2009 NFL Survivor Elimination Pool - Week 4 Pick Strategy

Entry Name	Week 1	Week 2	Week 3	Week 4 Recommendation
Tony/Cody	BAL	WAS	PHI	SF
Johnny	IND	WAS	BAL	SF
Ryan	GB	WAS	BAL	SF
Anonymous	NO	WAS	BAL	SF

My week 4 recommendation for everyone is to use San Franciso this week (They are playing at home against St. Louis. SF, according to the updated team scoring system posted yesterday, is the 17th best team and St. Louis is the 31st. The SF 49ers only other predicted blowout on the schedule is in week 16 when they host the Detroit Lions. But Detroit (despite recently losing about 20 games in a row over 3 calendar years) is actually a much better team right now than St. Louis, so there is no better week to use SF than right now in week 4.

Many of you will be wondering about the other options this week: Chicago (at home against Detroit), Indianapolis (at home against Seattle), and NY Giants (on road against KC) are all good options as well. However, using the updated team valuations from yesterday's post. The difference in scores between SF and St. Louis (after taking home field advanatage into account) is larger than the disparity in any of the other games mentioned.

If you are considering going with either the Giants, Bears, or Colts instead of the SF 49ers, here are some more things to consider:

I would save NY Giants for week 5 when they are home against the Oakland Raiders. The Raiders are almost just as bad as KC, so better to wait for week 5 to use the Giants when they are at home instead of this week on the road.

The model is saying to save the Bears for week 8 when they host the Cleveland Browns. The Browns are the worst team in the NFL by far. So the Bears are more likely to win in week 8 than this week against the Lions.

The model is saying to save the Colts (4th best team) for week 7 when they are on the road aginst St. Louis (31st best team). In week 7, they will be even more likely to win than they are this week against the Seahawks (18th best team).

Questions/Comments?

2009 NFL Survivor Elimination Pool - Updated Team Rankings

A few of the comments have questioned my model because it was based on ESPN's power rankings. For the first few weeks of the season, I think it was a good idea to base the optimization model on the power rankings because what else do you have to go by? It just doesn't make sense to use last year's team statistics or 2009 preseason statistics. So I went with the power rankings which told me that going into week 1 the defending Super Bowl champs Pittsburgh were the best team and the 0-16 Lions were the worst, which made a lot of sense at the time.

However, there are several areas in which using the power rankings leaves something to be desired. For example, what if two teams (such as Tampa Bay and St. Louis) are equally good or bad? Power rankings force one of them to be ranked better than the other. Also, rankings force teams to be given incremental ranks from 1 to 32, but what if the best team is twice as good as the 2nd best team? Power rankings force the model to believe that the best team is as equally likely to beat the 2nd best team as the 31st team is to beat the 32nd team.

So, 3 weeks into the season, I think it's time to incorporate some 2009 NFL team statistics into the model. So before calculating the week 4 picks, I've made a few enhancements to the model. Each team's score (which is the essential input to the model which gets updated every week) will now be calculated based on three metrics:

1. Yards gained minus yards allowed per game (Yard Differential)
2. Point differential per game
3. ESPN power rankings

I began with yard differential / game which is the yards gained per game minus the yards allowed per game. Denver (160 more yards gained than allowed per game) is the best and Cleveland (195 more yards allowed than gained per game) is the worst team at yard differential so far this season. But yard differential / game isn't everything. I wanted to incorporate turnovers and red zone offensive and defensive proficiency which is where the point per game differential comes into play. The Saints are the best team so far in point differential per game, averaging 21 more points scored than allowed through week 3.

