Rationality Will Ruin You
Perhaps you’ve seen this tweet at some point over the offseason. It surfaced most recently about 10 days ago via Dave Cameron in his piece about the Dodgers’ intentional inefficiency.
Andrew Friedman: “If you’re always rational about every free agent, you will finish third on every free agent.”
— Andy McCullough (@McCulloughTimes) December 6, 2016
What Friedman says absolutely applies to fantasy baseball. Especially auction drafts. Every year, I see people saying “I’m going to stick to my values.” I’ve been that guy too. Way back when I was the only person in the room with advanced analysis, sticking to my values was a winning strategy. Now it’s a good way to lose every key player.
Auction drafts tend to have a certain cadence. A few big names are nominated early to set the tone. You look at your cheat sheet. You read Kris Bryant, $41. You look back at the live bidding. Bryant is up to $40. You +1. Someone else taps +1. Now he’s $42. You say to yourself, $43 is close enough. +1. Somebody $10 bids. Some more +1’s follow. Bryant’s auction ends at $57. This repeats for several elite players.
Now comes the grumbling phase of the draft. The chat room fills with smug comments about the “crazy” bids. Insane. Unbelievable. [Insert Adjective]. These owners think they’re going to reap the rewards later in the draft. Surely, the early aggressors will adjust their pace so they don’t run out of money within their first 11 picks. Right?
Bad assumption. The aggressors are aggressing because they understand the nature of your league. Let’s say it’s a 12-team, 25 player roster format – 300 players are rostered. That’s fairly typical. The last 75 players drafted will cost about $1. There are way more than 300 fantasy leverageable players in the majors. More like 600.
With proper preparation, an aggressive drafter in this format can overpay for high floor, high ceiling talent then make up ground by drafting the best of the rest. If you’re in a keeper league based on player cost, this approach is doubly valuable. Every $1 Jonathan Villar found becomes a franchise asset. Even a $1 Victor Martinez could fuel a playoff run.
I digress. We were speaking of the point in the draft where everybody says, “y’all’s crazy.” And the crazy continues. It has to stop any time now. Then you realize the best player left is Adrian Beltre. He’s nominated. You’re still sitting on your full $260 so you desperately bid and win Beltre for $34. You had hoped he’d fall through the cracks a bit.
It’s at this point you realize you have two options. Join the crazies for the mid-rounds or finish the draft with $50 to spare. And so the crazy rounds extend. Other “rational” bidders make the same discovery. The crazy gets crazier. The early bidders bow out with $9 and eight roster spots to go. They’ll have to wait awhile before there’s a chance to draft Villar. The rest of you are too busy paying $19 for Jose Peraza and $11 for Matt Strahm. Strahm was supposed to be a $1 sleeper.
More often than not, that’s the cadence of an auction draft. At least, that’s my personal experience with a sample of N>100.
A handful of years ago, I observed a first mover advantage – that is, the first couple players of the draft could be had for a relative discount. Say Mike Trout goes for $50 with the first pick. Then we see Miguel Cabrera for $50. Then Andrew McCutchen for $50. Ryan Braun for $45. Jason Heyward for $40. Soon, one owner is gleeful about his “bargain” Trout, another guy is satisfied with Miggy, and three more are grumpy about overpaying for not Trout.
Exploiting the first mover advantage was a great way to stay rational while ensuring a high baseline of talent. Last year, I found no sign of this advantage in any of my auction drafts. I suspect the market inefficiency has been corrected.
If the first mover advantage is gone, then we’re left with one conclusion – rational bidding is a recipe for failure. That leaves us with two options. We can set the tone for the insanity, throwing ourselves at Trout and Bryant and Clayton Kershaw. In many things, it’s often best to be the aggressive one. It forces your opponents into making mistakes. In a fantasy draft with constrained resources…well, it’s a calculated gamble.
