Exploiting Market Inconsistencies to Win Fantasy Baseball

Most fantasy baseball players, including most fantasy baseball experts, don’t know how to determine a player’s actual value.  For example, which of the following two players do you believe is more valuable?

  • Player A:     85 Runs, 25 HR, 100 RBI, 20 SB, .250 average
  • Player B:      90 Runs, 10 HR, 65 RBI, 30 SB, .290 average

When assigning value to players, we tend to do so through a vague system of guesswork and hunches.  At best, we might use some underlying philosophies about fantasy baseball to inform our judgment.  My guess is that most readers would choose Player A as having the most value based on the belief that power is more valuable than speed.  At the same time, I’m sure there are at least a few managers scared off by Player A’s .250 batting average.

Until a few months ago, I didn’t know how to determine which player was more valuable, but I did have a general sense that the fantasy baseball community tended to overvalue certain categories at the expense of others.  Last season, I designed a draft strategy that aimed to exploit what I perceived as inconsistencies in categorical values, and I tested my strategy in three leagues.  For my efforts, I was rewarded with three fantasy championships and enough money to buy one and a half Chocolate Lionel Richie Heads.

Nearly as intrigued by my strategy’s success as I was the prospect of living in a world where you can buy a chocolate Lionel Richie head, I headed into this fantasy offseason determined to look deeper into the world of player valuation methods, a search that proved to be more or less fruitless until I came upon Larry Schechter’s fantastic book, Winning Fantasy BaseballIn his book, Schechter breaks down how the fantasy industry determines player value as well as how to calculate your own values.

As it turns out, the fantasy baseball market does an extremely poor job of capturing a player’s real value toward your fantasy team.  We all know that winning one category is no more or less valuable than winning another category, yet we still somehow manage to drastically overvalue some categories while devaluing others.  In the end, the market is so out of whack with reality that Schechter more or less disregards the idea of trying to determine a player’s real value because it’s not practical.  “When it comes to value formulas,” he writes, “being practical is more important than being right.”

While I agree with Schechter that being practical is more important than being right, this doesn’t mean there’s nothing to be gained by understanding real value.  Quite simply, if we understand where market value is inconsistent with real value, we can exploit these inconsistencies to maximize value on our team.  In this article, I will lay the groundwork for you to do exactly that.  In the following section, which is titled, “How I calculated projected real value (PRV),” I will explain how I determined player value.  If you’re more interested in what you can learn from real value than you are how to determine it, you can skip ahead to the section titled, “Examining projected real value,” where I will identify which categorical values are overvalued and undervalued.  Then, in the section titled, “Exploiting market inconsistencies,” I will briefly explain how you can use this information all of this information to win a chocolate Lionel Richie head of your own.

How I calculated projected real value (PRV)

As Schechter explains in Winning Fantasy Baseball, there are three basic approaches to calculating player values. You can calculate Standings Gain Points (SGPs), you can calculate value using a Percentage Value Method or you can use a Standard Deviation Method.

Schechter prefers SGPs because it is the most practical. In this method, you use historical league data in order to determine how many points each player will help you gain in the standings (thus, the name standings gain points).  For example, if you typically move up one place in your league standings for every seven home runs, a player who hits 35 home runs would be worth 5 SGPs. SGPs are logical and fairly easy to use, and as Schechter says, they are the most practical.  Unfortunately, the reason SGPs are so practical is that they mimic the inflation and inconsistencies that exist on the market.  What we want to do is remove inflation, leaving us with percentage value and standard deviation methodologies.

When you use determine player values using a percentage value method, you calculate the percentage of the total projected statistics that a player will acquire.  For instance, if you project that your league will have a total of 1000 saves, and you project Craig Kimbrel will have forty of those saves, then Kimbrel’s percentage value for saves is 4% (40 out of 1000).  The standard deviation is a very similar approach where you use deviation from the mean rather than percent value to determine a player’s worth.

