Describing and predicting the world of sports in the language of mathematics

### Week 4 NFL Picks from the Machine

Here we go again. It’s Wednesday and we are 24 hours from week 4 NFL Thursday night football. The Machine followed up it’s 12 win 4 losses in week two with 11 wins 5 losses and a tie last week. Some might say it’s crazy to be giving away picks that are running a 71% pick rate against the spread. But whatever. I’ve got like 600 followers on Twitter, and most people have been following the Rotoquant from the start. I like to hear my friends say they are using my picks to win their yearly competitions. Hell, my mom even won a pick-em contest in week two, picking more games accurately using the Machine than anyone else in the city of Boise, Id. You’re all welcome.

If you could share the blog, follow me on Twitter @TheRotoquant and/or sign up at Victiv: the next evolution in fantasy sports, it’d be greatly appreciated. Mom – I know you don’t have a Twitter account, it’s OK. To everyone else – please keep these good weeks in mind. When the picks inevitably go wrong one of these weeks, try and go easy on me … Yes, I built the Machine from scratch, and it’s a half-sentient complex blend of algorithms that search through thousands of variables to predict these games, and yes, I believe that it’s perhaps the most predictive NFL betting algorithm on earth. But this is still sports betting, and things still go wrong.

Now that that’s out of the way, here are the picks for week 4 sorted by least to most confident. The highlighted team is the suggested pick, the absolute value of the factor is an indication of the confidence in that pick – a larger absolute value indicates higher confidence:

Last week we had our first ‘lock’ and, as locks so often do, it went predictably wrong. So while my high confidence picks (any pick made with a factor greater than 3.0 or less than -3.0 is considered high-confidence) went 7 and 2 (78%) against the spread last week, my *most* confident pick was a horrible, horrible miss. It was on both my LVH Supercontest cards – where I’ve potentially got \$1 million on the line – so I was as disappointed as anyone to see them lose. If you’re confused about how such a high confidence pick could possibly go wrong, please revisit paragraph 2 of this post … Any time anyone uses the term ‘lock’ I cringe – I cringe hardest when I use the term. There’s simply no such thing. Stuff happens. If you think you really have an edge you should never, ever blow your bankroll on a single game, no matter how confident you are in the pick. Lesson learned.

My cards this week will be as follows:

TheSanction – BAL, NE, TEN, NYG, ATL

Rotoquant – BAL, NE, TEN, ATL, CHI

Good luck in all your contests this week – we’ve increased our payouts at Victiv for the 4th week in a row, running an eye-popping \$34,000 in guaranteed payouts and free contests this weekend… Let the fantasy football bonanza never end! Football season is only 18 short weeks so play now or forever hold your peace.  I’ll be back later in the week with an analysis of the value picks from last week, as well as a new crop of picks for this weekend. The top secret fantasy sports weapon of mass statistical destruction I’ve dubbed ‘TheMegatool’ is coming… stay tuned.

For sports, fantasy sports and sports betting analytics updates throughout the year Follow me on Twitter @TheRotoquant.

Cheers!

### Week 3 NFL Pick Results: 10 – 5 – 1 ATS

Another good week from the machine – picking 11 of 16 against the spread and 7 of 9 for those games over a 3.0 threshold. Over the last two weeks, excluding the always wonky week one games, the machine has gone 22 of 31 against the spread (71% with 1 push) – that’s pretty good and exactly commensurate with the expectations from backtesting (for the record, week 1 in backtesting was expected to be sup-par as well). The odds of flipping 22 heads in 31 tosses is only 1 in 100. Every winning week does, and should, increase confidence in the validity of the Machine. Still, it’s not entirely impossible to get these results by chance alone, so we shouldn’t get too excited just yet.

Here are the official week 3 results:

I’m obviously feeling much better about the LVH Supercontest. Both my cards are now in the top 10% of all the entries. They stand as follows:

TheSanction (tied for 130th of 1403): 10 wins, 5 losses

Rotoquant (tied for 130th of 1403): 10 wins, 5 losses

You can see the official rankings, along with my entries, here.