Finally, if I would have counted point differential/game and yard differential per game equally, the Saints would have come out as the best team going into week 4. This is why the power rankings are still incorporated into the model, b/c they provide a nice grounding (The saints are now power ranked 4th btw)

So these three metrics are all indexed meaning each of the 3 metrics is re-scaled so that the best team gets a score of 100 and the worst team a score of 0. A team's final score will be their average score across the 3 metrics. Here is how the final score was calculated for all 32 teams (remember, these scores will be re-calcualted every week):

Team	Yard Differential	Point Differential	ESPN Power Rank	Yard Index	Points Index	Power Rank Index	Final Score
BAL	147.6	16.6	1	96.6	89.1	100.0	95.3
NO	125.7	21.3	4	90.4	100.0	90.3	93.6
NYG	146.4	10.7	3	96.3	75.5	93.5	88.4
IND	90.7	9	2	80.6	71.6	96.8	83.0
DEN	159.6	15.4	13	100.0	86.4	61.3	82.6
MIN	57.6	10.3	5	71.2	74.6	87.1	77.6
NE	132.7	3.3	7	92.4	58.4	80.6	77.2
PHI	121.3	7.3	9	89.2	67.7	74.2	77.0
NYJ	59.0	10.3	6	71.6	74.6	83.9	76.7
SD	60.0	3	12	71.9	57.7	64.5	64.7
DAL	47.0	8.4	17	68.2	70.2	48.4	62.3
GB	-22.7	6	10	48.5	64.7	71.0	61.4
CHI	21.7	1	14	61.1	53.1	58.1	57.4
PIT	56.7	-1	16	71.0	48.5	51.6	57.0
CIN	-29.0	1.6	11	46.8	54.5	67.7	56.3
ATL	-78.3	1.3	8	32.9	53.8	77.4	54.7
SF	-43.7	4.6	15	42.6	61.4	54.8	53.0
SEA	43.6	3	23	67.3	57.7	29.0	51.3
TEN	16.4	-4.4	18	59.6	40.6	45.2	48.5
ARI	-25.0	-3.7	19	47.9	42.3	41.9	44.0
WAS	15.6	-3	26	59.4	43.9	19.4	40.9
JAC	-49.0	-3	22	41.1	43.9	32.3	39.1
BUF	-71.7	-2.7	21	34.7	44.6	35.5	38.3
MIA	-13.7	-8.7	24	51.1	30.7	25.8	35.9
HOU	-103.0	-7	20	25.9	34.6	38.7	33.1
DET	-97.7	-9	28	27.4	30.0	12.9	23.4
CAR	-69.0	-16.7	25	35.5	12.2	22.6	23.4
OAK	-143.0	-7	27	14.6	34.6	16.1	21.8
KC	-98.0	-12.3	30	27.3	22.4	6.5	18.7
TB	-136.0	-16.6	31	16.6	12.5	3.2	10.8
STL	-127.3	-16.3	32	19.0	13.2	0.0	10.7
CLE	-194.7	-22	29	0.0	0.0	9.7	3.2

So these are the week 4 scores which will go into the model. Let me know your comments. I will share the model's week 4 picks sometime tomorrow.

2009 NFL Survivor Elimination Pool - Week 4 update

All our week 3 picks won easily. Week 4 picks will be coming up either tomorrow (Tuesday) or Wednesday. If anyone else is alive and has a different set of winners than the ones listed below, then let me know which winners you have used and I'll let you know what the model says for your week 4 pick as well.

Entry Name	Week 1 Winner	Week 2 Winner	Week 3 Winner
Tony/Cody	Baltimore	Washington	Philadelphia
Johnny	Indianapolis	Washington	Baltimore
Ryan	Green Bay	Washington	Baltimore

2009 NFL Survivor Elimination Pool - Week 3 Pick Strategy

Thanks to comments and emails from several of you, we have 3 entries still alive and in need of a week 3 pick and strategy for all the remaining weeks. Here is a summary of our 3 entries:

Entry Name	Week 1 Winner	Week 2 Winner
Tony/Cody	Baltimore	Washington
Johnny	Indianapolis	Washington
Ryan	Green Bay	Washington

In week 3, I recommend that the Johnny and Ryan entries go with Baltimore for their winner. Following a win on the road at San Diego, the Ravens are now power ranked #1 and facing the 0-2 Cleveland Browns in week 3 who are ranked #29. The Browns are the second worst team the Ravens will face between weeks 3 and 17. The worst is during week 14 when the Detroit Lions (rank #31) play Baltimore in Baltimore. Although my model predicts that that week 14 game will be even more of a blowout than this week 3 affair, the week 14 matchup is down-weighted for being so far into the future, and the model says to use Baltimore now in week 3.

