Drafting for Value

If you’re like me, the day of your league’s fantasy football draft trumps Christmas morning.  It is the day that you get to join with close friends in fellowship due to your common interest.  It is the day you can wear a crappy team jersey, shit talk, and drink excessively, all without having to deal with girlfriends and wives.  And oh yeah, it is also the day that, after 4 months of tedious research and preparing (don’t judge me), you finally get to build your fantasy team’s roster.

But how are you planning to maximize the value of all that preparation?  You’ve already done the hard work.  You’ve got your final player rankings and know exactly what guys you want and what guys you want to avoid.  That’s great, but in order to really build the best fantasy roster you can, you need to know your league as well as you know the running backs on your top 25 list.  While you may have a quarterback ranked as a top 5 player going into this season, what makes him even more valuable is also knowing the perfect spot in your draft in which to select him.

One of the most useful things you can do to prepare for a draft is to spend time mockdrafting from whatever spot you have in the draft order.  This is beneficial because immediately you will begin to recognize the approximate draft position (ADP) of each player based on the preferences of the general public.  With this knowledge you can begin maximizing the value of the players on your own individual rankings by finding the latest round possible to select those players in.  Just remember that the ADP you find online must be very current and be derived from a league with similar lineups and scoring as your own.

Now once you’ve done some mocks and have a good idea of what the ADP of your ranked players is, you can start applying this to your own league.  This really gets easier and easier the longer you’ve been involved with a given league.  In fact, you should view your draft as not just a time to select this season’s players, but also research for future year’s drafts.  For example, in my league of 10 friends there are 6 Chargers fans, 2 Raiders fans, a Packers fan, and a bandwagon fan.  There are 2 guys who like to have top-tier QBs every year and 1 who wants to dominate the TE market.  4 guys like the RBs ‘early and often’ strategy while 3 are okay with spending early picks on the top-tier WRs.  Also I know to keep a watchful eye on the 3 guys who seem the least prepared, because they’re the true wild cards and could really draft any player at any time.  So taking all of this knowledge into account, I can make educated guesses throughout the draft as to what I think these other guys will do.  And with that, I can really pinpoint which spot to select which players.

This may sound like a lot of strategy just for drafting a fantasy team, but this is just the beginning.  Once you’ve mastered the ability to gauge approximate draft position of fantasy players, you can actually begin to visualize exactly how you want your team to look before the draft even begins.  Based on your draft position you can start going back to the mockdraft sites and determine the optimum draft style based on your position and the ADP of fantasy players.  For example, say you’re picking 10th out of 10 in the first round of your draft (therefore also 1st in round 2).  Chances are you will be looking the best 7 or 8 running backs being off the board, along with perhaps the top wide receiver and/or quarterback.  Well after several mock drafts you’ll be able to know before even going into your real draft whether you’re better off getting a tier 2 RB or starting out with 2 of the top tier WRs, because you know what RBs are likely to come back to you at round 3 and on.

Another really strong tactic is to ‘draft backwards’.  Find players who you believe will be impact players this fantasy season but are not being drafted until the very last rounds or not at all.  If you can find a QB for example who you believe can be your starter this season that isn’t going until about round 11 or later, make it a point to take him in round 8 or 9 of your draft.  You will guarantee getting your starting QB and it frees you up in rounds 1-7 to not even need to worry about that position, but instead load up on other positions.

In review… take a break from ranking those fantasy players!  More valuable than knowing how they’ll do this season is knowing when to get them in your draft.  Study ADP, study your league mates.  Be able to put together a strong strategy for how to maximize the value you get with each selection because doing that will give you the best possible outcome on draft day.


A Look At The Math Behind Fantasy Football – Part 3

Before reading, don’t miss Part 1 and Part 2.

So you know all that stuff I just rambled about in the last post about consistency being irrelevant?  Well I’m going to go ahead and contradict myself a bit here.  Not saying that consistency is good, but, I’m toying with combining consistency with another variable to show something even more meaningful.  This other variable is a player’s skew about their mean.  In English, is a player more likely to produce fantasy scores that are above or below his weekly average?