As I mentioned last week, the Supercontest only pays prizes to the top-30 finishers – both my cards are still 100 spots out of the money – but outside of confirming the predictive power of the Machine, this week was advanced my cards up the leader board several hundred spots. Their are no perfect cards, and only two cards with only 1 loss. At 10 and 5 I’m a mere 4 picks off the lead, a difference that could be made up in a single perfect week.  The top prize is \$736,575, and 2nd place will likely be half that – my goal is to take down both prizes for over \$1 million dollar take. Follow along – it’s going to be a wild ride.

The next evolution of fantasy sports – Victiv – also had an awesome week 3 of NFL action.  We’ve already are put up \$24,000 in free and guaranteed prizes for this weekend and are still offering a 100% drip-in deposit bonus up to \$1,000 if you sign up today. Draft a team and ride your fantasy knowledge to glory, power, money, bragging rights, and potentially more friends. Who knows, you might even get some chicks – take it from me, there’s nothing women love more than a guy who obsesses about fantasy sports and sports analytics 24 hours a day, spends 12 hours every Sunday screaming at the TV and drinking beer, and ends up making bank… what’s that mom? Women only like the last thing? Well crap … play anyway… because It’s really, really fun.

I’ll review the performance of my DFS value plays from last week, put out this week’s Supercontest picks and put out week 4 Victiv value plays later on this week. I also hope to be launching an interactive NFL MegaTool that will give you a dynamic UI to help make decisions on who to start / sit based off some harder-to-come-by data like participation, targets, and total team plays. Hoping to launch the RB version later this week – I’ll keep you posted.

Follow me on Twitter @TheRotoquant, and please share these picks and this blog with your friends. If all goes well I wont be able to announce them so publicly for long.

Thanks so much, and good luck to everyone in their contests this week!

### Week 3 NFL: VICTIV Value Plays

Week 2’s value plays were mixed bag. First the disappointments.

Jake Locker was a flop putting up only 14 points. Shonn Greene at 4 points and T.Y. Hilton at 6.5 were both less than stellar plays.

On the positive side Alfred Morris put up 20.5, Delanie Walker was the 2nd highest scoring TE of the week with 20.2 points. The Cardinals, one of my recommended defensive picks, put up a solid 21 points tied for the second highest scoring defense of the week.

Dez, Forsett and the Texans D were all just OK.

Moving right into week 3, here are the value picks at Victiv:

QB:

Cam Newton (\$8,600) – Cam looked good last week coming off an injury, and made plenty of good throws to demonstrate the ribs feel fine. At home against an average Pittsburg D, could have a nice game.

Geno Smith (\$7,100) – One of Smith’s greatest assets from a fantasy perspective is the Jet’s porous secondary. I expect Cutler, with a healthier Marshall and Jeffery, to score quickly, forcing Smith and the Jets to throw the rock to play catch-up.

RB:

Marshawn Lynch (\$8,100) – Through two games, Lynch is the highest scoring back in fantasy football. While I don’t think this trend lasts all season, in week 3 I expect Seattle to push the run, control the clock, and try to keep Manning’s high-powered offense off the field. More touches for Lynch means good things for fantasy owners.

Pierre Thomas (\$6,700) – Thomas has had a rough start to the season, being eclipsed by Ingram who was in top form prior to being injured. With Ingram out in week 3, NO will lean on Thomas and Khiry Robinson to shoulder the load.

WR:

Marques Colston (\$6,100) – Colston was shut out in week 2 and has been eclipsed by Graham and the rookie Cooks so far in 2014. His price has fallen accordingly. This week NO loses Ingram, and returns home to play indoors where Brees thrives. Expecting a bounce-back performance.

Mohamed Sanu (\$4,800) – If A.J. Green is sidelined, Sanu is an obvious value play stepping into the WR1 role in the serviceable Bengals offense.

Sammy Watkins (\$ 3,800) – After a quiet week 1, Watkins found the end zone in week 2 and potentially some chemistry with EJ Manuel. He’s a very risky deep value play, but the price is right and the upside is real.