With Baltimore not available, the Tony Cody entry has a tougher decision in week 3. For week 3 for this entry the model is going with the Philadelphia Eagles (rank #9) who host the Kansas City Chiefs (rank #30). The Eagles did just get blownout at home in week 2. However, if you're going to win your pool, you can't use one of the very best (top-5) teams every single week, and you're going to have to use the 9th best team at some point. In the case of the Eagles, the KC Cheifs are the worst team remaining on their entire schedule and the Eagles remain at home in week 3. If you wanted to avoid the Eagles, the only other predicted blowout winner you could consider using is the NY Giants who are on the road at Tampa Bay. However, I would definitely recommend saving the Giants until week 5 when they host the lowly (even worse than the Tampa Bay Bucs) Oakland Raiders at the Meadowlands. Also, you really need to save the New York Giants for week 5, because my model is telling me that their really aren't any other good options in week 5. Eagles stock may be low heading into week 3, but to give yourself the best chance at winning your entire pool I'd go with the Eagles in week 3.

2009 NFL Survivor Elimination Pool - Week 2 Update

My week 2 winner choice, Green Bay, did not win, so I am out. If anyone out there has advanced to week 3 and enjoys reading my weekly pick post, comment to this post with who your week 1 and week 2 correctly picked winners were and I will input them into my model and continue sharing picks for week 3 using your set of available winners to choose from. If I get no replies, then I guess it will be time for me to start thinking about fantasy basketball draft strategy.

NFL survivor elimination pool debate - Part 2

Continued from part 1...

If you go along with the other 4 competitors and use team A as your pick, then a 5 way tie for 1st place is guaranteed and you will get 20% of the $100 prize pool.

Can picking one of the other teams, with less of a chance to win their game actually lead to higher expected winnings? Let's find out.

If you pick team B instead of team A, then you will still finish in a 5 way tie if both teams win or both teams lose. As stated in part 1, team A has an 80% chance of winning and team B has a 70% chance of winning. The probability that both teams win is 80% * 70% = 56% and the probability that both teams lose is 20% * 30% = 6%. So 56% + 6% = 62% of the time you will still finish in a tie and win 20% of the pot even though you picked a different team.

If team A wins and team B loses, you, of course, will win nothing. This will happen 80% * 30% = 24% of the time.

But, in the event that team B wins and team A loses, you will get 100% of the prize pool all to yourself. This will happen 20% * 70% = 14% of the time.

So to summarize, 62% of the time you will get 20% of the prize, 24% of the time you will win nothing, and 14% of the time you will win 100% of the prize pool. To calculate the total expected winnings, you take the sum product of the probabilities of each case and the winnings from each case: 62%*$20 + 24%*$0 + 14% * $100 = $26.4. $26.4 is more than $20 so picking team B, which actually had a lower chance of winning than team A, is the best strategy here.

Case	Probability	Winnings	Expected Winnings
A & B win	80% * 70% = 56%	$20	$11.2
A & B lose	20% * 30% = 6%	$20	$1.2
A wins & B loses	80% * 30% = 24%	$0	$0
B wins & A loses	70% * 20% = 14%	$100	$14
Total	100%	-	$26.4

Before the start of the weekend, the distribution of picks across all 32 possible winners is published by sites such as Yahoo! How can this information be used to increase your expected earnings similarly to in this example? Let's tackle that in another post. To be continued...

