Jonathan is the author of the Fantasy Football for Smart People book series (most recently “Daily Fantasy Pros Reveal Their Money-Making Secrets”), and a member of Team DraftKings.
In my latest book on daily fantasy sports, I interviewed some of the game’s top players to learn how they’re so profitable playing daily fantasy sports on DraftKings. One of those players was a really smart guy named Mirage88, a.k.a. A.J. Bessette. He’s actually written some ridiculously awesome articles here at RotoViz on how to use Monte Carlo simulations to win best ball leagues.
Mirage88 (if you think I’m going to continue to refer to a grown man with his real name, you’re sorely mistaken) is a big proponent of using the Vegas lines to create daily fantasy lineups, and I spoke with him about how he does that.
First, he discussed why the Vegas lines can be trusted.
I think it’s important to understand how the Vegas lines are created, which then aids us in figuring out how useful they are. There’s a perception that Vegas sets lines solely to get 50/50 action on each side of the bet. And to some degree they probably want that in many situations since they’ll guarantee themselves profit just from the juice (the commission they charge to play). But what happens is people will sometimes use that as a reason that Vegas shouldn’t be used in projections, saying something like “Oh, they just care about whatever popular opinion might be and just getting in the middle of that.”
The problem with that is that there are a lot of sharp bettors out there with a lot of money, so if Vegas indeed produces a line to equalize bets but it’s weak, those sharps are just going to pound that bet and Vegas will be in a really poor situation in terms of expected value.
So the way I like to think about Vegas is that it’s really where the most risk is in terms of projecting any player results—at least the most financial risk from one entity making projections, anyway. So if Vegas posts a poor line—let’s say they post a total that’s way too low—then all of a sudden anyone who can bet on that who is relatively sharp will just start hitting the over, and Vegas will realize that the bet isn’t really balanced.
Vegas will compensate for that by moving the total up to get more action on the under. That’s fine, but then there’s this area in the middle which was over the initial line but under the new line movement bet that’s now a really bad place for Vegas. If the game ends up in that spot, they could theoretically lose a whole lot of bets to sharps who bet the original over, but at the same time lose late bets that came in on the under when the total was higher.
Vegas doesn’t want to put themselves in that position where they can be arbitraged, so it’s really important for them to create an accurate line from the start. Even if they don’t guarantee a profit by getting equal money on each side, they can limit their downside—their risk of ruin—by making the line accurate. They really don’t want to be in a situation where they set a bad line that moves a whole lot and they could potentially lose their share on both sides of a bet.
Ultimately, making accurate lines is just a safer way for Vegas to make money than trying to predict public opinion, especially when there are sharks out there who might not agree with public opinion. Vegas has a very clear financial incentive to make accurate lines, and they do. So that’s my little rant on why we can trust the lines and why the idea that all Vegas wants is to balance bets is false.
And this is how Mirage88 actually uses the lines.
In football, I’m usually looking at projected totals the most. The easiest way to use over/unders in football is to look at the total and the spread and calculate the projected total for each team.
Once I have the total for each team, I look at some historical scoring rates—what percentage of scoring has typically gone to each position for certain teams. So let’s say we’re looking at the Packers and the Giants, who Vegas has projected at 24 points each. By looking at historical scoring, we’d see that we should project Aaron Rodgers with more touchdowns than Eli Manning just because Green Bay scores a higher percentage of touchdowns to field goals, and Rodgers also accounts for a much higher percentage of the Packers’ scores than Manning for the Giants, even with the same projected points.
You need to be careful there, too, because there can be a lot of turnover in the NFL, so things change. For example, the Giants have a new offensive coordinator and an entirely new offensive philosophy, so that data on how their touchdowns are usually allocated might change. For the Packers, on the other hand, we can pretty much assume the same scoring rates since not much has changed for them in terms of coaching or personnel.