TE:

Dennis Pitta (\$4,000) – Daniels proved he’s not to be ignored in Kubiak’s offense, but Pitta is still the guy in Baltimore. With Cleveland’s Haden distracted by Steve and Torrey Smith, I expect both tight ends to put up decent numbers here.

DEF:

Texans (\$3,900) – The New York Giants are bad at football. Really bad. A solid Texans’ defense should be able to cause havoc for Eli Manning and put up some respectable fantasy numbers.

To convert your football knowledge into real cash prizes each week, check out the action at Victiv.com. This week get your piece of our \$28,000 in promotional contests – they are filling fast!

Good luck to all in your week 3 action. Follow me on Twitter @TheRotoquant for more picks, insights, content and updates every week.

### Week 3 NFL Picks from the Machine

Here we go again. It’s Wednesday and we are 36 hours from week 3 NFL Thursday night football. The afterglow of last week’s 12 of 16 performance against the spread has faded. There’s no such thing as momentum in sports betting, if this week turns out 4 of 16 I’ll be right back where I started – sitting down to a piping hot plate of humble pie and wondering where the bugs are in my algorithms. Am I excited? Hell yeah I’m excited! Am I still nervous?! Absolutely.

Here are the picks for week 3 sorted by least to most confident. The highlighted team is the suggested pick, the absolute value of the factor is an indication of the confidence in that pick – a larger absolute value indicates higher confidence:

You might notice that the factors are much higher this week than last, including a -13.2 factor for the Titans game. In backtesting a factor that high only occurred 15 times last year, and picks above that threshold went 15 for 15. So the Machine is making the numerical equivalent of predicting a ‘lock’, which is terrifying. Can’t wait to see how it turns out.

My cards will be as follows:

TheSanction – TEN, ATL, SD, ARI, IND

Rotoquant – TEN, ATL, SD, ARI, HOU

Good luck in all your contests this week – we’ve got \$28,000 in guaranteed payouts running at Victiv this weekend – it’s going to be a freaking bonanza! I’ll be back later in the week with an analysis of the value picks from last week, as well as a new crop of picks for this weekend.

For sports, fantasy sports and sports betting analytics updates throughout the year Follow me on Twitter @TheRotoquant.

Cheers!

### Week 2 NFL Pick Results: 12 of 16 ATS

No one has become rich picking 8 of 16 winners against the spread. In fact, many people have probably gone broke doing just that. Being bad, but not bad enough to realize how bad you really are, is one of my biggest fears when it comes to sports betting. I’d rather lose huge and know I’m no good than be perpetually teased by untimely variance making me believe I have an edge. In sports betting specifically, and predictive analytics more generally, a false belief in one’s ability to predict the future is an extraordinarily dangerous thing.

So last week, when the Machine went 8 of 16, I wasn’t thrilled. Very valid questions crept into my mind – maybe something was missed during backtesting and the model is bogus… Maybe the method I’ve dreamed up just isn’t that valuable… maybe my \$3,600 of LVH Supercontest entry fees was a complete and utter waste… maybe, maybe, maybe. The reason I entered the contest in the first place was to prove out the method in a public forum that cannot be manipulated, a living testament to the predictive power of the Machine. Maybe after so much fan-fare it simple wasn’t predictive. It’s a possibility that every data scientist needs to be  ready to accept: the data is in, your model is no good, back to the drawing board.

Of course, variance works both ways. We can’t go abandoning methods we believe in at the first sign of weakness, the first bad pick, or bad week. Going 50% for 16 games was well within the expected variance, albeit obviously lower than my expectations. With all these doubts going into week 2 it’s difficult to overstate how happy I was to bounce back in such an authoritative fashion, with the Machine going an impressive 12 of 16 against the spread for the week. My 7 unique Supercontest picks went 6 of 7 agains the spread. I always like to put sports-betting performance in perspective by comparing it to the probabilities of seeing the same level of performance if, instead of using the Machine, we were to pick games at random by flipping a coin. The probability of flipping a coin to achieve the mediocre week 1 performance, flipping heads 8 times or more in 16 tosses, is approximately 60%. We expect a coin to get 8 games or more correct most of the time. So going 8 and 8 isn’t an inditement of the Machine, but it also isn’t much of an endorsement.

