In my last post, I looked at what draft strategies worked in snake drafts this past season. While there were a few interesting findings, I think the most striking for me was looking at when the first QB was drafted, I've reposted the graph below.
What is striking here is that it goes against the idea of a late-round QB. While you still should not be drafting a QB in the first or second round, once you get to the third round you have a better than random chance at winning. This actually peaks in the 4th round and drops with each subsequent round until you are back to random at round 12 and drop below random in 13 or 14. This result was supported when looking at team point totals, win totals, or final standing. Today I want to dive into the QBs of 2020 and try to understand why we see this pattern.
The Data
As with most of my posts, we will use data collected from public leagues. In order to breakdown the results by specific players, we will expand our dataset from last week to include a total of almost 9500 leagues and more than 90,000 teams. We will limit the dataset to only include leagues that had 8, 10, or 12 teams in the league (just to ensure the draft rounds are reasonable), and when looking at specific players in rounds, we will limit it to only include combinations of player and round that occurred in at least 300 leagues (to remove weird outliers like Joe Burrow in the first round). Also as with the post last week, I am only counting when a team drafts their first QB, many of these teams likely drafted a second but we will ignore that for now.
Who was actually drafted
The first thing to understand is who was actually drafted in each of the rounds. Below are the underlying quarterbacks drafted in each round:
Scatterplot of QBs drafted in each round. The size of the dot is relative to the number of teams that drafted that QB in that round. The color represents the value of the pick vs that player's average draft position with red being drafted earlier than ADP and green being drafted after ADP
Man was Josh Allen a value this year. He was good enough that even teams that took him well above ADP (for instance in the dark red dot in the 7th round) significantly improved their chances of winning the championship. Regardless of the round, Josh Allen helped his fantasy team win more often than average and looking at him this way he really appears as an outlier. He also seems to be driving a lot of the effect we saw at the start of this post. If we ignore Josh Allen, the rest of the players seem to be bimodal in how they work. There is an initial peak in the 4th and 5th rounds and there is a second one around the 12th or 13th round. This is more in line with the general consensus although even taking a QB in the 4th or 5th round is earlier than most experts would recommend. Interestingly, even that bimodal-ness is pretty subtle, suggesting that for the most part QBs were drafter where they should have been and where they provided value to their manager.
Note: One interesting comparison here is that drafting Dak Prescot in the 4th to later was a better choice than taking Mahomes in the 2nd, despite Dak's injury.
Billing Department
Consistency Breakdown
Another aspect to consider is how consistent QB's taken in specific rounds were. This past week, the Fantasy Footballers went through their "truth" series on the QB, which included going through their consistency. They broke down a QB's performance by how many points they scored, specifically defining 3 levels, Great games mean more than 26 points, Good games are 20 - 26 20 points, and Bad games are less than 15 Points. I am going to add OK games which are the games between 15 and 20 to my analysis but will keep the other categories the same. Also please note that any game where a player did not play (either due to injury or benching) was excluded from this analysis.
As we can see, QB's drafted earlier were more consistent, producing a good or great game on almost 60% of their games. Interestingly, this 60% mark holds stead for the first 4 rounds and doesn't significantly start to taper off until the 7th round. This is in line with what we saw above with a 4th or 5th round QB being the best time to target (the most bang for your buck).
If you are curious about any specific QB's consistency, you can look it up below
Value vs ADP
A final aspect to look at is how much value was created based on ADP. The easiest way to look at value is simply how many points the player scored on average. We can graph that versus the players ADP below. (Note again I am going to ignore games where a player was benched or injured)
Oh, what could have been with Dak Prescott. I know he was not going to continue at that pace, but as someone who drafted Dak, this hurts. The other takeaway from this is that Josh Allen and Aaron Rodgers were the clear breakout winners this year, outperformed their ADP and performing similar to Patrick Mahomes despite the 10th round ADP. (Note Justin Herbert was not really drafted so isn't included here, but he averaged 21 points per game or similar to Ryan Tannehill).
Limitations
The big limitation for this analysis is I am not looking at who was actually started week to week. I am taking 2 snapshots, who was drafted at the beginning of the season, and who won the league at the end of the season. For instance, As I noted above drafting Dak Prescott in the 4th gave you a better chance of winning than drafting Patrick Mahomes in the 2nd. That does not mean the person started Dak each week, but that overall those players won their league more often (perhaps by pivoting to a Justin Herbert). The Draft is how you set yourself up for success, but it does not guarantee it.
The other important limitation is this is just looking at 2020. As mentioned in previous posts, there are lots of reasons not to commit to conclusions drawn from 1 season, especially a season as weird as 2020.
Conclusions
Overall, my interpretation of this is still the same, that we (or maybe just I) have swung too far on the late-round QB. While I still don't think you need the best of the best (don't spend a 2nd round pick on a QB), starting in the 4th round or so, the pre-draft value matched the post-draft value pretty well, so just grab one when it feels right and don't be afraid to go after that second tier of QBs.
Questions? Comments? Let me know at ac@alexcates.com. Want to read more breakdowns like this? sign up for my newsletter. Finally, like what I do? Consider supporting me on buy me a coffee.