Fantasy Freestyle: Watching Mike Make the Sausage

For the past three years, I have published my bid limits for auction formats for AL-only, NL-only, and mixed leagues here at Baseball Prospectus. While I do talk quite a bit about how I change bid limits for individual players over the course of spring training due to breaking news, I spend little if any time talking about my larger process, and how I construct a series of bid limits across the board for an entire player pool. This article is a step-by-step review of this process. It is not intended to be a granular review of what I do but it is meant to be more than merely the cursory look at how I arrive at the bid limits that I provide in March.

While there are a number of steps that I do take, I have to start with a step that I don’t take, because otherwise it will trip everyone up completely.

I Do Not Project Player Statistics
Many fantasy analysts find player projections useful, and some would go so far as to say that it is essential to have player projections at your disposal to maximize your auction success. While I like to look at player projections and do use them as a barometer during my process (more on this later), I don’t believe in using projections as a baseline. So while I have the ability to project player stats (and have done so in the past) it is labor intensive, and this is time that could be better spent on superior avenues of preparation.

Additionally, player projections can create a false sense of security about what a player’s value is. This is particularly true with pitchers.

Pitcher A: 8 wins, 205 IP, 193 hits, 53 walks, 80 earned runs, 207 strikeouts, $15.
Pitcher B: 11 wins, 205 IP, 183 hits, 47 walks, 76 earned runs, 207 strikeouts, $20.

Pitcher B is Jon Lester. Pitcher A is Jon Lester with some slightly modified statistics. A tiny fluctuation in BABIP, some calls at the plate not going Lester’s way, and some even worse luck in wins would have shaved five dollars off of Lester’s value. It is not uncommon to see two projections like this in the spring that are inherently similar but present a radically different idea of a player’s value. If I simply leaned on one set of projections (whether they were mine or someone else’s) I could be fooled into the false idea that Lester is worth $20 or $15 and get locked into what is in reality merely one possible outcome.

Instead, the first thing I look at in attempting to come up with a baseline for player bids isn’t what a projection says but rather what the player has earned in the past. While this may sound akin to trying to fit a conclusion into a thesis, it’s really an attempt to do is recognize what the true earnings baseline is for a player. Too often, I see auction players bid past a player’s possible earnings ceiling. While there are a select handful of cases where this could make sense, in many instances it is because fantasy players don’t have a fundamental grasp of earnings theory and aren’t cognizant that they are proffering a poor bid. This does not mean that I will never set a bid limit that is higher than what a player has earned in the past, but I like to use prior earnings as my foundation.

Once I have done this, there are other factors that I will look at, including but not limited to the following:

Position Scarcity: This concept is overblown in some circles, but at a minimum in two-catcher leagues I like to push my bids up a dollar or two for the top 10-15 catchers. I also like to add a little money to elite middle infielders, particularly in only leagues.

Proven Stars: For elite players, I will bump their bid limits up based on reliability. Mike Trout only earned $32 last year in AL-only 5×5. However, even in a down year Trout was a money-in-the-bank $30+ player. My bid limit for him in 2016 will reflect this and be much closer to $40 than what he earned in 2015.

Category Bias: I do not recommend dumping stolen bases or saves in your auction. However, I don’t want to pay par value for steals and I especially do not want to pay par value for saves. Paul Goldschmidt and Dee Gordon both earned $41 in NL-only 5×5 in 2015. I am far more likely to set a $41 bid limit on Goldschmidt than I am on Gordon in 2016 (although it is extremely likely that both players will have lower bid limits).

Rookies: I always hedge my bets with rookies, especially in redraft leagues. It will be particularly tempting in 2016 to look at 2015’s rookie class and bid aggressively on the next crop of rookies. This is because memories are short and people tend to focus on past success rather than past failure. Kris Bryant earned $28 last year and turned an $11 profit. His teammates Arismendy Alcantara and Javier Baez cost a combined $22 and earned $1. Guess which of these two stories is going to be pushed harder by the fantasy analyst community this coming spring?

Part-timer bias: While a part-time player can earn money in deeper leagues, there is playing time risk to consider in addition to performance risk. It is possible for a part-time hitter to earn $10-15 in deep league formats (tip your cap, Franklin Gutierrez), but even a small swoon for a player with 150-200 plate appearances can have a significant impact on his value.

Injury history: I am not a big believer in the idea that a player is “injury prone” or that past injuries will lead to future injuries. On the other hand, with all other things being equal I will lean toward the $20 player with a track record of health over a $20 player who has missed chunks of the season for the last two years.

Beyond what a player has done in the past, the second most important factor in setting my bid limits revolves around market value. This is why all of the factors above are taken into consideration and why in addition there are the additional steps of seeing how the expert market is ranking players throughout the run up to Opening Day as well as what the various evaluation models are projecting.

To use an obvious example, the best middle relievers frequently exceed the earnings of the bottom third of starting pitchers. However, because most major league bullpens are seven men deep, middle relievers are a dime a dozen, both in auctions and in the free agent pool once the season begins. Michael Blazek could very well earn eight dollars in NL-only once again in 2016; however, there is no reason for me to pay more than $1-2, as there are several middle relievers floating around in the free agent pool who could earn eight dollars this year.