Now, the probability of flipping 12 or more heads in 16 tosses – the week 2 Machine performance – is a meager 3.8%. Certainly 3.8% is a non-trivial probability. Even going 12 of 16 there is not sufficient statistical evidence to rule out the possibility that the process which I used to pick 12 of 16 against the spread isn’t a random process no better than a coin flip. With such a small sample size the jury is still out statistically speaking. But a random process is much less likely to go 12 of 16 than is a process that does have edge. So after week 2 I remain a skeptic, but my dread about having potentially wasted my LVH Supercontest entry fees are assuaged. I’m moving into week 3 with a modicum of optimism…

Here are the official week 2 results:

My cards now stand as follows:

TheSanction (tied for 406th of 1403): 6 wins, 4 losses

Rotoquant (tied for 223rd of 1403): 7 wins, 3 losses

You can see the official rankings here.

The Supercontest only pays prizes to the top-30 finishers, and my best card Rotoquant is still almost 200 spots out of the money, but this week was a big step in the right direction. It doesn’t take too many 5-0 weeks to rise to the top of the leaderboard. The top prize is \$736,575, and 2nd place will likely be half that. My goal, as ambitious and potentially silly as it sounds, is to take 1st and 2nd for a cool million pre-tax. Hope you enjoy following along, it’s already the most exciting football season I’ve ever experienced.

The other bit of good news is that our site Victiv – the next evolution of daily fantasy sports – had a successful week 2 of the NFL season and will be rolling out it’s biggest set of contests yet in week 3. We are putting up \$28,000 in free and guaranteed prizes this weekend and offering a 100% drip-in deposit bonus up to \$1,000 if you sign up today.

I’ll review the performance of my DFS value plays from last week, put out this week’s Supercontest picks and value plays later on this week.

Follow me on Twitter @TheRotoquant, and please share these picks and this blog with your friends. If all goes well I wont be able to announce them so publicly for long.

Thanks so much, and good luck to everyone in their contests this week!

### Victiv Week Two Value Plays

Fantasy football contests and freerolls are going off tonight and all weekend at Victiv – we will be paying out \$27,000 in free and guaranteed prizes through Monday.

I wrote up an analysis of our value plays just in case you’re having trouble setting your lineups. Also, if you’re new to the game, check out the Victron tab on the contest enter page. It’s a daily fantasy noob’s best friend.

Victiv Value Plays:

QB

Jake Locker (\$8900) – Going against the abysmal Cowboys defense. Locker has the ability to make things happen with his legs. Could put up big numbers.

RB

Alfred Morris (\$5,800) – Alf averaged 6.5 yards per carry last week against a tough Houston front. At this price point any true feature back represents value.

Shonn Greene (\$5,100) – People forget that Shonn Greene is the starting back for the Titans. Greene outplayed his backup Bishop Sankey last week, receiving more than twice Sankey’s workload. Up against the Cowboy’s defense.

Justin Forsett (\$3800) – With Ray Rice out of the league, Forsett stands to split the load with Bernard Pierce this week who should also be widely owned. Forsett should be the change of pace back, but the price is right. Thursday games only.

WR

Dez Bryant (\$6,600) – Tony Romo looked terrible last week. That being said – Dez is an elite talent and it would be a surprise if the Cowboys don’t try to get him more involved in the offense against the Titans.

T.Y. Hilton (\$5,300) – Hilton is a major boom or bust play this week. Going against Philly in what is expected to be a high-scoring affair (over under at 53.5, highest of the week), I think Hilton makes a nice large-field tournament play here.

TE

Delanie Walker (\$3,200) – Walker will be one of the most added TE’s in season-long formats having a solid week one showing. Again, going against the Cowboys there should be plenty of points to spread around.

DEF

Cardinals (\$5,100) – Despite being hampered by injury, Arizona looked tough and ready to play against an above average San Diego offense on Monday night. This week they face the Giants offense that is apparently totally lost. I expect big things.

