How To Get Started With Daily Fantasy Baseball – Part 1 – Background Reading

Each site is going to have a different scoring system, meaning players will be valued differently on one site than on another. This means that any “player-specific” DFS advice you see on the web better be listing site prices. Advice needs to be both player- and site-specific to really be applicable.

The thinking is that Vegas has A LOT on the line and they’re professionals at handicapping just about any sporting event. Rather than using a lot of your own time and resources to create your own projections (when you’re not an expert at it), you might be better off using Vegas’ information and using your time on other things (information gathering, checking weather, studying lineups, etc.).

Not only is this information accurate (if it wasn’t, bettors would be robbing the casinos blind), it’s also free! Over-under information is a great indicator of the expected offensive output for games and the betting lines are an indicator of what starting pitcher is expected to win the game. Some services even offer prop bets about the number of strikeouts a pitcher will have for a game (that’s like a free mini-projection for a player).

I get a definite sense that many figureheads have earned success playing DFS as a whole, but they don’t necessarily know a ton about baseball and advanced stats. It seems like they’re teaching themselves baseball.

I could be misguided, but I get this impression by the way they explain baseball terms. Things are just a little off. Not the way I’m used to hearing things explained. Almost like a hint of a foreign accent from someone that you can’t otherwise tell speaks English as a second language.

This is not to say that these DFS Experts won’t be helpful. Quite the contrary. I can tell that their expertise from other games, understanding of DFS tactics, and game theory will be very helpful along the way. But it’s worth keeping in mind that you might have to weed through their “baseball advice”, which could be somewhat misguided and oversimplified. It’s promising to think that someone without a lot of technical baseball knowledge can succeed as long as they have a great knowledge of DFS strategy and tactics.

I think this indicates there is a great opportunity for success if you can excel (I had to get that word in here somewhere) at both the DFS game/strategy element AND the advanced baseball statistics element.

• Derek Carty Follow @derekcarty – Former writer at Baseball Prospectus, the Hardball Times, and others. Not just a fantasy baseball writer either, I’m talking advanced metrics he has created on his own. He’s very sabermetrically inclined but he has also played in LABR and Tout Wars. He’s been focused on DFS for a few years now and has his own DFS projection model available for sale.
• Todd Zola Follow @toddzola – If you like my site, I’m afraid to even tell you about Todd. I’ll probably lose you as a reader. He’s more well known for his rotisserie baseball expertise and his site Mastersball.com, but he’s writing about DFS now too. Analytically-minded and data driven. You can find his season-long rotisserie work all over the web and you should go out of your way to read whatever he writes.

• Jonathan Bales Follow @BalesFootball – Here’s an example of someone whose expertise is more with DFS game play and strategy than it is with baseball analytics. With that said, Bales is very analytically focused (if you play fantasy football, I highly recommend his work) and is bringing that skill set to baseball now.

  So as not to be misleading, I should mention that Bales is the creator of the “Fantasy Sports for Smart People” series. Although it sounds similar to my site, I have nothing to do with those books, although I have read several of them and do recommend them.

• Peter JenningsFollow @CSURAM88 – A contributor to RotoGrinders.com, he created much of the “GrindersU” content I link to below and he’s also quoted in Bales book mentioned above.

The argument I’ve seen made against fully projected stat lines are twofold:

• Inaccuracy – Because we’re dealing with a sample size of only one game, hitting stat lines are somewhat True/False or Black/White propositions. A hitter can’t get 1.3 hits in 4.2 at bats with 0.18 HR on a given night. They’re probably going to be 0-4 or 3-5 with a HR and a double.


• Efficiency – Developing a fully projected stat line for a player would be very time consuming unless you had very technical automated systems in place.

My problem with these two arguments is that I haven’t seen research to back up the inaccuracy claims. They make sense to me. But I can’t dismiss all the weight sabermetricians put on projections and the research I’ve previously seen that suggests we can have somewhere between 65% – 70% accuracy with projections for the entire season.

Just because they may be inaccurate on a given night, if I use something that’s 65 – 70% accurate over the course of the season, wouldn’t I be better off over the long run? It seems like wOBA is going to be inaccurate on any given night too. A player’s wOBA could be anywhere from .000 to .800 on a given day.