NFL survivor elimination pool debate - Part 1

Let's look 15 weeks into the future and imagine that you have survived the first 16 weeks of the NFL season and are contemplating your week 17 pick. Also imagine that along with you, 4 other competitors from your original pool of 100 competitors have survived and that the 5 of you have the same set of 16 teams available to choose from for week 17. Using the Las Vegas moneylines, you determine each of these 16 teams' percent chance of winning in week 17. One of the available 16 teams (Team A) has an 80% chance of winning, one team (team B) has a 70% chance of winning, and the 14 other teams are all extreme underdogs with less than 20% chance of winning. Which team should you pick, team A, team B, or one of the 14 underdogs? Let's assume that:

1. $100 will be awarded to the person who survives the most weeks
2. If there is a tie, the prize will be split amongst those in the tie
3. All 4 of your competitors will be picking team A

To be continued in the next post.

Thanks to Bill D. for asking me a question that was the inspiration for this post.

2009 NFL Survivor Elimination Pool - Week 2 Pick Strategy

My week one pick Indianapolis defeated the jacksonville Jaguars on Sunday, meaning I have advanced to week 2 of this year's 2009 NFL survivor pool. Here is my strategy for this week's pick:

My week 2 pick is the Green Bay Packers to defeat the Cincinnati Bengals (in Green Bay). A Green Bay victory is actually one of 5 projected blowouts on the NFL week 2 schedule. I also expect 3 other home teams (Atlanta over Carolina, Tennesee over Houston, and Washington over St. Louis) and one road team (Minnesota over Detroit) to also win in blowout fashion along with Green Bay. However, Green Bay is a better choice for a week 2 winner than those other 4 options because:

1. Green Bay has 2 other projected blowout wins on the calendar. Week 3 on the road against St. Louis and week 6 at home against Detroit. However, in week 3, the Eagles are an even more likely winner at home against Kansas City. And in week 6, the Pittsburgh Steelers are an ever more likely winner at home against the Cleveland Browns. So we can use Green Bay now because we're planning to use Philadelphia and Pittsburgh those other weeks.

2. Not going to pick Minnesota on the road this week against Detroit because we're saving Minnesota for week 10 when they play Detroit at home (and are even more likely to win).

3. Not going to use Atlanta at home this week against Carolina because we're saving Atlanta for week 12 when they are home against Tampa Bay. The updated (week 2) ESPN power rankings say that Carolina is a much better team than Tampa Bay, so Atlanta will be even more likely to win in week 12 than this coming week.

4. Not going with Tennesee over Houston this week because we're planning to save Tennessee for week 14 when they are playing at home against St. Louis. St. Louis is a much worse team than Houston (at least according to the ESPN power rankings).

5. The last blowout option I am passing on is Washington to beat St. Louis in Washington. Unlike with the other 3 blowouts we are passing on, in the case of Washington, this is their most likely week to win as a favorite in their entire remaining schedule. However, using my formula based on the ESPN power rankings, the differential between Green Bay and Cincinnati in the coming week is greater than the differential between Washington and St. Louis.

Notes: I developed an Excel model to optimize winner picks for the entire NFL schedule so that the sum of the differential of all the games used is the largest. On Tuesday of each week, I update my Excel model with the latest ESPN power rankings, which are also updated on Tuesdays. From week to week, as the power rankings fluctuate, the optimal set of winners the model says to use for all remaining games will change. That is why I am not publishing my picks now for every single remaining week. I will just take the latest information, and use it to come up with the best possible pick for this week, but always keep the entire NFL schedule in mind. For more on the philosophy / algorithm / approach / methodology of the Excel picking model, please see the post announcing my winner pick from week 1.

Fantasy Sports Insurance

An article on CNN.com today was about the booming fantasy football industry and how different companies are cashing in. One example, is FSI FantasySportsInsurance.com. Their motto is "Armor for your fantasy investment". You can buy an insurance policy such that the company will reimburse your league entry fees if your key player(s) gets injured. On their website, You pick how many players you want to insure (1 to 3), which specific players you are insuring, how many games they need to miss combined in order for your policy to pay, and finally how much $$$ your policy will payout if your players do miss that many games. After selecting these 4 inputs, the website then gives you a quote and you can continue onward to enter your billing information if you wish to purchase the policy.