Texans (\$3,500) – Loosing Clowney is a hit, and Carr looked decent in his debut for Oakland. That being said they face a rookie QB on a below average Raiders team and the price is right.

Good luck to all in your week 2 action. Follow me on Twitter @TheRotoquant for picks, insights, content and updates every week.

### NFL Week Two Picks from the Machine

Week one went by in a blur, and week two is now staring us dead in the eyes. I’m finally feeling better after a full year of devastating withdrawal symptoms from the lack of the Red Zone channel. On Sunday the crack cocaine of sports entertainment was back in full effect, with a compliment of new graphics, octo-screen simulcast madness and a stream of sweet glorious touchdowns. Many of the devs, myself included, were in the office all Sunday at Victiv glued to our big screen TV’s where we monitor all the action, soaking up every play.

Since the end of Monday night I’ve been running numbers and prepping for this week. The lines are set, and the picks are in. Without further ado, here are the Machine’s week 2 picks:

The factors are pretty low by comparison to last week – but anything over 3.0 has backtested (and shown historically) to perform at 70% or better against the spread. Let’s hope this performance continues this week.

My Supercontest picks TheSanction:

NE, WAS, BUF, NYG, DAL

My Supercontest picks RotoQuant:

NE, WAS, BUF, HOU, PHI

Good luck to all! I’ll be posting value plays on Victiv later on today!

### NFL Week One Pick Performance

To my knowledge, I am not a father. Hell, I can barely take care of my dog. You see – my world is currently wrapped up in sports and sports analytics. Girls? Kids? Pets? Ain’t nobody got time for that! It’s football season!

To take this misguided metaphor further – the closest thing I have to children are my analytical methods. These algorithms I’ve spent months developing and testing are now out in the wild, making predictions and doing their best to make a name for themselves. So it’s with fatherly apprehension that I approached week one; excited to see how the algorithms would fare in the most coolheaded and objective sense, but also wanting so badly for them to perform.

As with everything that we build up in our minds, week one was nowhere near as good as my greatest hopes nor as bad as my worst fears. The algorithm went a pedestrian 8 for 8. In a whacky week, this was enough to put me well in the top half of most of my handicapping competitions. The results are as follows:

So what? It was a shrug of a week. 8 and 8 is certainly within the bounds of my expected performance for backtesting. The top 5 Supercontest picks went exactly 50-50 on both cards, with 2 wins, 2 losses and a tie. Unfortunately in the actual Supercontest the line on the NYJ’s game was -5.5 instead of the -5 on which the projection was based, so currently both Supercontest cards stand at 2-3. Amazingly even 2-3 is enough to have me in a many-way tie for 609th in a record breaking field of 1,395 competitors.

In other news, we’ve got \$27,000 in promotional contests this week at Victiv running throughout the weekend. Sign up this week, and get in to play some fantasy football for real cash prizes! Stay tuned for the machine’s week 2 picks and some Victiv value plays and follow me on Twitter for more updates @TheRotoquant.

Good luck in all your week one action!

### Week 1 NFL Picks

Football is here at last. And, as if I wasn’t committed to following the NFL action before – now I’m totally pot-committed, 100% full bore, neck deep in The League. I can’t believe the sports-centric changes that have taken over my life in the last year: I left my position as director of technology at Wolfram Solutions, moved 1800 miles from Boise Id to Austin Tx, and joined Victiv.com as CIO in charge of, amongst other things, athlete pricing and using analytics to optimize the user experience.

Overshadowed by all of this is the original purpose of the Rotoquant: the application of advanced analytics in the context of sports. I’ve hardly even mentioned that I entered the maximum of 2 entries into The LVH Supercontest, where once again I’ll be using the machine’s picks to hopefully coast my way to a cool \$500K. I haven’t had time to write up my exciting backtesting results – 70% on all games last year! Honestly, I haven’t slept in weeks, and a shower would do me and my coworkers some good.

It’s better to be busy than broke; I’m certainly not complaining. I just wanted to let anyone who has been following along from the beginning of the Rotoquant know that I haven’t forgotten my roots – the picks will still be coming weekly (unless, of course, I’m in a position to win some serious cash), and I’ll be posting additional data driven sports content throughout the season.