For counting stats such as saves and RBIs, percentage value and standard deviation methods make a ton of sense.  Unfortunately, you can’t use the same approach for batting average, ERA, or WHIP.  According to Schechter, those who use PVM calculate batting average values by subtracting a player’s batting average from the median average for the player pool, and then multiply that product by the player’s at bats.  Quite frankly, this seems like one of the most arbitrary ways to calculate batting average values imaginable.

I took a different approach.  Instead of focusing on batting average, I used a players’ batting average and his at bats to calculate how much each player would affect a team’s final batting average.  For example, if you had a team of .230 hitters and then you added Mike Trout to the roster, how much would he improve your hypothetical team’s batting average?  If he raised the average from .230 to .255, I gave him a batting average effect score of 25 points.  After converting each player’s batting average to his batting average effect, I had transformed this category into a counting stat, and I could then treat it the same way as runs, home runs, RBIs, and stolen bases.  (It’s worth noting that this isn’t an entirely revolutionary idea; this is the same approach we use to calculate SGPs, I’ve simply applied it to the percentage value method.)

The last thing I did in order to determine a player’s real value was remove artificial inflation.  For reasons unknown to man, every fantasy baseball ranking system assigns somewhere between 66% and 70% of the total league resources to hitters while leaving the remaining 30% to 34% of resources for the league’s pitchers. Think of it this way: for a 12 team league with a $260 salary cap, the total salary available to be spent in your auction is $3,210.  When the fantasy baseball industry ranks players, they assign 70% of that money, or $2,184, to the hitters in the player pool, and they disperse the remaining $936 amongst the pitchers.

The question you are hopefully asking is why would anyone do that?  Pitching categories and hitting categories are weighted the same, so you could argue that you should spend 50% of your resources trying to win pitching categories and 50% of your resources trying to win hitting categories. You could also make the argument that you should spend roughly 57% of your salary on hitters because 57% of the starters in standard leagues are hitters (55% if you play in Yahoo leagues).  You could even argue that pitchers are at a higher risk of injury than hitters, so perhaps you should shift a little value towards hitters to offset injury risk.  But 70%?

While there is no reason to do this from a mathematical standpoint, we are in a situation where we now more or less have to do it.  To understand why, simply imagine what would happen if you used a 50/50 split in your next draft.  You would end up valuing pitchers far more than all of your league mates, and you’d end up with a team completely dominant in the pitching stats and completely atrocious when it came to hitting.  While using a 50/50 split isn’t practical for your next draft, if we want to gain insight into player’s real value, all categories must be weighted the same.  For this reason, I used a 50/50 split when projecting real value.

Examining projected real value

In the last section, I explained how I calculated what I call projected real value, or PRV.  I use the term real value because I am trying to suggest a player’s actual value in a world free from categorical inflation/deflation.  Simply put, we pay a lot more for some categories than others, and PRV allows us to look at a player value when all categories are valued equally.

That said, please understand that PRV is not to be used as a big board.  I don’t use PRV to draft a team, and I would never recommending doing so.  PRV is valuable as a way to identify trends in the way we overvalue and undervalue certain statistics, and in my opinion, its value starts and ends right there.

It’s also important to note that all of the rankings you are about to see were based primarily off of projections I made in early February.  Don’t put too much stock into who is ranked where – all you need to see are general trends in the rankings, and those trends should be clear enough.  Take, for example, the top ten players based on PRV:

  1. Craig Kimbrel
  2. Kenley Jansen
  3. Greg Holland
  4. Mike Trout
  5. Koji Uehara
  6. Clayton Kershaw
  7. Trevor Rosenthal
  8. Dave Robertson
  9. Addison Reed
  10. Joe Nathan

You don’t need to look at that list more than once to notice the most striking trend. Eight of the top ten players according to PRV are closers.  But every serious fantasy baseball manager knows that you shouldn’t pay for saves, so how can this be?

