What is Position Scarcity, Continued
I first want to thank everyone who read What is Position Scarcity, Really? and commented, as it led to an interesting discussion with lots of good points and observations. I quickly realized how in-depth this concept could go and that I may not have explained myself as well as I would have liked. It happens all too often, as I sometimes suck at clearly stating my argument and finding the words to use so that everyone understands what I am trying to convey. I have been emailing with a reader/listener of my radio show who read the aforementioned post and it was evident that I had not done a good enough job convincing him either. Luckily, as I was typing my response to him, it suddenly hit me, as I thought of something I should have included in my original article, but failed to.
Most of the argument against taking catchers early or the concept of position scarcity in general is that the fantasy owner is much better off with a 30 home run, 100 RBI first baseman as opposed to a 20 home run, 80 RBI catcher in the early rounds. “Why would I draft a catcher early when I can get a first baseman who is better in every category? It is about maximizing team output, not any specific position’s output”, the anti-camp says. Well, of course, in a world where you have a roster of 14 utility spots, that would be a perfectly sound argument. However, I have never heard of such a league format. Fantasy leagues require that we draft a catcher that actually catches (or qualifies at the position) and a first baseman that qualifies in that slot.
For some reason, the anti-camp seems intent on comparing the raw stats of a hitter at one position (like catcher) to another position (such as first base). This is a mistake and appears to be one of the primary reasons why fantasy owners misunderstand the idea behind position scarcity.
Let’s use an example. You’re in Japan and find an item that you want to buy and need to determine if it is cheaper there or in the U.S. You notice that the price is 15,410 Japanese yen, but have no idea what the exchange rate is. Do you think that just because the item requires you to pay significantly more yen than you would need to pay in the U.S. using dollars, that the item is more expensive in Japan? I would hope not! This is no different than comparing a catcher’s raw stats with a first baseman’s. Just because a first baseman is projected to produce better in all categories than a catcher, doesn’t automatically mean that his projected value is higher.
Without the exchange rate, it is impossible to compare the price in dollars with the price in yen. What needs to happen is an adjustment based on the exchange rate to either the yen price or the dollar price, to put both numbers on the same scale. In fantasy baseball, an adjustment needs to be made to the projected stats of every hitter. In the valuation system that I use, this adjustment would require that you take a replacement level hitter (because you should only be paying for production that is better than what you can get freely out of the free agent pool) at each position and subtract those projected stats from the projected stats of the rest of the hitters at each position. After adjusting your projected stats, then finally, you are comparing apples to apples and can accurately compare Mark Teixeira to Carlos Santana. Without this adjustment, you are truly comparing apples to oranges, and will not be properly valuing players.
If a catcher and first baseman’s projected stats were identical, you would always draft the catcher first, or pay more for him, right? That’s easy to answer, but by admitting this, you are acknowledging that position scarcity does in fact exist and bumps up the values of the hitters at that position. Where it gets difficult is figuring out exactly how much to increase the value of a player at a so-called scarce position. Unfortunately, this cannot be done in your head and requires a full-fledged valuation system. This will quantify exactly how much of a jump a player at a scarce position experiences and there will no longer be a need to guess at which point the superior stats from a first baseman versus a catcher trumps the effect of position scarcity.
There is a whole lot more that can be discussed, such as size of tiers within a position, drop-offs between tiers, etc, and how that affects player values and draft strategy. This kind of stuff is also very interesting and is worthy of its own posts, but the concepts are different than position scarcity as I have defined in this and my previous post. So keep that in mind when commenting!
Mike Podhorzer is the founder of ProjectingX IQ, an advanced fantasy baseball analytics platform that transforms projection data and in-season performance signals into actionable intelligence. He is the 2015 Fantasy Sports Writers Association Baseball Writer of the Year and three-time Tout Wars champion. He is the author of the eBook Projecting X 2.0: How to Forecast Baseball Player Performance, which teaches you how to project players yourself. Follow Mike on X@MikePodhorzer and contact him via email.
“Do you think that just because the item requires you to pay significantly more yen than you would need to pay in the U.S. using dollars, that the item is more expensive in Japan? ”
Should be “valuable” not “expensive”. But the point is valid cost!=value.
I like the exchange rate analogy, and the method you discuss is a useful one. I use it in a linear weights points league to great effect every season. I’m seemingly the only owner in our league who has realized that elite pitchers are the most valuable commodities in the league.
Brad – sounds interesting. Where do you play in that advanced of a league? Would love to join.