Without further ado, here are the (new and improved) machine’s week-1 NFL picks:

Teams in bold are projected to cover, the higher the factor the greater confidence the machine has in the pick. Fair warning – week 1 contests are the least statistically tractable, without incorporating data from all the trades, roster adjustments, drafts, etc., that have improved many teams during the offseason. I wouldn’t recommend diving right in with these picks until at least week 2 or 3, when the machine has had time to adjust to what the league looks like in 2014.

The LVH super contest are as follows:

Card 1, TheSanction: NYJ, SEA, NO, CAR, DAL

Card 2 TheRotoquant: NYJ, SEA, NO, DAL, JAX

Looking forward to an amazing 2014 season. Check in periodically to follow the progress – it’s going to be an awesome ride.

If you didn’t have the time to to go out to Vegas and sign up for the Superconest but still want to use your sports knowledge to make a buck, give daily fantasy sports (DFS) a try! Check it out at Victiv.com and take advantage of our launch promos with \$19,000 in guaranteed payouts, and \$4,000 in freerolls!

Good luck in all your contests and leagues – till next time.

### The Value of QB’s in Daily vs Season Long Fantasy Football

I’ve played DFS for several years now and seen many DFS sites come and go (this is particularly true more recently). Each site has had its own unique take on the player pool and player pricing. Some offered expansive player pools including irrelevant 5th string WR’s and backup K’s in an effort to make sure they didn’t leave any athletes out. Some had sharp salaries, some not so sharp. Some had tight caps, some loose. I have always been interested in what went on behind the scenes at these sites; what was the thought process behind salary/cap/player pool determinations? Were they trying to be predictive? Were they trying to anticipate user beliefs about the athletes to increase team diversity in large-field contests? Were they optimizing some other metric I hadn’t thought of? Were they just throwing darts?

Now I’ve stepped into the CIO role at Victiv.com and have to make those determinations myself. I had to think long and hard about what I felt the salaries setters should be doing – it’s not immediately clear. Hanging in the balance of these decisions is nearly every aspect of the user experience, from roster creation to the large field in-contest sweat. After weighing the plusses and minuses of different player pricing strategies I arrived at the following goal: salaries should be as sharp (i.e. predictive) as possible, and the cap should be set in such a way that the user is always forced to make tradeoffs when filling out their roster, but can still roster a few of the highest priced athletes (studs) if they are willing to roster some value (potentially deep value) plays.

Making salaries tighter or looser is easy – just lower or raise the salaries as a whole (or, tackling the problem the other way around, lower or raise the salary cap). But having sharp, predictive salaries is complicated – really complicated in fact. Do you have a single equation, in which the athletes position is just another variable? Do you have one equation per position? Per athlete?! What algorithm is best suited for this kind of problem? How often does it need to be retrained? How do you measure accuracy and avoid overfitting? How do you handle injury and other news that’s not mathematically tractable?  Do you filter outliers? What counts as an outlier? And the list of concerns goes on, and on.

As I started building models, creating and testing different variables for predictiveness, trying different algorithms etc., one thing continued to crop up: the QB position was just much more valuable than the others. Consistently the top 5-10 players projected for each week would be QB’s. From my DFS experience I knew this just wasn’t how salaries worked, on a given week there may be 1-2 QB’s that were priced well beyond any other players (usually Drew Brees and Payton Manning), and then inevitably the 3rd highest priced athlete would be a RB (usually Adrian Peterson) followed by a mix of QB’s, more RB’s and somewhere in the top 10 Calvin Johnson and maybe one other stud WR. There’s always a RB or two in the top 5 and a WR in the top 10. But, according to my historical data explorations, this just wasn’t realistic.

I obviously can’t divulge the methods employed to create our salaries, but to demonstrate my point that QB’s are being underpriced in DFS I wont need to. Below is a chart of the top 15 scoring athletes (sorted by per-game-played-average) for the last 3 years at each of the big 4 fantasy positions: QB, RB, WR and TE.