Before I give us both a hard on with more exciting Excel Spreadsheets, lets just think about this intuitively for a moment.  Say you have 2 players, and both players score the exact same total points over the entire season, Say 160 points.  If they both play every game, that’s an average of 10.  Now, imagine that one player is so freaking consistent that he gets exactly 10 points every single week (ya right, but just play along k?).  The other guy’s scores are more chaotic; sometimes 5, sometimes 15, etc.  Now think about this: is this player likely to be above 10 and below 10 the exact same number of times?  Well its very possible, but there is also a very good chance it is something like 9 times above and 7 times below, or vice versa.  So if you’ve got these 2 players playing each other 1 on 1, both score the same points over the season, but one player beats his average more times than the other guy does, who do you think will win more games?

Ex-freakin-actly.  The player who has a more positive skew of production is desirable.

Here’s a visual of what we just talked about:

Note that both players have the same total points, and exact same weekly average.  Player 1 has a higher standard deviation and therefore could be labeled more inconsistent.  However, Player 1 is more desirable in every sense.  In individual matchups, Player 1 has a far superior total hand of ‘trump cards’, so to say, the majority of his scores will beat the majority of scores for Player 2, as seen when sorted below:

Now before I move on I’d just like to acknowledge that at this point in time I cannot think of any way to predict how a player’s future production skew will look, and so as of now this doesn’t have any value to your draft preparation.  But I find this interesting as shit and it is a concept I will continue to work with and research over the coming months.  However, when the season is winding down and you can start comparing 2 players’ performances over the last 10-12 weeks, it might be worth considering that the player with greater positive production skew may be more likely to continue that trend in future games.  (In reality, you’ll notice a high positive skew when a player has a real stinker or 2 in a couple games, which bring his weekly average down, but in general he scores much better than his average would indicate).

Besides choosing between 2 players to start or sit when the season is nearing its conclusion, what practical purpose does this have for us?  Well, I’m not sure.  But I believe it may be out there and so I’ll be spending a lot of time looking for it.  If anything, for now, I just wanted to go a step further in showing that going out of your way to choose a player who is more consistent can be detrimental to you on draft day.  I mean, Player 2 was the more consistent player in the above example, but he got pwnd by Player 1. Don’t draft Player 2!!

That should sum up my consistency theory bashing for a while.  In my next post I’m going to take a look at how understanding the tendencies of the guys you draft with is going to make you dominant on draft day.

A Look At The Math Behind Fantasy Football – Part 2

Before reading, don’t miss: Part 1.

So, you know that whole ‘Consistency is King’ mantra?  That one where all the ‘experts’ on fantasy sites and magazines talk about how players are better when they are consistent?   In my opinion this line of thinking currently plagues fantasy sports.  Around this time last year I began debating someone who was trying to show mathematically (using the standard deviation of a player’s historical week-to-week stats) which players are more desirable because of their relative consistency (lower standard deviation indicating greater consistency, in this case).  I thought about this for a while because his initial point made sense: if a player has lower standard deviation there are more weeks that you can expect his minimum point output to be greater than a player with higher standard deviation, week-to-week averages being equal.  (We’re not to the confusing part yet).  This is also ‘useful’ because it provides you with some confidence that your player won’t drop a real stinker in an upcoming game.

But, upon further study into this napkin analysis of consistency it is clear that we’d be ignoring the flip side to expected values: what is the MOST you could expect from a ‘consistent’ vs. ‘inconsistent’ player using their standard deviations as an indicator (averages being equal).  Obviously, you would expect the player with greater standard deviation to probably have the lowest AND highest weekly scores on the entire spectrum.

In short, this purpose of this post is to show mathematically that any value as derived from the consistency in a player’s week-to-week scoring average is all psychological; consistency provides no added value in terms of winning fantasy football games.

To demonstrate my position, I am going to illustrate with a test by pitting two teams against each other, ‘Team Consistent’ and ‘Team Boom/Bust’.  I’ve constructed the two teams so that they have exactly the same total scoring output over the entire season, the only difference is that the players on Team Consistent have a lower standard deviation in their week-to-week scoring than do the players making up Team Boom/Bust.  My intent is to show that consistency gains no edge when it comes to winning head-to-head games.