The efficiency argument might have some merit, but hopefully I can find some ways around it by implementing things in Excel to calculate things more efficiently.

This becomes less important for sites that have you start multiple pitchers (some sites only start one, others as many as three), but the reasoning behind investing heavily in a starting pitcher is due to the predictability of pitchers as compared to hitters.

This is the exact OPPOSITE of what we hear in rotisserie baseball. We are told that “pitchers are unpredictable, so draft reliable hitters early.” So now being told that “pitchers are reliable, spend big on them”, is a difficult pill to swallow.

I reconcile this in a few ways:

• To liken back to the example hitter stat line I used above, pitchers’ projections are not so black and white. On a given night we see 1/162nd of a hitter’s performance for the season, whereas we’re seeing 1/32nd of the pitchers season. This larger relative sample size smooths out a lot of the unpredictability for pitchers.
• When we are told “pitchers are unreliable”, people are not saying, “their projections are unreliable”. They’re really saying, “We are not good at projecting who the top 108 pitchers are going to be at the end of the season”.
• This is also a function of the time horizon we are dealing with. For season-long performances, it’s fairly easy to project a hitter’s stat line. But if you shorten the horizon to just the next four at bats, the task is much more difficult.
• Injury risk plays a large role in the valuation and projections for pitchers when you’re playing the season long game. But the risk of injury on one given night is not significant and in DFS you’re only investing for one day if the player were to get hurt. For season long leagues you’re stuck with the hole of an injured player indefinitely.

The bottom line is that a hitter might only put the ball into play a couple times on a given night; while the pitcher is going to throw 100+ pitches. All of those extra opportunities (larger sample size) a pitcher has on a day in which they start allow us to make more accurate projections.

But depending on the contest type you choose for DFS, the 6th place analogy (in a 12-team league) above is actually enough to win (50-50 games). While in the larger prize contests you do need to finish much higher in the pool cash.

If you only had to finish in the top 6 of your 12-team roto league, you would not be drafting that trendy minor league prospect in the fifth round. Instead, you would pay for the unsexy but steady veteran. You might even pursue that same approach all the way through a draft, just loading up on proven veterans and solid players. You would do this to build up a solid predictable base of statistics likely to get you into the top half of the standings. You would be less likely to chase highly variable players when you can rely upon the proven and safe investments.

If you were in a winner take all roto league, you would take more chances. You wouldn’t be foolish about it. You’d stick with the consensus rankings for the first 5-7 rounds, but after that you might begin to embrace more risk. You might sneak in a few players you expect to break out, taking them earlier than recommended. You wouldn’t ignore veterans, but you would be more inclined to gamble on youth and the unproven later in the draft, in hopes of hitting big on a potential star.

The takeaway here is that you should construct lineups considering the type of contest you’re playing. We’ll need to use this line of thinking to determine our optimal strategy for the DFS contest type we’re playing.

Not only do we want to choose our optimal lineup, but we also need to beat other participants. In an extreme example, assume everyone had access to a super powerful Excel spreadsheet that spit out the optimal lineup each day, and everyone chose a very similar lineup in a DFS contest… Your chance of winning would be slim. So not only do we need to pick a strong lineup, we need to be cognizant of what others might do and act accordingly.

In the season long game, the strategies of our opponents are transparent. We can go and look at their starting roster at any time and develop an expectation of our best strategy against them in the standings.

In DFS, we’re not afforded that luxury. A large part of being successful is understanding game theory (how to make optimal decisions in light of what we expect our opponents to do).

The web is flooded with DFS information and it’s becoming more and more mainstream. I mentioned Derek Carty and Todd Zola above. Their work is being featured on ESPN now! It seems like there will be an opportunity to get a sense for what the crowd will do by following the advice the most popular sites will be giving.

“How to Evaluate Batter Matchups” by Derek Carty

“How to Evaluate Pitcher Matchups” by Derek Carty

I’ve already mentioned a lot of my ability to play DFS is going to depend on my ability to create quality lineups in a short amount of time. I’ll try to do that by building a spreadsheet to automatically bring information I need into one centralized place, while simultaneously processing salaries, projections, and identifying good buys.

That research routine and decision making process will need to include whatever stats I decide to use (probably some sort of handedness split), ballpark factors, weather, Vegas lines, and lineup information.