To make sense of this, consider that after creating my player projections for more than 400 players, my player pool had roughly 4,000 home runs, 2,000 stolen bases, and 1,000 saves. Stolen bases, in other words, were roughly twice as scarce as home runs, and saves were twice as scarce as stolen bases. Mathematically, this means that a player who had 30 saves would be twice as valuable as a player with 30 stolen bases, and a player with 30 stolen bases would be twice as valuable as a player with 30 home runs.

Or think about in another way: if a typical 12 team league accumulates 1,000 saves, 2,000 SB, and 4,000 HRs over the course of a season, the average team in the league would have 83 saves, 167 SB, and 333 HR.  We tend to value 30 HR hitters very highly, but if the average team hits 333 HR, a 30 HR hitter isn’t even giving you 10% of the total number of home runs that your team needs to be average.  At the same time, if a player saves 42 games, such as my projection has Craig Kimbrel doing, he would account for more than 50% of the saves that your team needs.  When you look at categorical value in this way, it starts to become clear how much we undervalue saves.

From a mathematical standpoint, it would actually make sense for the beginning of a fantasy baseball draft to be as heavy on closers as your fantasy football draft is on running backs.  Still, be careful not to read too much into it.  Just because saves are incredibly undervalued isn’t a reason to grab Craig Kimbrel, Kenley Jansen, and Greg Holland in every league. The mantra “Don’t pay for saves” still applies, and it applies because you can always find great bargains on closers late in drafts and auctions.  Think of it this way: instead of spending $20 on Craig Kimbrel, you might be able to build a staff of guys like Steve Cishek, Jose Veras, John Axford, and Tommy Hunter for around $15.  Even though Kimbrel is undervalue, low-end closers are being undervalued by a much greater margin.  Therefore, if you want to maximize value, you should target the John Axfords of the fantasy landscape.

At this point, I think we’ve dissected saves enough to demonstrate that closers are wildly undervalued by the fantasy baseball community.  Let’s now remove closers from the picture altogether and look at PRV again.  Doing so allows us to examine the top ranked hitters and starting pitchers.  Examine the following breakdown of the top thirty remaining players, and see what trends you can identify from this list:

  1. Mike Trout
  2. Clayton Kershaw
  3. Miguel Cabrera
  4. Jacoby Ellsbury
  5. Billy Hamilton
  6. Andrew McCutchen
  7. Madison Bumgarner
  8. Alex Rios
  9. Carlos Gonzalez
  10. Hanley Ramirez
  11. Yu Darvish
  12. Carlos Gomez
  13. Paul Goldschmidt
  14. Adam Jones
  15. Everth Cabrera
  16. Chris Sale
  17. Adam Wainwright
  18. Edwin Encarnacion
  19. Jean Segura
  20. Jose Fernandez
  21. Justin Verlander
  22. Stephen Strasburg
  23. Chris Davis
  24. Jason Kipnis
  25. Hunter Pence
  26. Max Scherzer
  27. Ryan Braun
  28. Cliff Lee
  29. David Price
  30. Ben Revere

There are three things worth noting here. First, you can see the clear influx of base stealers in the top thirty. Stolen bases, as it turns out, are undervalued for the same reason that saves are undervalued – they are a scarce commodity.  My February projections for Billy Hamilton were for 65 R, 4 HR, 40 RBI, 66 SB, and a .242 average. In other words, Hamilton doesn’t check in as the fourth most valuable hitter because I have wildly inflated projections for him – he projects as the fourth most valuable hitter because stealing 60 bases would instantly make your team competitive in the category.

The second trend worth noting is the absence of power hitters on this list.  Edwin Encarnacion and Chris Davis are the only two hitters who cracked the top 30 with projections for less than 12 stolen bases.  Are home runs valuable?  Yes, but they are far less valuable than we treat them.  Adrian Beltre, who I projected for 80 runs, 28 HR, 100 RBIs, 1 SB, and a .308 average.  28 HR is the 14th most of any of my projections, and 100 RBIs ties him for the 5th most.  Yet Beltre’s inability to steal bases makes him the 27th most valuable hitter according to real value.  Simply put, we overvalue HRs and RBIs.

