601.15 28
103970 600.26 23
108047 600.2 18
107622 600.17 21
100244 600.08 21
Last edited by a moderator:
370th Most Likely to get into Top 250. 3 week average (14-16) of 552.
A crude attempt at combining likelihood of getting into the top 250 and avg 3 weeks score in a power rankings, (Score = the rank of getting into top 250 + rank of Avg 3 week score, so lower is better)
Code:
1 105076 7 22
2 110146 8 22
3 108122 10 20
4 100086 11 30
5 100205 16 28
6 101754 17 21
7 107775 30 30
8 100189 31 23
9 100492 32 25
10 102003 32 18
11 108047 33 18
12 103802 34 30
13 108757 35 22
14 102584 36 25
15 107622 39 21
16 102992 43 22
17 108226 43 19
18 105017 51 20
19 101771 52 28
20 100634 56 24
21 100679 57 21
22 101483 64 22
23 103970 66 23
24 107044 66 30
25 107528 66 25
26 101087 69 22
27 111098 70 29
28 103771 74 18
29 109213 77 19
30 100244 79 21
31 100257 80 26
32 101690 82 29
33 101170 87 26
34 104745 105 26
35 102206 107 18
36 100927 109 22
37 109798 109 23
38 100201 118 18
39 105620 118 18
40 100433 120 23
41 107908 121 29
42 108703 123 18
43 107241 127 19
44 107036 132 21
45 104832 133 18
46 109773 134 22
47 101473 136 24
48 103276 136 24
49 101539 139 24
50 109578 139 28Thanks for the vote of confidence. Should be a fun next few weeks. Good Luck!
QuizGuy66
Footballguy
Go DC Yourself
Footballguy
Are you sure the high scorers the last three weeks weren't just teams that had most of their byes weeks 4 to 6?
Care to explain why you think this?
TheChairman
Footballguy
People with heavy byes weeks 8,9,10 will be underrepresented. So your last 3 weeks score will be flawed.
Ignoratio Elenchi
Footballguy
I think by "last week weeks" he meant weeks 14-16, not weeks 8-10. Hey Modog, I'm pretty sure I asked this before but I don't remember what the answer was - are you actually running a bunch of iterations of a sim, using distributions of player scores (like Doug does) or are you just plugging in a set of projections for those weeks?
Oh, I get what he's saying. Iggy is correct, by last 3 weeks, I meant weeks 14-16, not 8-10.
The last time you asked the answer was plugging in a set of projections for those weeks.This time the answer is, running a bunch of iterations of a sim. Each week is simmed 250 times (play by play, game by game) with opponent, performance to date and play calling/situational tendencies factored in. For the road to the top 250, I combined each of the 3 weeks 250 scenarios into 187,500 "paths". For example, path 1 would be Wk11 Scn.1, Wk12 Scn.1, Wk13 Scn1....path 2 would be Wk11 Scn.1, Wk12 Scn.1, Wk13 Scn2, so on. Then counted up how many times a team survived all three weeks.For the average final 3 week score (wks 14-16), I again simmed each week 250 times, but I just kept all the paths aligned, so Path 1 = Wk14 Sn1, Wk15 Sn1, Wk16 Sn1...Path 2 = Wk14 Sn2, Wk15 Sn2, Wk16 Sn2 So there's only 250 total "paths". I could do more, but I'm not sure how much the results were change or if it's worth the additional computation time being that these predictions are 6 weeks out.
Ignoratio Elenchi
Footballguy
Is that some kind of service you subscribe to?
Nope, I use a simulator I created to do this.
rock753
Footballguy
bad news for 82 owners on D'Angelo Williams
could it help the 6 owners with Stewart? Possibly (Full disclosure, I own neither)
I'm still amazed by the fact that over 1/4 of the field has Finley (424 entries)
14 have both Finley & Williams
23 have Romo
115 have Mark Clayton
1 has Donnie Avery
1 has Ramses Barden
1 has Antonio Bryant
69 have Jake Delhomme
11 have Trent Edwards
9 have Kevin Faulk
29 - Anthony Gonzalez
14 - Stephen Gostowksi
31 - Ryan Grant
332 - Jason Hanson
2 - Montario Hardesty
6 - Steve Hauscuka
62 - Chad Henne
1 - Torry Holt
60 - Larry Johnson
16 - Matt Moore
56 - Mike Nugent
64 - Jeff Reed
38 - Alex Smith
252 - Matthew Stafford (including me)
2 - Ben Tate
18 - Pierre Thomas
there's most of the dead weight, I'm sure I missed a few......
could it help the 6 owners with Stewart? Possibly (Full disclosure, I own neither)
I'm still amazed by the fact that over 1/4 of the field has Finley (424 entries)
14 have both Finley & Williams
23 have Romo
115 have Mark Clayton
1 has Donnie Avery
1 has Ramses Barden
1 has Antonio Bryant
69 have Jake Delhomme
11 have Trent Edwards
9 have Kevin Faulk
29 - Anthony Gonzalez
14 - Stephen Gostowksi
31 - Ryan Grant
332 - Jason Hanson
2 - Montario Hardesty
6 - Steve Hauscuka
62 - Chad Henne
1 - Torry Holt
60 - Larry Johnson
16 - Matt Moore
56 - Mike Nugent
64 - Jeff Reed
38 - Alex Smith
252 - Matthew Stafford (including me)
2 - Ben Tate
18 - Pierre Thomas
there's most of the dead weight, I'm sure I missed a few......
⚡DEADHEAD⚡
Footballguy
I have Finley but I picked three TEs. In this scoring system, I don't see you can't have at least three TEs on your roster. Next year, I may go four.
TheChairman
Footballguy
Wow, #1 in sim only has 93.2 survival %
TheChairman
Footballguy
I don't think it's been posted, but the link works.
QuizGuy66
Footballguy
Updated.-QGGoing into week 10, these are the entries with the most players left at each position.
6 QB Guy OUTThere's 13 5-QB guys.
10 RB Guy #1 STILL IN
10 RB Guy #2 STILL IN
10 RB Guy #3 STILL IN
10 RB Guy #4 OUT
18 WR Guy!!!Side note - this entry has had the most WRs the entire way.
STILL IN!!!!!!!!!!!!!!!!!!!!!!
7 TE Guy STILL IN
8 PK Guy.OUT There are 3 7-PK guys left.
7 Defense Guy
So there you go
-QG