Alternatively, you can sit back and react. Every once and awhile, an auction draft will be surprisingly civil. One hundred players will be picked at or near their value before anybody starts to panic. Taking a reactive approach is best in this type of auction. You would never have known it was an option if you go hyper-aggressive.
Whatever you do on draft day, be prepared to overpay for at least one top 12 caliber player. The alternative is to overpay for a bunch of top 150 players. Don’t worry, you’ll still get to do that too if you pay for the stud.
You can follow me on twitter @BaseballATeam
Excellent work Brad.
This has taken me years to finally figure out.
Decades for me.
Inherently this makes sense. After all, what is the difference between say Stahm and I don’t know, post hype Bruce Rondon. Anyone? When it gets past about 150-200 top players the difference between players becomes negligible and many of them will end up on the wire anyway. You can’t build a roster of ‘value’ picks from 150-300. If you think things might get very crazy though it may pay to focus on the ‘value’ high floor guys to overpay for and this will allow you to save some $$ through the middle rounds ($11 for Stahm?!? hahahah no) and basically have your choice of the bargain guys where your $2-4 bid will win almost whoever you want.
As someone who just made an aggressive trade for Kelvin Herrera, the suggestion that Strahm may be valued at $11 almost gave me a heart attack.
I find there is often a lull after the most aggressive bidding, and that’s the time the best bargains happen before the $1 round. I also find at least one person is sitting on enough $$ to snipe several $2-$3 bargains.
I used to find that, but not anymore… I think because many plan on that, and those “many” then find that they aren’t alone and have to overpay during what they expected to be an easy lull.
This is the best analysis of studs and duds that I’ve seen. Nice work! Last year, my 10-12 team redraft league (23 man rosters, but only 9 active hitter spots) converted to auction format for the first time. I bid aggressive on the first two guys, Trout (who I got for $45) and Rizzo (who I got for $29). After that, Harper went for $52, Correa went for $39, MadBum went for $42, etc.
I’ve also been reviewing my rosters at year end compared to post-draft and have noticed that only the top 5-9 guys end up staying on my team all year round (the others being waivers or trade acquisitions). This year we’re going keeper, and I definitely plan to spend whatever it may take to acquire at least 3-5 top 20 bats.
It seems in particular that it is always easy to find nearly a full staff’s worth of SPs to buoy your ace if keeping a watchful eye on the wire (and Rotographs!)
your approach works in a 10-12 team league, more pitchers to choose from later on. i’m in a 14 team keper roto and i know this year i’ll be paying for 2-3 top 20 hitters and 1 top 5 starter.
Then again, your “crazy price” is just another owner deciding that he expects the player to perform better than your projections. To him it might not be an overpay at all.
You should have a number of guys like that also, where your price valuation exceeds the consensus.
As I often find, you’ve again posted an article where I go “Damn, I hope the rest of my league doesn’t read this”.
There’s an interesting, informative article that could be written from this called “Poor Pricing Curves Will Ruin You”.
Every few years, an article is written about how a 6WAR player is worth more than two 3WAR players. This is the fantasy baseball version of that.
I fully agree for a standard 12 team league about paying for top talent. Does your approach differ at all for a deeper or only league where replacement level is much lower. Do you only pay more for one higher level guy knowing that fielding a team of 6-8 $1 guys is a high risk high reward proposition?
That’s a good point, in my limited “only” experience. Kind of the reverse of an earlier comment where I was shocked that Trout went for only $45 in a 10-12 team league with only nine active hitters. I’d think more like ~ $70 in the latter case.
Personally, I have more success finding small improvements in mid-tier players than large improvements in lower tiers. It seems too reliant on sniping the right $1 guys. I have always found my teams better having 4 quality $18 players and a single dollar dream than a $50, $15, $5, and 2 $1 tickets.