Some notes to the spreadsheet: each player in this spreadsheet is a real NFL player, and these are real fantasy stats taken from 2002-2009.  I have handpicked the players so that each position matches up closely in total point production with his counterpart on the other team, IE- the RB1 for Team Consistent and Team Boom/Bust will have nearly the same exact total yearly fantasy points, with main difference being their standard deviation in weekly scores.  The year end point total for each team is the exact same down to the tenth of a point.  Furthermore, only players that played in all 16 games that season were chosen.  The ‘Min’ and ‘Max’ at the last 2 columns are the minimum and maximum weekly fantasy scores that you would expect from the player by subtracting/adding their standard deviation from their weekly scoring average.

Test 1:

(Rosters: 1 quarterback, 2 running backs, 3 wide receivers, 1 tight end)

Test 2:

(Rosters: 2 quarterbacks, 2 running backs, 3 wide receivers, 1 tight end, 1 kicker, 1 defense)

As you can see I have performed the same test with 2 different roster requirements.  My intent with this is only to show that league settings do not play any role in skewing the results.  Results have Team Boom/Bust actually winning the season 9-7 in test 1, and in test 2 the teams tied 8-8.  I am actually surprised that either team ended up winning more games than the other in this test, but I attribute it to the fact that 16 games is simply not a large enough sample size to perform statistically relevant analysis.  In reality, as these teams continue to play head-to-head on into infinity each should win exactly 50% of the time.

So what do these numbers mean?  Well, its like I said before: scoring consistency from week to week provides no added value when it comes to winning games.  Consistency backers will argue with me because I am using hindsight to prove these results, and that in practice you need to have confidence that your players will perform above a minimum level each week.  I can only say that this too defensive an approach (playing not to lose).  If its the championship game then I want to play to win, which means that if anything I’m going to start the player who has the potential to go off for a huge game, not the player who is least likely to give me a pile of crap.  I mean you have to assume that the team you are playing is going to score points too, right?  But that is all post-draft talk involving who to start on a week-to-week basis.  In regards to valuing players for your fantasy draft, I don’t even recommend trying to ‘predict’ which players will be consistent and which won’t, there are just too many things that go into point production that make it impossible to predict such a thing.  Having said that, you will likely find many other guys in your own league who like consistent players, and therefore will ‘reach’ for them on draft day.  If you take anything from this post, take the knowledge that on draft day you can completely disregard a player’s tendency to be consistent or inconsistent.  Just let the other guys in your league worry about that stuff because you know that you can sit back and take the player who will score the most points for you that year, whether the points come in steadily or in jumbles doesn’t matter on draft day.

The next post in this series goes a step further by combining a player’s standard deviation with their tendency to perform above or below their weekly scoring average: Part 3.

A Look At The Math Behind Fantasy Football – Part 1

Fantasy Football season is finally back upon us, and that means I am back into Ultimate Nerd Mode until about mid January.  You might say “but Brian, every guy nowadays plays fantasy football, its not really that nerdy is it?”.  In actuality this might be the nerdiest blog post about one of the nerdiest hobbies ever.  In fact, if you find it not just educational but entertaining, you are one hell of a nerd.  Anyway what was I talking about.  Oh ya, this fantasy football thing appeals to me on multiple levels.  First, I absolutely love football, so in this sense I can justify in my head all the time that I waste doing fantasy football.  Second, I am really into things that require me to think strategically, or mathematically, or really anything intellectual that helps me win a competition because that is how I live with the fact I could never become a starting quarterback in the NFL.  I’m kind of like Uncle Rico but I was too lazy to film the videotapes.  But enough about me, lets talk numbers ya?

Okay.  So I’m going to organize this beast into a several part series on how math plays a role in arriving at a player’s value to your fantasy football team.  This first post is going to focus solely on the basic understanding and application, and I’ll then move onto a bit more confusing topics regarding scoring consistency and skew of production.

