MLB DFS Strategy Session: Site Specific Scoring (Hitters)

Most DFS analysis focuses on variables. Because that day is the only thing that matters, we tend to focus on what park the player is playing in that day, what the weather is like that day or what kind of matchup the player has that day.

But as Nick talked about in his DFS Strategy article from last week, part of being successful is playing consistently to mitigate the random variation that can occur on any particular day. By playing consistently with solid research, you can lessen the impact of the day-to-day variance and make DFS play a bit more like a season-long rotisserie league.

With that said, the day-to-day research and analysis still needs to focus on variables. But in these strategy articles we can take more of a long-term approach. Today we’re going to look at how scoring differs on different DFS sites to see what kinds of players might have more or less value given a specific scoring system.

The idea here is that the daily variables aren’t the only factors that matter, and a player’s general skill set should not be overlooked. The hypothesis is that certain types of players may be over- or under-valued in specific scoring systems. There may be little market inefficiencies that can give us a leg up on the competition that’s only looking at the day-to-day.

It should be noted that this approach is geared toward cash game analysis. In GPPs you need to focus on the daily factors to try and identify players with significant upside. But in cash games finding a little extra value here and there can be key.

To test this hypothesis, we’ll simply take the results from 2016 and calculate how hitters performed on per plate appearance basis on different sites (min. 300 PA). We’ll include DraftKings and FanDuel obviously, but we’ll also look at Yahoo! and FantasyDraft since you can find projections for those sites on our research dashboard.

Once we break it down to a per plate appearance basis, we’ll look at how a player performed relative to average (his z-score) on each site. And finally, we’ll compare a player’s z-scores on different sites to find which scoring system best and least suits his game.

DraftKings

When you compare the z-scores for players in DraftKings scoring to their average z-score on all four sites, it’s clear what kind of player gets a bump on DK. Here are the players with the biggest positive difference between DK and average.

Each one of those guys swiped at least 30 bases last season, so it’s clear that DK’s scoring system rewards speed more than others. On the flip side, it’s clear what DK’s scoring system doesn’t reward more than others. Here is a list of the guys with biggest negative difference between DK and average:

Among our sample of 268 hitters with at least 300 PA last season, Chris Davis, Ryan Schimpf and Chris Carter finished first, fourth and fifth in three-true outcome percentage, respectively. If you don’t know, the “three true outcomes” are strikeouts, walks and home runs. That means DK rewards putting the ball in play, which also explains Bautista and Coghlan being on this list. Bautista finished third in walk rate last year, and Coghlan finished next to last in batting average and third to last in BABIP.

Tl;dr Hitters who hit their way on base and then run to the next one have more value on DK.

FanDuel

Here are hitters with the biggest positive difference between FD and average:

Look familiar? The top three guys showed up on the list of guys with the biggest negative differences on DK, so it’s obvious that FD is more favorable to the high three-true outcome guys.

This is a good time to point out that this does not mean these guys are playable on FD but not DK. For example, Russell Martin ranked 139th out of our 268-hitter sample on FD and 165th on DK. In other words, he was below average relative to this sample on both sites. If Martin was super cheap on both sites and facing a bad, homer-prone left-handed pitcher in a good ball park, he could be considered as a salary relief option on either site. It’s just that he might be a little more viable on FD if you’re on the fence about using him.

At the other end of the spectrum, we don’t even need to post the list of players with the biggest negative differences on FD. It’s virtually all the speed guys we saw with the biggest positive differences on DK.

Yahoo!

Below are the guys with the biggest positive difference between Yahoo! and average. Tell me if this looks familiar.

As you can see, Yahoo! scoring more closely resembles FD than DK. To put a finer point on this, these sites simply reward high walk/high OBP batters more. The hitters listed above ranked as follows in walk rate among our 268-hitter sample: 1st, 12th, 23rd, 3rd and 34th. When you look at the difference in their rank in batting average to their rank in OBP, three of them have the biggest differences in our sample and four of five rank among the 10 biggest differences.

To compare it to seasonal leagues, just imagine FD and Yahoo! are OBP leagues.

FantasyDraft

FantasyDraft scoring is identical to DK’s with one small but important difference. DK has no penalty for being caught stealing whereas FantasyDraft has a two-point deduction for being caught stealing. Given that the players with the biggest positive differences on DK were base stealers, we see a slightly different list of players with the biggest positive differences on FantasyDraft.

There’s a little speed on that list with Turner stealing 33 bases last year and Owings stealing 24. But the real common thread among them is a low walk rate (all under five percent last year) and a high BABIP (all above .334 last year).

The take away is that the DK and FantasyDraft scoring systems are more susceptible to variance (BABIP), though DK’s favorability towards speed covers that issue up a bit. But FD and Yahoo! reward more repeatable skills like walk rate and should thus see less variance and be a bit more predictable. You can’t go so far as to say DK/FantasyDraft make more sense for GPPs and FD/Yahoo! make more sense for cash games, but their scoring systems do skew a little in each of those directions.