I think this is dependent on knowing your league’s tendencies. One keeper league I’m in I can take that strategy. Years ago a joined a different league, and it took me a couple of years to adjust to those owners’ style of aggression. The $18 players I was paying for were worth $10-$12. Thus, I lacked cornerstones to produce my points, and I rarely had great keepers for the following years from that group. My other keeper league those $18 guys I might have felt were worth $22 with upside. It just depends a lot on your fellow owners. I.e., be flexible.
That doesn’t work too well in our keeper league as those “mid tier” players typically have inflated prices and don’t go for $18….more like mid 20s. The $18 we pay are normally for $10 worth of stats.
I disagree. If you set the values of your top (however many players get drafted) to equal the total money available in your league with the bottom (however many you think should or will go for $1) all equal to $1, then your values will work out. If you properly value your players, for every dollar spent above what you think a player should cost, there will be another player that goes for at least* a dollar less.
*- at least because if there is money left at the end of the draft, that’s all additional money for players that went below what they “should”.
What you think players should cost doesn’t matter in an auction. The highest bidder sets the market value, and at a certain point you will be forced to “overpay” otherwise you will have no top-30 players. I initially thought like you but after a couple of auctions it was clear your theory just doesn’t work in practice.
My point is that if you do your valuations correctly, based on the available money in the league, you won’t have to overpay for a top 30 player. The problem is in people’s valuations, not in their aggressiveness.
Not completely true. If you stick to your valuations, you absolutely will get bargains. However, the vast majority of these will be in the endgame, and you’ll end up with a team full of guys you think are worth $5-7 that you paid $1-3 for (and $150 left unspent).
To give a simple example, your team (worth ~$150) will be much worse (even by *your* valuations) than the $210 team of the guy who spent $170 on 4 studs that you think are worth a combined $120, and $100 on the rest of his team that you think are worth a combined $90.
(That’s just a long-winded way of saying you have to spend your money, and if you do your valuations correctly and stick to them looking for bargains, you won’t.)
There’s no reason that your league won’t overpay on the top 30 (or 50, or 100) players. In fact, I think it’s very likely that the top 30-50 will be overpaid. And holding back on those guys, looking for slight bargains in the 50-100 range is a VERY risky strategy.
Sorry, slight error in my math above. That should read “$160 on 4 studs that you think are worth a combined $120”.
In my values list currently, of the top 30 players I have 3 that are projected to be “worth” $10+ higher than their average auction price, and four more that are $4-$8 higher than their average auction price(per yahoo where my leagues are). Now, it is theoretically possible that all 7 of those players in the particular auction I am in will go higher, but it is unlikely. In a typical auction, you will be able to get at least 2-3 of those players at less than their “worth”.
If your top 30 player values are all below what players are actually going for, then what good are your values doing you? At that point isn’t that just a ranking?
I guess the issue is with the word overpay. I’m saying that if that’s what the players are going for, the rest of the league isn’t overpaying, you are just undervaluing the players.
Absolutely nailed it, seems like every auction I’m in goes like this. After the star inflation followed by young player/prospect inflation, you can clean up on bargain vets. If you don’t join in it is akin to skipping the 1st 3 rounds of your draft and joining in with extra picks in rounds 4-6. Who wants to do that?
I’ve always found there are good bargains if you nominate players that are 10-15th on the depth chart of currently available at that position. For instance, you like Jake Lamb, put Jake Lamb up for bid in the 1st round.
Most people will be rational about his price, because there are 12-15 3B they like more.
That’s the best way to remove the overbidding… ‘removing’ the scarcity
What I gather from this article is you better know who you’re drafting against. Or, be very confident in your strategy… that is where being the aggressor will over the length of a draft pay off. I know of many who bid up players with no intention of rostering them. Leaving bargains as others have to wait on those $1 marginal players. For my back pocket when I consider an auction draft.
Great stuff! Auctions are amazingly fascinating when there are no auto-bids. Where is a good place to do some mocks this early in the year?