My first thought is that this has nothing to do with insurance... FSI does not administer fantasy sports league, they cannot confirm that you are a fantasy owner of a specific player, and they cannot confirm how much (if any) your league entry fees actually are. You don't even need a fantasy sports team to buy one of these insurance policies.

I just went to FSI's website, and typed in some inputs so that i was offered a policy that would cost me $12.88 and payout $100 if Tom Brady misses 10 games this season. This seems like gambling / sports betting to me. How is this legal? FSI is based out of New York state. You don't even have to be 21 to buy a policy (make a bet)

My second thought is that no one should buy any policies from them. FSI is likely in business to make money. This implies that if people buy policies from them, in the long run, FSI will make money (not the people buying policies). So why should anyone buy a policy if their expected return on that policy is to lose money?

But people buy insurance policies all the time, such as extended warranties on their washing machine, travel insurance when going on vacation, jewelry insurance on engagement rings, and health insurance. So why do people buy insurance if in the long run it makes insurance companies money and costs them money? Well, it's important to buy policies to insure against things you cannot afford to have happen, like getting very sick or becoming disabled and not being able to work anymore. Your washing machine breaking and having to buy a new one may or may not fall into the category of events you cannot afford to have happen. However, I really hope that for everyone playing fantasy sports, not winning their league and losing their entry fee (if there was one) does not fall into the category of catastrophic events you cannot afford to have happen. If it does, you have a gambling problem.

2009 NFL Survivor Pool - Week 1 Pick

I'll assume you are familiar with the basic rules of an NFL survivior/elimination pool. If not, read here: http://en.wikipedia.org/wiki/Survivor_fantasy_leagues or here: http://football.fantasysports.yahoo.com/survival

For my posts on survivor football, let's assume the ESPN power rankings (which are updated weekly) are the ultimate source of how good a team is and that the absolute value of the difference between two team's power rankings is perfectly correlated to how likely the favorite is to actually win the game.

Yahoo has a nice feature that shows you the distribution of everyone else's pick for a given week. For week 1 of the NFL season, 40% of America is picking New Orleans (currently power ranked #19) to beat Detroit (currently power ranked #32). That's a power ranking difference of 13. And also, for my system we will penalize the visiting team 5 ranks. So the adjusted difference becomes 18.

So is this the most lopsided matchup on the week 1 schedule? Nope. Jacksonville @ Indinapolis, KC @ Baltimore, Washington @ NY Giants, Buffalo @ New England, and SD @ Oakland all have power rank differentials (adjusting for home field advantage) between 19 and 25.

Jacksonville (rank 25) @ Indianapolis (rank 5) has the largest adjusted differential of 25. If all of this power rank differential stuff was accurate and was, in fact, the best possible prediction system, then you might be thinking that Indianapolis is the best survivor pool choice for week 1. But there's more to it than that.

Say in week 2, Indianapolis once again is predicted to be the most lopsided winner with a week 2 differential of 26 points. And also imagine that in week 2 all of the other 15 games on the NFL schedule are very close matchups with power ranking differentials all less than 5.

Now thinking back to week 1 again, in this scenario should you pick Indianapolis in week 1 (and leave yourself with no good options for week 2) or go with one of the other good options for week 1 and save Indianapolis for week 2. I think everyone will agree that in this case the latter strategy is better.

So now you know that you should look ahead when making your picks. --But you also don't want to look ahead to far. For example, what's the point of saving a great pick for week 17 if the entire surivivor pool is likely to end by week 12? For my model, I used a "decay factor" that down weights the power rank differential of all games in future weeks proportionally to the week of the game on the schedule (There will be more on my model in a later post).

Anyway, for my week 1 pick I am going with Indianapolis to win. It turns out that in this case, there is no better week so save Indianapolis for and now is the best week to use them. Yahoo is showing me that Indianapolis is the week 1 pick of 1.03% of Americans. I am guessing I am the only one of them not living in Indiana.