When it comes to applying any kind of formula to something in order to give it a value, the most important factor is your input variables.  In the coming posts I am going to be using historical statistics in order to show how various formulas/methods yield certain results.  Historical statistics’  input (a player’s fantasy point production) is 100% correct and therefore helps us prove the validity of a formula, but this is of course all theoretical stuff.  To actually make use of a formula you’d have to use future stat projections because obviously you are drafting players based on what you think they’ll do in the future, not the past.  Any finance person will tell you that in business school you can learn all sorts of fancy equations for arriving at the exact value of a company’s equity, but when it comes down to it these equations all rely on estimations of what you believe that company’s future profits will be (in the end, most day traders just look at trends in charts rather than waste their time projecting the company’s future cash inflows).

Projecting future statistical production is extremely difficult to do.  Indeed, this aspect of fantasy football is what separates the guys who make their league’s playoffs every year from the guys who don’t.  It is my belief that, playoff teams are more or less decided upon by how well each team drafted.  It is then the ability to make the right waiver pickups and weekly sits/starts, especially down the stretch, that make a playoff team a champion.  Long story short, your league draft is extremely important to your not sucking this season.  My next few posts will focus on the use of math in determining player value on draft day.  Again, a lot of this is going to be heavy on the theory and thus difficult to accurately apply in real life, but even if you just understand the concepts you might be able to avoid some damaging preconceptions in your analysis of a player’s value.  Alright, lets get to some number crunching!!

#musicmonday – N.A.S.A. Gifted (masuka remix)

I don’t personally enjoy anything relating to dubstep, but this beat that one of my friends posted on Facebook doesn’t suck:


#musicmonday – The Arcade Fire

This is a special Music Monday post because tomorrow is the release of The Arcade Fire’s new album, The Suburbs.  In preparation of the release, here are some awesome songs to get you through your day:

No Cars Go:

Wake Up:

Neighborhood #3 (Power Out):



What movie characters should always be-WOODERSON

Wooderson was brought to you by Matthew McConaughey.  He reached his acting zenith at the beginning which I respect because I can’t think of anyone else who has done that.  Ole’ Matt can be proud of his performance here, even though his acting choices went down hill after that.

This movie is amazing.  Wooderson has sooo many great quotes.  He makes pedophilia seem cool.

Just keep on livin. L I V I N


Bad Ass Movie Characters – pt 2

What happens when you take a really bad ass character and insert an amazing actor? Academy awards happen bitches. Our character in question is There Will Be Blood’s Daniel Plainview, played by Daniel Day-Lewis.  Realistically, we could give this guy the bad ass award based solely on his mustache and the end scene in which he drinks Paul Dano’s milkshake.

Here is a ‘stache close-up:


Now moving on to more important things, like actual acting.  There are few scenes in a movie that have you walking away with such a remembrance as does the final scene in TWBB.  Shockingly I cant find a better version of this clip on YouTube, but if its quality you are anal about than I don’t know why you’re still reading this blog.  Anyway, if you can watch this clip and somehow not find a soundbite of it to use as your ringtone for several months, you’ve already surpassed me:

“OMG WHY DIDN’T YOU SAY SPOILER ALERT!?!?” Fuck you! This movie came out 4 years ago and that wasn’t even the part where he killed him!  Oops.  Anyways if you really have been dumb enough not to see this movie until now do it in a hurry.  There are plenty more scenes as bad ass as this one.  Oh and the story and all that other necessary bullshit is real good too.

I’m Finished.


Bad Ass Movie Characters – pt 1

For no reason other than I have major man-love for bad ass movie characters, I am going to tell you about them.  Many times, it isn’t always the film’s main hero or villain who is most awesome.  And this is true in the case of The Departed.  Mark Wahlberg is in only a handful of scenes in this great Scorsese film, but you have to imagine that ole’ Martin had the most fun shooting Mark’s than any other scenes in the movie. (suck it Jack and Leo).  If you haven’t seen the movie, here is indisputable evidence for why Wahlberg’s Staff Sgt. Dignam is incredibly bad ass:

You’re welcome.

The next character in this series may or may not have the best mustache in the history of cinema.


email from Jake

I just got this email from Jake.  I laughed.