I’ve found that you’re best bet is to go into the draft with a team more or less “built”. In other words, you know exactly who you want, and how much your plan changes depending on how ‘crazy’ the bidding gets on your “must have” players. For example, if you end up paying $60 for Betts instead of $55, then you adjust your $25 Cueto to a $17 Tanaka. So on and so forth. Granted, the draft doesn’t always play out in any semblance of ‘order’ but if you have a list of players at each position that include multiple tiers, you’ll never find yourself bidding on a player you really don’t want just because you feel you have to. Obviously it won’t play out like this for the entire draft, you’re not going to get every single player you want, but if you fill 75% of your starting roster with players you feel comfortable with at prices that are more or less within your value range, you’ll find that you’re still in great shape towards the end. Its the old adage of “plan for the worst and hope for the best”. If you plan on having to overpay for every player you truly want, then you’ll never be disappointed, and you’ll have more freedom come the end of the draft if things go better than expected. Anyways, after years and years of auctions, I’ve found this is really the only way to go about it.
Its really like Brad said, NEVER pigeonhole yourself into refusing to overbid on players that make your team work.
Couldn’t this be a result of setting replacement level too low? Doing so will compress the values of players and result in overvaluing guys close to replacement value.
I agree, mostly. But there is an additional consideration regarding how to react if the league is overbidding the top 30 players in a huge way. At some point, you need to jump in, unless you think the trade market is so liquid that you can convert great depth into some top players.
I’ve seen this quite a bit in my main auction league recently with bids going well into the 40s and 50s for the top players (8 keepers, add $5 to price per year). I’ve had a lot of success the last few years by banking on my keepers at reasonable prices and then staying out altogether. I don’t think I won a player until about the 30th guy came off the board after keepers but ended up taking 10 of the next 15 players and absolutely dominated the middle of the draft. Each league and draft is different, but this is a strategy for keeper leagues that I’ve liked a lot
Related point: there are ALWAYS at least a dozen guys I value at $5+, who go for $1 in the end game. That’s due to ‘overpaying’ for the top tiers, but also differences in valuation. 12 guys have different views on a player, there will always be a few available you like more than others.
So, you WANT to have 3-5 slots reserved for $1 guys in the end game, because you can count on getting good value from those slots. That money should be repurposed at the top end.
Great Article. It confirms something I experienced in last year’s draft. Most high caliber players were going around $10-12 more than their listed price. It didn’t affect my bidding much though. I had read a great Matthew Berry article a few years ago when I was starting out in fantasy baseball. He broke down a formula for valuing your keepers vs the projected value that was available at the draft. To Mr. Berry’s Credit, he said that he did not create this formula. Berry called it inflation calculation. It allows you to put a better value on the free agents. Here’s a link http://www.espn.com/fantasy/baseball/flb/story?page=mlbdk2k10manifesto
And in case the link doesn’t work for some reason, here’s the meat of the article:
“Let’s say it’s a standard 10-team league with 25-man rosters and a $260 cap. That means there is a total budget of $2,600 (10 x $260) of available money to spend in your league. Now, you add up how much each team has spent on keepers. For simplicity’s sake, let’s say each team has kept 10 players at $10 apiece. So each team spent a total of $100, for a total spent of $1,000 (10 x $100).
OK, here’s where we get even nerdier. Take whatever price list you have decided to use and calculate how much “value” is being protected. For example, my Justin Upton is projected to earn $23 this year. So although I have him at a $5 price, his value is $23.
So you add up all the total value on the teams. Again, for simplicity’s sake, let’s say every team is protecting $160 worth of value. The value protected is $1,600 (10 x 160), but the total spent is only $1,000.
Subtract both numbers from $2,600 (the total budget);
• $2,600 minus $1,600 (value protected) = $1,000 of value left.
• $2,600 minus $1,000 (total spent) = $1,600 of money left.