The third and final trend worth noticing is the influx of starting pitching on my rankings for projected real value.  On Fantasy Pros, there are currently six pitchers ranked in the top thirty overall players, but our rankings have eleven pitchers in the top thirty.  This time, the difference between real value and market value stems not from the economics of scarcity, but from removing artificially adjusted pricing.  Recall that the fantasy baseball industry assigns 70% of all value towards the hitting categories, thereby inflating the value of hitters on the fantasy baseball market.  When you remove this adjustment, pitchers naturally become roughly as valuable as hitters.

Exploiting market inconsistencies

Knowing that categories like HR and RBI tend to be overvalued, while saves, stolen bases, and starting pitching are undervalued can be extremely helpful if you know how to exploit these inconsistencies, but how you exploit each of these categories will depend heavily upon what type of league you play.

In rotisserie leagues, you can exploit inconsistencies in categorical value by choosing where to spend the majority of your resources.  Many fantasy pundits preach the need to target power early, and they suggest this because power dries up quickly.  But ask yourself why power disappears so quickly on draft day?  We know there are roughly twice as many home runs in a typical mixed league than there are stolen bases, and there are roughly four times more home runs than saves.  So why can we find players projected for 25 saves littered across our draft board in round ten, but most of the 25 HR hitters have been taken long ago?

When you construct your team, you have a decision to make about where you will allocate your resources.  You can target categories which have inflated costs, such as home runs and RBIs, or you can target categories which come at a discount, like saves and steals.  In rotisserie leagues, it’s very risky to punt any categories, and because of the correlation between HR and RBIs, punting HRs would be a foolish endeavor.  At the same time, if there are two categories that you should avoid trying to win, it would be HR and RBIs.  The best approach would be to build a team that is in the middle of the pack in the power categories, while trying to win the categories that will cost you the least resources, such as stolen bases and saves.  To do this, simply bypass one or two of the overpriced power hitters that you would usually draft and invest your savings on players that will assure your team finishes at the top of the pitching categories and stolen bases.

While I’m confident that you can exploit market inconsistencies to help you in your rotisserie league, I believe that exploiting market inconsistencies in head to head leagues is an absolute game changer.  If the rest of your league is driving up the price of power hitters, the obvious solution is to punt HR and RBIs and to spread the entirety of your resources across the eight categories which are being undervalued.  While the rest of the league is sinking all of their money into the two categories which come at the greatest expense, you will be investing your money in the categories which come at the greatest discount.

This is exactly what I did last season, and as I stated in the beginning of this article, it was an extremely effective approach netting me three championships in three tries.  While I’m not nearly so naïve to suggest this strategy will always work, I believe that it gives you a clear upper hand against your opposition.

Still, a word to the wise.  Don’t think this approach means that you don’t need to do any research or that you can enter your draft without a plan.  There are many pitfalls that could derail your team, so in my next article, I will break down how to punt home runs and RBIs in detail.  I may have won all three of my leagues where I tried this approach, but I prepared heavily before hand and I learned a lot of lessons on the way.

In the meantime, here is a quick peak at the last team I drafted using this strategy.  I still have some work to do to flush this team out, but this gives you a general idea of what such a team would look like.

C – Carlos Santana
1B – Billy Butler
2B – Dustin Pedroia
3B – Aramis Ramirez
SS – Jean Segura
OF – Shin-soo Choo
OF – Norichika Aoki
OF – Ben Revere
U – Kolten Wong
U – Adam Eaton

SP – Clayton Kershaw
SP – Stephen Strasburg
SP – James Shields
SP – Hyun-jin Ryu
SP – Dan Straily
SP – Doug Fister
SP – Spot reserved for streaming pitchers
RP – Kenley Jansen
RP – Greg Holland
RP – John Axford
RP – Matt Lindstrom
RP – Latroy Hawkins
DL – Bobby Parnell