I think that’s also a good example of projections sometimes being really scientific and other times being more of an art. With Green Bay, I’d be more likely to rely on the numbers; I can look back at however many years the same sort of scheme was in place and say, “Okay, 20 percent of touchdowns go to Jordy Nelson, 35 percent go to running backs, and so on.”
You can actually do the same sort of thing when projecting kickers, looking at a combination of the line and then what percentage of points the kicker has produced, assuming there haven’t been giant shifts in offensive philosophy or personnel.
After that, you still need to adjust for other factors, specifically the opponent. Maybe the Packers’ wide receivers account for 50 percent of all touchdowns, but they’re facing a defense that has really short cornerbacks who get picked on in the red zone, so they allow 65 percent of touchdowns to opposing receivers. Then you’d expect an even greater rate of the scoring to come from Nelson & Co.
But that’s the general idea behind what I do to at least get a baseline projection.
Vegas and Team Totals
Many sportsbooks offer individual team totals, but as Mirage88 mentioned, you can calculate a team’s projected total very easily. You take the total for the game, subtract the spread, and divide by two.
For example, this week the Seahawks are 4.5-point favorites against the Broncos in a game with a total of 49. To project Seattle’s points, we can just subtract the spread (-4.5 for the Seahawks, since they’re the favorite) from the total, which gives us 53.5 points. Divide that number by two, and we get the Seahawks’ projected total of 26.75 points.
If we do the same calculation for Denver, it would be 49 minus 4.5 (44.5), divided by two (22.25). Thus, we can effectively project this week’s Denver-Seattle game at 26.75-22.25 in favor of the Seahawks.
Anyone want to do a side bet with me that neither team will hit those totals? I’ll give you 1,000,000-to-1 on it.
Top Week 3 Projected Team Totals
Using the current Vegas lines, here are the teams projected to score the most points in Week 3.
- Saints: 30.25 points
- Patriots: 30.25 points
- Eagles: 28.5 points
- Lions: 27.5 points
- Seahawks: 26.75 points
- Colts: 26 points
- Falcons: 25.5 points
- Bengals: 25 points
- Packers: 25 points
I think the common argument against using the Vegas lines is, “Well that’s just common sense,” but I don’t really think that’s the case. If I could project games as accurately as Vegas without even doing any research and getting by with common sense alone, I think I’d kick in the research time, beat Vegas, and stop waking up every morning to write about how a 21-year-old’s hand size is going to help him throw a ball.
Plus, I don’t know about you, but I wouldn’t have ranked the teams above like they’re currently listed. I would have Denver in there, for sure, even facing Seattle. I’d probably have the Eagles as the top-projected team, and I don’t think I’d have the Lions projected ahead of Green Bay.
This is pretty meaningful stuff, I think, because it has an obvious impact on fantasy output. In my book, one of daily fantasy sports’ top players (and fantasy millionaire) CSURAM88 discussed how he uses the lines to go after specific teams instead of starting with players that he likes.
I’m a huge proponent of using the Vegas lines for information because they’ve proven to be so accurate. One thing the lines allow me to do which is different from some other players is target teams over players.
That is, I look at the Vegas lines and use those to help me figure out which teams are going to be able to score a lot in a game. Then, I try to predict which players on that team are going to be the main beneficiaries of that. That’s in opposition to some other daily fantasy players who start by looking at individual players.
Overall, though, I am 100 percent a Vegas-based daily fantasy player; I use the lines as a very strong foundation for my projections and lineups, and I don’t think there’s a more efficient and useful way to go about playing daily fantasy. In addition to game totals, I look at team totals, line movement, player props—anything I can get my hands on.
If actual team scoring is a good proxy for fantasy production, we can assume that, over the long run (i.e. if we could simulate Week 3 like 10,000 times), New Orleans will post around 20 percent more fantasy points than the Packers. I wouldn’t have guessed that to be the case, which I think suggests that the Vegas lines can give us really useful, accurate insights (in a hurry) that are extremely actionable, especially when playing DraftKings’ weekly version of fantasy football.
It’s also very possible that I’m just an idiot.