An ‘advanced league” that doesn’t realize how it’s own player pool is valued? :p
The league is advanced not the owners
Mike, I think your 30hr/100rbi 1b vs 20hr/80rbi catcher is kind of a straw man argument. I think the “anti” camp that you are referring to is saying they’d prefer the combination of a top 1b (.300/30/100) and a weak catcher (.250/15/65) than a weak 1b (.280/22/85) and a strong catcher (.290/22.90), especially given the higher degree of uncertainty and injury risk at the catcher position. Also the HR/RBI upside for a top 1b is much higher.
I was merely giving an example illustrating that just because a first baseman is projected for better stats in each category does not mean he is more valuable to a fantasy team, ignoring all other variables.
In your example, assuming the stats of the two pairs were exactly the same, then the overall projected dollar values of the pairs should be the same. However, I would absolutely agree that I would still rather the pair with the top first baseman. You hit it on the head with your reasoning, but this is not an argument against scarcity.
So just draft the biggest outliers at every position
So do you think this “exchange rate” can be defined in some sort of number valuation? I guess there would be a lot of external factors determining relative value of positions due to league settings, which would make it difficult.
I wasn’t sure in your conclusion whether or not you were planning to follow up on the exchange rate concept in terms of explicitly coming up with an actual rate.
Well, it then gets into valuation systems and methods used to calculate dollar values to use for your drafts. That stuff gets very technical, boring, and Excel-y, and I simply follow instructions I have from a system developed like 10 years ago. So I am not sure it makes sense to discuss the system in a separate post, especially since I didn’t even develop it.
My issue arises in leagues with utility spots, or CM or MI spots. The replacement level system says that the last player picked at a position is essentially your replacement level for that position (perhaps it is the best player not picked, but the same issue arises regardless). But a utility spot confuses this. Imagine a twelve team league that has one SS, one 1B, and one U spot. Let’s say you have six first basemen and one shortstop who are good enough to roster in the utility spot after filling the 12 SS and 1B positions. But the sixth utility first basemen is actually worse than the one SS, so the base value for a first basemen would be worse than the base value of a SS. But that is blatantly wrong. The multi-position spots greatly change this as far as I can tell.
When making a valuation system, how do you accommodate for this fact? The average production of a position, not just the replacement for that position, has to affect a player’s value. Is there a single replacement level that is used for all positions that you value a player based off of, and then a separate position adjustment you need to make? I’m making my own valuation system, and any help would be greatly appreciated.
Great observation. My long-running league has these positions as well and I haven’t figured out a perfect way of doing it. In my 12-team league, I generally just assume that the utility slot will solely be filled by 1B, 3B and OFers. Middle infielders are rarely good enough to be worth using at Util and obviously catchers never are. So I lower my replacement level for these positions to accomodate the extra players that will be drafted and starting.
For MI and CI, I just combine the two positions and so use the 37th 1B/3B as replacement, for example (12 1B, 12 3B, 12 CI). Actually, I don’t truly do that, because I then include Util players, and so usually end up with another 6 CI players at Util, so suddenly my replacement is the 43rd CI.
Send me an email with any other valuation system questions!
What you’ve got to do is apply replacement level as a fictitious concept and not as the last player picked. Pick a stat level below which you consider everyone fungible, and call that your baseline. If it so happens that more or fewer players fall on each side of that line than your league actually picks up, then so be it. Sometimes your league mates are going to pick up sub-replacement level players. That shouldn’t move your definition of replacement level, it just means they drafted poorly.
Also, you’ve got to rank each player individually at each position he’s eligible. That includes UT. You can’t scale the positional baselines against one another.
Putting Position Scarcity aside I’d like to just looking at the catching position. Every year I seem to find someone near the end of the draft or on waivers during the season that fills the void just fine, sort of the same way people find closers. Using one of my keeper leagues as an example, here are the catchers I either drafted or picked up each year: Avila (11), Y.Molina (10), K.Suzuki (09), G.Soto (08), B.Molina (07), K.Johjima (06), M.Barrett (2004).
With the exception of 2005 (where I played musical catchers off the waiver wire) I was able to land a catcher who finished in the top 10, using the same stratagy people use when picking closers late or grabbing a guy once the primary guy loses his job or gets hurt. Point being you can get a top 10 catcher for cheap if you do your homework and pay attention, the same thing can’t be said for first base.
If you look at the final stats for first basemen at the end of the season, more than likely you are gonna see all the guys who were drafted early sitting on top of the leader board (baring an injury). Now do the same thing for catchers, but the difference is there are a number of guys up there with the McCanns & Mauers of the world that weren’t even in your draft plans (just like with closers).
You’re right, you can’t compare the raw stats of a first baseman to a catcher, but you can compare catchers and using last year as an example:
Catcher A) 22 HR’s, 68 RBI’s, 72 Runs, .262 Ave
Catcher B) 19 HR’s, 82 RBI’s, 63 Runs, .295 Ave
Catcher A is Matt Weiters who was I’d say on average the 6th catcher taken off the board. Player B was Alex Avila, someone who was taken very late (if at all) in most leagues.
