Bias & DFS
Let’s play a game. I’m going to provide two pitcher stat lines, and I want you to try and guess who they belong to. No cheating:
Line 1: 43.1 innings, 5.40 ERA, 7.5 K/9, 9.3% Swinging Strike Rate
Line 2: 41.1 innings, 3.27 ERA, 11.1 K/9, 13.7% Swinging Strike Rate
Got your guess ready?
If you guessed Trevor Cahill, you’re right (and you probably cheated because nobody should know that answer without Google). In fact, both of the lines belong to Trevor Cahill. Line 1 is from his 2015 season, and Line 2 is from this 2017 season.
Now I’m going to ask you a question and I want you to be honest. Here it is: Is Trevor Cahill good?
If you answered “yes”, would you have answered “yes” back in 2015? Would you have even answered “yes” during this past Spring Training? If you answered “no”, at what point in 2017 would you change your answer to “yes” if he continued his 11.1 K/9 ratio and 13.7% swinging strike rate? Do you need more data, a larger sample size, or do you just refuse to believe Cahill is good?
I was compelled to write this article because now that we’re nearly a month and a half into the 2017 season and we’re starting to get decent sample sizes of player data, I’ve had to come to grips with biases I’ve held onto since the start of the season and evaluate whether they are fair to keep holding onto. Regarding my question above, it’s been years since I’ve considered Trevor Cahill a good pitcher. But these past few slates, we’ve had to deal with Trevor Cahill chalk, and it’s clear something has changed in either Cahill’s skills, our perception of Cahill in the DFS community, or both. Trevor Cahill doesn’t become chalk for no good reason.
So what’s different? I love what JMToWin wrote in his recent MLB Edge article: “Cahill has legitimately been a thing this year and Cheeseisgood was the first to turn me onto him a few weeks ago. It’s been mentioned he’s shown increased velocity on his sinker this year and the results have been a bump in strikeout rate.” So there we have it, Cahill’s results have changed because there’s actually been a change in his skillset. So is it now fair to say Trevor Cahill is good, even though my biases have told me to ignore him since 2013?
When I talk about bias, I’m talking about a predisposition we have towards something. In terms of DFS, our biases can be shaped by a variety of factors. Maybe we have a positive bias towards a certain player because a tout we respect is hyping him up. Maybe we have a negative bias towards a certain player because we once watched a game on tv, saw him strike out three times, and have since labeled that player as a fringe Major Leaguer who can’t make contact. Maybe it’s recency bias where we refuse to roster players who have been the WOAT recently, and we use box scores to assume a player won’t produce today because they didn’t produce yesterday. Maybe we only stack teams with high-implied Vegas totals without giving much thought to the idea that perhaps Vegas is wrong today. Biases manifest themselves in many other ways: perhaps we refuse to roster players who are over a certain age, or refuse to roster players on a rival team (I’m a San Francisco native so rostering Dodgers felt extremely awkward and disloyal when I first started playing DFS until I came to realize the goal of DFS is profit, not loyalty).
Yonder Alonso is another perfect example of bias. Check out Alonso’s stat lines from the past three seasons:
2015: .282 batting average, 5 HR, 31 RBI, .327 wOBA, .099 ISO (103 Games)
2016: .253 batting average, 7 HR, 56 RBI, .299 wOBA, .114 ISO (156 Games)
2017: .302 batting average, 12 HR, 29 RBI, .444 wOBA, .387 ISO (34 Games)
In only 34 games, Alonso has tied his home run total from the past 259 games, which is simply amazing. It’s been well documented now through an article by Fangraphs that Alonso has changed the mechanics behind his swing to generate more lift and increase his launch angle. So here’s the question: is Yonder Alonso now good? Is this 34 game sample enough to make us believe Alonso is a legitimate power hitter we need to roster in DFS, or do your biases plead with you that this is simply a fluky hot streak?
This Spring Training, I had the privilege of writing team previews for every MLB team. I went back to read some of them and my biases were very clear. I mentioned that Daniel Murphy would probably regress to his pre-2016 averages (he’s currently batting .331 and on-pace for another 20+ home run season). I said the Yankees might have some usefullness for DFS purposes but aren’t very appealing (they’ve hit the 4th most home runs in the majors thus far). I had my doubts about Eric Thames being a threat (we all know how that has gone). But here’s the thing – all these statements were influenced by my biases. I conditioned myself to view Murphy’s breakout season as simply a flash in the pan, rather than account for the fact that like Alonso, he has changed his launch angle. I wrote off the Yankees as a team lacking DFS potential without fully understanding the impact Aaron Judge and Matt Holliday could have on the team (and not realizing how much freedom a team could have by getting rid of Alex Rodriguez). I refused to give Eric Thames’ massive home run numbers in Korea the credit they deserved because I’ve seen so many players come from overseas and fail in the majors. Those were my biases, and as we keep rolling on into 2017, I have to be okay with changing those biases if the data tell me otherwise.
One thing I appreciate about RotoGrinders is that the content forces me to think about my roster decisions. As some would say, the content is teaching me how to fish. It’s one thing to read an article that says play Trevor Cahill, and that’s all I go off of. It’s another thing to tell me Cahill’s velocity has increased so that I can make an informed decision on whether I need to re-evaluate my bias towards Cahill. I enjoy reading Gimino’s/Seth’s Trendspotting article because it digs into numbers and data that can reinforce a change in launch angle or exit velocity. In other words, part of my evolution as a DFS player lies in asking a simple question, “Why do I believe what I believe, and how likely is it that I’m wrong?” And by being able to stop and ask myself that question, it allowed me to click the name Trevor Cahill and Yonder Alonso before other DFS players may have caught on.
So I’ll leave with this – what are your DFS biases? Have you ever stopped to sit down and think what your DFS biases are and how they are impacting your daily roster decisions? And lastly, if you’re presented with data that contradicts your DFS biases, at what point are you willing to change them and admit your biases are wrong?
I would love to hear in the comments below what biases you struggle with, or feel free to hit me up on Twitter here. Thanks for reading.