Only once in the last three years was a RB snuck into the top 5 of average fantasy points per game. Only 3 RB’s have made the top 10 in as many years, and only 4 have made the top-15. The story is much, much worse for WR’s. In the last 3 years only 3 WR’s have cracked the top-20! in terms of average fantasy points per game. Keep in mind that these numbers are per-game-played, the calculation here doesn’t punish players who missed entire games due to injury (we are assuming DFS’ers are not playing an athlete if the athlete isn’t playing – this is harder to do than you might think). Obviously RB’s and WR’s are much more likely to be injured in a game than are QB’s, and much more likely to be game-time decisions that end up not playing. It’s no wonder the algorithms are pricing QB’s at a premium to the other positions – QB’s tend to put up substantially more fantasy points!

But wait… Anyone who plays season long leagues seriously knows that you shouldn’t draft a QB in the first round. And if you miss out on the big 3 (maybe big 4 if you include Stafford as some analysts are this year) you should likely just wait and pick up whoever is available in the later rounds. So if we don’t care about QB’s in season long, why should we care about them in DFS?

The answer is obvious – in season long leagues you are drafting from a pool of players, against other drafters, where your picks prevent other drafters from taking an athlete. With the typical roster you are drafting 2 RB’s, 2 WR’s, a TE and a FLEX – with 6 non-QB skill positions to fill, plus a bench to accommodate bye weeks and inevitable injuries,  the advantage gained over the competition by having a consistent difference maker like Calvin Johnson is enormous. In daily fantasy football everyone can draft Calvin Johnson if they think he’s a good value for the week; we aren’t going to run out of quality RB/WR/TE’s.  Another way to look at this is from the point-of-view of objectives. In daily fantasy roster creation you simply want to put together the team that will get you the most points in the coming week. That’s a very different objective from trying to outmaneuver your league mates in a snake draft to put together the best roster that can win week-in week-out for 17 weeks.

Alright, so QB’s should be valued a bit higher in DFS based on average output. But what about the second piece of any predictive puzzle: variance? Maybe it’s the case that QB’s average more points, but it is inherently unpredictable when they will have big weeks and when they will flop. Below is that same top-15 list, but this time sorted by their coeficient of variation – a measure of each athlete’s mean adjusted variation week-to-week:

Again – this data only includes games in which the athlete played. What does the above table show? Not only are RB’s and WR’s scoring fewer points on average, they are doing so with generally greater variance than the QB position. In terms of pricing this would imply that QB’s should be priced higher based on average output and the fact that we can better anticipate when they will do well compared to their RB, WR and TE counterparts.

Any set of salaries has the potential to cause a stir in the DFS community. Pricing QB’s predictively means that QB’s will cost a bit more and RB’s and WR’s will generally be more affordable. I have no idea how users will react to this shift. But the question should be answered or at least discussed: if salary makers aren’t making predictive salaries, then what are they doing? If the QB is the most productive (and most predictably productive!) position, shouldn’t we price it as such?

If you’ve never played DFS, check out what all the fuss is about at Victiv.com. If you are a DFS fan and have never been to Victiv.com, come check out what we’ve got going on. We launched publicly this week and we feel that we’ve put together a state of the art DFS experience in everything from the real-time in-contest sweat to to the salary structure. Feel free to vent at me about the salaries on Twitter @TheRotoquant – looking forward to some honest feedback. I’d love to know how the different approach to player pricing effected your DFS experience (if at all). We will always be optimizing our experience for our users. Any and all feedback will be taken into consideration as we consistently strive to create the best DFS product in the market.

This week will be the biggest week in DFS history, and Victiv.com is happy to be a part of it with our official launch. We are offering a promotional launch series of guaranteed contests and freerolls to celebrate.

Thursday, September 4th:

• \$1,000 Countdown Freeroll

Sunday, September 7th:

• \$7,500 Bunker Buster Guaranteed (with \$1,500 guaranteed overlay)
• \$5,000 Little Berth Guaranteed (with \$1,000 guaranteed overlay)
• \$2,000 Missile Launch Freeroll

Deposit today and take advantage of our 100% first-deposit bonus up to \$1,000!

Good luck in all your week-one contests! Till next time.