Now look at the top 10 first baseman for last year: Cabrera, A.Gonzalez, Pujols, Fielder, Votto, Teixeira, M.Young, Berkman, Konerko, Howard. The top 6 were all more than likely gone by the end of the first round if not early in the second, and I can include Howard in that group. If you were left with a concellation prize of Konerko a few rounds later you weren’t upset. The only surprise here was the monster year Berkman put up, and while he wasn’t a top choice he was still drafted in the middle rounds. No late round or waiver pickups in this group.
You can make a case for taking a catcher early, but I can easily make the opposite argument for not taking one. For the rest of you, it all comes down to what you are more comfortable doing. I say do a little homework and wait (just like with closers), there is a catcher out there for you.
Yes, this is a great point. You probably get more surprises in the Top 10 catchers at the end of the year compared to other positions, and if you are in a 1-catcher league those surprises are available as FAs. But the position becomes a lot more “scarce” in a 2-catcher league. That’s where it makes a lot of sense to spend heavily at catcher. (Though I’m just not sure how many leagues still use 2 catchers.)
What I find is that occasionally there is a catcher who has a great year (a la Mauer 2009) who seems to warrant being picked alongside the top 1B because he is so far ahead of the rest of the catchers. But I think it’s rare for catchers — with injuries and days off — to put together great seasons with reliability. So I drop my expectations for catchers way down, and while I still like to get a top 5 catcher, you have to be careful not to go crazy.
“For some reason, the anti-camp seems intent on comparing the raw stats of a hitter at one position (like catcher) to another position (such as first base). This is a mistake and appears to be one of the primary reasons why fantasy owners misunderstand the idea behind position scarcity. ”
This is a gross mischaracterization of the “anti” camp. It doesn’t matter if a position is “scarce” if there is not a huge divergence in values within the position. It also doesn’t matter if a position is “scarce” if the price in draft order or dollars is such that you’re turning down much better players. Again, it is a logical fallacy to think you will maximize a team’s output by maximizing output at putatively “scarce” positions. You must, absolutely must, consider the opportunity cost of the much higher actual level of performance a position player can be expected to generate.
Let’s take a look at a current ADP: http://rotochamp.com/
To get Mike Napoli you have to turn down Mike Stanton or Justin Verlander. To get Carlos Santana, you have to turn down Andrew McCutchen. To get Jesus Montero you have to turn down Starlin Castro. To get Buster Posey you have to turn down Jose Reyes, Jay Bruce, and Eric Hosmer. To get Brian McCann, you’re leaving Tim Lincecum and Josh Hamiliton.
So let’s say you do that. You decide you want Carlos Santana instead of McCutchen. Congratulations, later in the draft you just won the Alex Presley sweepstakes with Jonathan Lucroy still on the board!
What would you think of a trade of Santana and Presley for McCutchen and Lucroy? Put that way, the Santana-first draft strategy sounds like the truly terrible idea it is.
There’s no ADP data on RotoChamp yet. You’re just disagreeing with their system’s projections.
I like the concept of replacement value, and agree it applies rather well here. Basically, position scarcity occurs when the leaguewide required number of players at a position is greater than the number of positive contributors at that same position. For such positions, failing to buy a positive contributor means you are buying a negative value player, and even if the cost is only $1, you are still overpaying. So you should be willing to pay a premium at that position, up to a level equivalent to the cost differential between $1 and the value of a replacement player.
Specific examples always depend on league rules (for example catcher is not a scarce position in single catcher leagues), and under certain league structure/size combinations odd scarcities can occur (large leagues I’ve found can sometimes make closers be purchased at premium prices – a practice I find dubious).
My big money league is a 13 team 15 starting hitter live draft league, (8 standard, MI,CI,2 add’l OF, 3 UTIL), and with such a set up I find oddly enough that there isn’t a major scarcity anywhere, but the overall replacement level can be quite low and I think there is a inefficient premium put on B level players which depresses the cost of stars. This example has nothing to do with position scarcity, but dammit if you can’t ramble on in an internet comment, where can you ramble on anymore?
Sorry to reply to my own comment…
I disagree with the exchange rate analogy, I think it’s better to think of position scarcity effect as an additive premium and not a multiple.
For example, if you have $25 to spend on 2 players, and you buy let’s say Texiera and Saltamacchia ($22 and $3) or Santana and Ike Davis (say $18 and $7) you end up with a slightly better stat line with Santana and Ike Davis (at least by Bill James’s projections). Tex is obviously better value than Santana when we ignore position considerations (he’s better in all five categories), but at the end of the draft, when you’re trying to swing a guy who won’t hurt you you do much better on the 1B front than the C front.