I want my bid limits not only to “predict” what I think a player is worth but to also be a realistic reflection of the marketplace. Last year, I was enthused about Lucas Duda’s fantasy potential and wanted him to be on as many of my fantasy teams as possible. However, after seeing that the market was far less excited than I was about Duda I moved my bid limit down from $24 to $22 in NL-only. Sure enough, my price for Duda was still aggressive and I purchased or drafted him in all but one of my leagues last year.

Another way of explaining this is that I do not want a bid to be so far out of line with the market that I fool myself into believing that I have a bargain based on an unrealistic bid. If after listening to the most recent edition of Sahadev Sharma and Brett Taylor’s award winning¹ Limited Range podcast I am convinced that Jason Heyward will be an offensive monster and earn $35 in fantasy in 2016 this is fine, but if I see everyone else bidding in the $23-25 range for Heyward there is no need for me to set a $35 bid limit. $30 should be more than sufficient. I do not want to take precious auction resources away from other players where there will likely be more competitive bidding by placing a bid on Heyward that is out of line.

A Note on Metric-Driven Data
I will take metrics into account when it comes to setting bid limits on players. However, I will generally err on the side of caution and try not to overemphasize a specific metric and overthink a bid limit. As an example, if a hitter’s BABIP for the last three years were .275, .303, and .294, it is unlikely that I would put too much weight into the .275 result unless there was some other trend in the data that I could specifically pin down. There is always the danger of overanalyzing underlying metrics and unless there is a clear flag I do not want to overthink a player’s value.

Once I am done considering all of the factors above, I will start putting together my bid lists.

As an example, here is how my process would work with Todd Frazier.

Todd Frazier will be a 30-year-old third baseman for the Reds in 2016.

Table 1: Todd Frazier NL-only Earnings Scan, 2013-2015

Frazier was the third best third baseman in NL-only 5×5 in 2015 (behind Nolan Arenado and Kris Bryant). His scan shows that while he did slip in earnings in 2015, his power wasn’t a fluke and while his steals did drop off as many predicted they didn’t disappear entirely. His batting average did drop, and judging by the scan Frazier is more likely to hit .240 then he is to hit .280. The Reds haven’t entirely gutted their team, but if they do go this route Frazier will lose some RBI and run opportunities.

My first pass bid on Frazier is $24. If this seems low for a hitter who averaged $28 in earnings from 2014-2015, it is because my initial bids typically err on the side of caution. I will have opportunities at the tail end of my process to add or subtract bid value from the player pool.

I will go through this process for all 30 major-league teams. I will assign bid limits to the projected Opening Day lineups, benches, rotations, and most of the bullpens. In leagues where you can bid on minor leaguers, I will assign bids or draft positions to some minor league players as well.

Once I complete this process, I will tally up the total bids in each league. My bids should add up to $3,120 for the hitters and pitchers combined in 12-team only leagues, and I like to generally have somewhere in the neighborhood of $2,160 allocated for hitters and $960 allocated for pitchers. This is league dependent and is more suited toward expert leagues. In my home leagues, my experience is that $1,020 or so is spent on pitchers so I will adjust accordingly. Again, I want my bid limits to reflect reality. A $100 difference between the hitters and pitchers may not sound like a lot across an entire player pool but you can easily run the risk of purchasing an imbalanced squad if your bid limits are too far off in one direction or the other.

After my first pass, I am typically anywhere from $50-100 above or below the $3,120 cap. This is where I begin tweaking and really diving into my pricing. If I have two pitchers with a $15 bid limit, I will move a pitcher down by one dollar based on player preferences. I will do this across the player pool until my bids add up to $3,120.

If I were simply building bid limits for my own auction, this could very well be the end of the process. But since I am publishing bid limits in late February or early March my next steps are to tweak. And tweak. And then, after that, to tweak again.

What I am attempting to do with my tweaks is adjust to realities in terms of projected playing time changes and market value. What I do not want to do is overthink my bids and make adjustments $1 in one direction or the other simply on a whim.

My goal with all of this preparation is to enter my auctions with a comfortable frame of reference to build a fantasy team. The bid limits are not designed to be a rigid framework that I cannot deviate from during an auction but rather a reliable frame of reference. If I have purchased two closers at solid prices early in the auction, I will make a mental note to divert money from other closers at pay a little more for other categories. If I do not like the prices for steals early, I might decide to toss the category in the auction and push for a nine-category team.

The goal of my bid limits is not to trap myself within a dogmatic framework but rather to allow for a quick reference during the auction so that I do not bid past my comfort level. When using my bid limits, you should feel free to adjust them to your auction conditions and as you otherwise see fit. I don’t have the “right” answer any more than you do and if you believe I am off base on a player, by all means adjust his price. Just make sure that your bids add up to $3,120 for your league pool. Regardless of whether you are using your framework or mine, you want to enter your auction with a decent handle on reality.

^{1 – As far as I am aware, Sahadev and Brett’s podcast has not captured any awards. But it has captured… my heart.}