Exploiting Middle Infield Bias

“… pros were more likely to ride a wave of irrational exuberance than to fight it. One reason is that it is risky to be a contrarian. ‘Worldly wisdom teaches that it is better for reputation to fail conventionally than to succeed unconventionally’” – Richard Thaler in Misbehaving

At the root level, fantasy baseball is about acquiring more undervalued assists than your opponents. Everyone wants a first-round talent for a last round price (e.g. Aaron Judge). With teams clamoring to acquire every advantage, they are insistent on wasting away an early draft advantage. In early 2018 drafts I’ve participated in, an early emphasis on middle infielders is inflating their value way beyond their projected production. Is the observation wrong? If so why? If not, how can an owner take advantage of this mispricing?

Note: For this article, I will lump second basemen and shortstops together into one middle infield position. Neither position has more talent than the other and the bottom players will be used to fill a middle infield position.

For those who have recently created mixed-league valuations, positional scarcity doesn’t exist besides with catcher. I use the method outlined in Larry Schechter’s book, Winning Fantasy Baseball to determine my values. I’m not going to go into the process’s exact details but it’s the standard procedure used by fantasy experts to prep for auctions. Even a couple years ago a small amount of positional scarcity existed but a huge influx of good middle infielders has raised the group’s overall talent level up to the other positions.

Here’s an example of this bias. In a 12-team mock draft I just completed at CBS Sportsline, the average projected Steamer stats for the last three players I ranked at middle infield (MI), corner field (CI) and outfield (OF) are compared.

The stat mix is unique for each but the overall production is similar because stolen bases having about twice the value of home runs.

And the following table is where the middle infielders got acquired compared to the draft pick’s valuation.

Note: I know owners will have differing opinions on various projections. I disagree with some but for this analysis, there is a systematic bias being observed that is not being caused by a couple projections being off. If there is a point to argue, it’s the position replacement level.

On average, the picks went 1.5 rounds earlier than valued with the median difference at eight picks. Several ideas need to be discussed about this table.

First, owners may consider the top picks to be when to reach for middle infielders with less than a round’s difference between the value and cost. The problem is that the hitter value doesn’t decline linearly. It’s more of a logarithmic curve with the first picks valued higher.

For example, here are the hitters ranked by Standings Gained Points (SGP) with the last player at 0 SGP and everyone else with a level above zero. SGP (explained in Schechter’s Book or at SmartFantasyBaseball) is the number of points in a Roto standings a player should gain compared to another player using each of their projections.

So, Mike Trout is 2 SGP better than Giancarlo Stanton who is then 2 SGP better than Bryce Harper. The next 2 SGP drop is 10 spots to #12 overall, Trea Turner. Now, if we work our way up from the bottom, it takes 70 players to make the 2 SGP jump from #168 up #99. The overreach for Corey Seager (valued at 5.1 SGP) going at pick #20 (valued at 8.1 SGP) cost his owner about three points in the standings. He can’t make up this difference once the talent begins leveling off. It’s best to fill out a roster’s positional needs after the top 100 hitters since their values are similar. The goal is to grab as many of the top guys as possible and worry about positions later.

So why does this discrepancy exist? Here are my thoughts.

• Owners still think middle infield is still a scarce position. I believe most experts have been making this mental adjustment for years and can’t stop. This adjustment is not going to end with many top prospects at the middle infield positions (e.g. Gleyber Torres, Brendan Rodgers, Bo Bichette). Owners need to fight and try to eliminate this internal bias when constructing their team. The source of production is irrelevant if owners grab as much possible talent before the curve flattens out.
• Owners are drafting off a positional ranking list instead of an overall ranking. Owners may panic when they see all the top infielders being drafted but shouldn’t. There are enough middle infielders to go around. An owner may reach early for a shortstop who is worth 2 SGP less than an available outfielder because the other “top” shortstops are being drafted. Take the better value now and wait to fill your shortstop at the same time the Seager owner is filling up his outfield a 0 SGP value. Again, production early, positions later. It’s sometimes tough to follow this approach as Richard Thaler stated in this article’s beginning quote.
• Owners transferring real-world defensive value to fantasy value. It’s true that slick-fielding middle infielders have more real-world value. For fantasy purposes, only offensive stats count in the most standings. Using the incorrect metrics to value players can also be seen with speedsters going too high in points leagues or high AVG, low walk hitters in OBP leagues.

No matter why the bias is occurring, it can be exploited. I’ve already mentioned the key factor is to start with an actual valuation system (e.g. SGP, PVM, Z-score), which represents a player’s contribution to the standings, and a set of projections. Actually, once this step is complete, owners can verify for themselves if positional scarcity exists. The more unique the league (e.g. single league, RF/CF/LF designations, 24+ teams), the higher the chance some position (e.g. right field or third base) may be scarce. I’ve only examined 12 and 15-team formats and haven’t found one yet but that doesn’t mean one doesn’t exist.

Creating non-ordinal values helps to put a single value on all the counted stats. Creating these single valuations takes time and most casual owners won’t perform the step. Instead, they’ll use a player ranking to draft players. I’ve heard one analyst talking about changing his rankings to be more in line with the rest of the industry. Differences in values are what owners need to find. By going the extra step and not following positional rankings, inefficiencies can be pinpointed and exploited.

Another huge advantage of creating valuations is the projections can be personally adjusted. An owner may believe Alex Bregman will steal 20 instead of the 12 projected bases. The owner knows he needs to bump up his ranking some but the adjustment would be a guesstimate. With a valuation system, the adjustment’s effect can move the player up 40 spots, if valued after the top 100, or possibly none, if he’s near the top.

I haven’t observed middle infielder bias as much in auctions because owners need to spend the time figuring out dollar values. These owners know each player’s potential. While they may pay more for stars, all the stars get equal love.

In leagues with no position scarcity, owners are still overvaluing middle infielders. These owners are giving up the opportunity to pick up better talent which will help them win. They may have a case of “FOMO” or still believe good middle infielders are scarce. By using a reliable player valuation system, true values can be created and exploited. Owners need to determine if they want to fail conventionally or succeed unconventionally.