Last week I published an article recommending a TEx5 start to your Scott Fish Bowl drafts. This week we added SFBX scoring settings to the Win the Flex app, and it disagreed with my take in no uncertain terms.

According to the Win the Flex numbers (using SFBX scoring and FFPC Superflex ADP), tight ends don’t become the best pick until about the eighth round. In fact, picking a tight end anywhere in the first three rounds appears to be the worst pick you can make. Come on, Win the Flex, why would you do this to me?

How Win the Flex Works

To understand what’s going on, we need to understand a bit about how the Win the Flex tool generates projections. Behind the scenes, it uses what is known as a locally estimated scatterplot smoothing (LOESS) regression. The name isn’t important — what you need to know is that it works by dividing the entire population into “neighborhoods” and then running local regressions within each neighborhood before combining them back into the projections you see represented above.

This means that when we ask it to project Travis Kelce (the TE1 by ADP), it looks not only at what the TE1 by ADP has done in prior seasons, but also what the TE1’s neighbors have done — the TE2, the TE3, and so on for all the TEs in the neighborhood.^[1]The algorithm weights neighbors according to their distance from the player in question. This means Kelce’s projection derives not only from what Kelce did last year (as 2019’s TE1) but also in part from Rob Gronkowski’s 2018 season (he was 2018’s TE1), Jordan Reed’s 2017 (he was the TE3), Evan Engram’s 2019 (TE4), and Greg Olsen’s 2018 and 2017 seasons (TE4 both times). Apart from Kelce’s 2019, all of these players scored fewer than 170 SFBX points.

ADP-Based Projections and Injury Risks

Tyler Eifert scored 10.6 total SFBX points in 2017 after being drafted as the TE6. Obviously any position can have these sorts of outcomes where players drafted early get injured in the first game and score very few points — consider David Johnson’s 2017. But at TE, where there are fewer early options and a lot of injuries, these outcomes get exaggerated.

So does that mean you should avoid early TEs, as the Win the Flex tool suggests? Are TEs too risky to use your early picks on? I won’t attempt to predict injuries, but there’s no doubt Kelce and Zach Ertz have proven more durable over their careers than Gronkowski and Reed. So there are good reasons to think we might know something that the Win the Flex tool doesn’t.

However, there’s little doubt early TEs suffer injuries more often than other positions. Among players with top-12 positional ADPs, only RBs are close to TEs in terms of the average number of games missed per season.

From this perspective, TE would appear to be the riskiest position to draft early, whereas loading up on quarterbacks looks to be not only the safest option but also the option with the most value.

Should We Avoid Risks Early in Drafts?

However, as we know, things are more complicated. Although it’s true that early TEs are risky — that they bust more frequently than other positions — still drafting a TE early appears to be the right move. In FFPC Superflex leagues, drafting a TE with one of your first three picks almost always gives you a sizable win-rate advantage. If you pick near the end of the first round, taking a TE with your first pick is probably your best option.

First-round running backs actually decrease your expected win rate from most slots after Pick 5. Even in those slots at which we do see an advantage, the win-rate bump is minimal.