This means at the auction, $1,600 of money is chasing only $1,000 of value. So you now divide money left by value left; $1,600/$1,000 = 1.6. This is your draft inflation rate. This means that at the beginning of the auction, you can spend $1.60 on every $1 of value and still break even. So let’s say Ramirez comes up for auction in your NL-only league. And your trusty ESPN draft kit has him listed at $34. You quickly multiply $34 by 1.6 to come up with $54. That’s his value in this league at this point in time.
The bidding gets to $40, and your competition, who all have him as a $34-38 player, drop out. That’s six bucks more than he’s worth, people say. But you know that’s actually a bargain for HanRam. You’re saving $14!”
Great article by Berry if you are just starting out (although it’s pretty dated at this point)
That is a nice description of inflation. If you are fortunate, you might be bidding against people who do not understand inflation, and you will be able to get studs below their inflated prices. But there are also leagues where almost everyone at the table is accounting for inflammation. In these leagues you will have to go above the inflated prices to land a stud. This is because the extra $600 (from your example above) is not allocated equally among all the players, but directed disproportionately to the top players.
The key point is to prepare exactly as you did last year since it worked, but also be prepared to adjust if you find the top players are going for more than you’d expect them to. Brad’s whole point is that an approach that worked well in the past won’t necessarily work in the future when others catch up to the math and strategy.
“accounting for inflammation”… not what I meant to type, lol
No disagreement to anything you said. I’m lucky enough to play in a league where I’m fairly certain I’m one of the few people that does the calculation (Too much free time, I guess). In fact, last year one of the owners tried a new strategy out. My league puts a disproportionate amount of value on wins and saves. This particular owner sunk most of his money into pitching and very little into hitting. Surprisingly the pitchers he got were pretty reasonable in value. He also created a shifted market for pitching by taking most of the big names off the board. As Brad pointed out, the market shift then put a premium on players that should not have went for as much as they did.
PS – it’s not my description. Matthew Berry wrote that article. I can’t steal the credit.
spot on. thats how all my auction leagues work
SHHHHHHHHHHHHH
Ok, but where does the money for all the overspending come from? There’s a fixed amount of money in the auction, and if people are overpaying for the studs and for the mid-range guys, that has to leave less money somewhere.
Maybe instead of counting on just the bottom guys going for $1, you need to think of, say, the #150 player as $1 and put all the marginal money into your values for the guys above that?
True, but I think what Johnson is getting at is that auction prices are not linear at the high end (they rise well above Points Above Replacement, even if you set the replacement level high). And if the majority of the league is overpaying, you pretty much have to as well. If only a few owners are overpaying, just relax, sit back and accumulate value. That never seems to happen, though.
My advice is to map ADP/AAV to actual prices paid in your league for several years and develop an empirical pricing curve to estimate what people will actually be paying for each player.
That’s Step One. Step Two is figuring out what to bid when your values are well below the expected price of all the top players. Not sure what is optimal. In football, I tend to set a max bid that is somewhere between true value and expected price for the studs, then shoot for discounts on below-average starters.
Disclaimer: my drafts almost never go very smoothly, so clearly I haven’t figured out a great solution. I’m hoping Joshnon has a follow-up to the this article with more thoughts on the topic.
One piece to add to your thought process that he touches on. Those $1 dollar picks often turn over as undrafted quality emerges. I also would caution that to use “the marginal money” as i understand your statement might have you overspending on marginal talent. I agree with the basic premiss of the article… lead the draft or risk being left behind.
All this means is that the “rational” owners are setting replacement level too low when computing values. In a 12-team MLB league, replacement level is going to be not far below average.
I control for this when I establish my draft values, giving more dollars to the sum of points above last/bench/starters/elite. I guess it all depends on your definition of “rational.”
BINGO! Nicely said. Our keeper league is over 20 years old and for at least the last 10 this is exactly how it unfolds. Better be willing to pay up. I’ve seen owners leave the auction with $10+($260 total) left. Owners who opt to wait it out for the last half in hopes of building a team with a host of players in the teens inevitably end up in bidding wars of their own as 1 or more other owners chose that strategy.