JohanFranzen 

 

Why, all things being equal, the odds say you take the older established guy over the kid with upside.



You are in a one year pool.  It’s your pick and you’ve narrowed it down to 2 players. Player A is universally acclaimed as a potential sleeper pick due for a breakthrough year.  Player B is a veteran whose numbers have hit a plateau and maybe even declined in the last few years.  Who do you pick?

 

Most will take Player A for two reasons.

 

First, we are culturally conditioned to think that the new is going to overtake the old.  You have subconscious reasons to draft younger.  Your bias for youthfulness and newness is reinforced every day.

 

Second, you will have pressure from other people in your pool.  The cool picks will be:

-       The newly traded player who finds himself farther up the depth chart (Frolik in Winnipeg).  

-       The average player who lucks out and gets a surprise addition to his line (Anisimov with Gaborik on the wing in Columbus)  

-       The easily pushed off the puck lightweight who opts to train with Gary Roberts over summer and is in “the best shape of his life”  (seemingly half the league). 

 

These are the picks that receive the grunt of affirmation from other ‘in the know’ poolies or even better, the rewarding, “Who?  I didn’t even have him on my list!”

 

But you might be wiser to fight the impulse for the new and the potential and go with the fading veteran because of something called the “Lindy effect”.

 

Simply put, the Lindy effect states that the longer a technology has been around, the longer it is likely to stay around. So the Lindy effect argues that in 200 years we will likely still be using wheels (an ancient technology) but there is a lower probability we will be using cell phones (a new technology). There are exceptions to this effect – that is why it is an effect and not a rule.

 

Let’s apply a modified version of the Lindy effect to fantasy hockey.  Here it is:

 

The failure rate of the up and coming player is going to be higher than the failure rate of the veteran.

 

Let’s look at a specific example to illustrate the point.

 

Dobber's Fantasy Guide has Anaheim LW Silfverberg projected for 47 points and Detroit LW Johan Franzen getting 53. If you are torn between the two of them the modified Lindy effect pushes you towards Franzen.  This is because the failure rate of the new (Silfverberg) is going to be higher than the failure rate of the old (Franzen).  Or to put it another way, Franzen is more likely to keep doing what he has been doing than Silfverberg is to do it for the first time.

 

Now, there are reasons to go with Silfverberg,  New team, new linemates, and untapped potential mean he could conceivably blow expectations out of the water and hit 70 points. It’s highly doubtful, but way more likely than the “The Mule” hitting 70.  No way Franzen surges like that.  Is he going to get more ice time or better linemates this season than before?  No. So if you want to take a home run swing, go with Silfverberg. He could get 70.  But my point is that while he could get 70, he might not get 47.  You know Franzen will get his 53. 

 

It is not about potential but about probability.  It’s not about upside but about minimizing the downside. My advice to you? When most everything else is equal go with the fading star over the sleeper. 

 

 


Write comment
Comments (14)add comment

number54 said:

number54
RE: PuckPastor You're right to say I'm nitpicking in my assessment, providing specific contexts in which Weselake's article might miss a beat. However, the examples are there to illustrate the fact that the article did not account for the context in which picks are made, because it ignored the relative value of successes and failures. This applies not only to deep picks, or leagues where there are bench slots -- it applies everywhere. The thing is, it wouldn't make my case too compelling if I only gave examples where players' expected values matched the Lindy effect's predictions. My aim was to show that the Lindy effect tells you that the conservative pick is best, when that's NOT always the case. Sure, sometimes you're best suited to taking a conservative pick -- heck, maybe even most of the time, but NOT every time.
August 23, 2013
Votes: +0

OceanMon said:

OceanMon
now i see this explains why autodraft teams usually do so well, in my experience.
August 23, 2013
Votes: +0

PensInThree said:

PensInThree
... In my opinion this is a misuse of the Lindy Effect. As I understand it, the Lindy Effect describes the expected lifespan of non-perishable things. Such 'things' need not be concrete, as the Lindy Effect applies to the expected lifespan of inventions (IE the wheel, cell phones - as the article describes), but also the 'fame' of musical groups (where fame might be understood to be an approximation of lifespan), the dominance of a particular ideology, etc.

The wheel has been around for about 5000 years. The Lindy Effect states that it's likely to be around for another 5000. This is because in the absence of evidence suggesting a particular lifespan for a relevant 'thing', the longer that thing has been around the more essential/durable/able to weather competition/etc it is.

To apply the Lindy Effect to a perishable thing (IE a tomato, a duck, the career of a hockey player) is fundamentally problematic, as we have evidence in respect of the expected lifespan of that thing. To put the point generally, perishable things are biological things that, based on their composition, break down over time.

Applying the Lindy Effect to a 40 year old Teemu Selanne would suggest that the Finnish Flash would continue producing .9 ppg seasons for another 18 or 19 years. This is an absurd result.

While I certainly don't disagree with the thrust of this article (proven vs potential), I'm not sure that the Lindy Effect can effectively be modified so that it applies to this particular context.

Cheers
August 21, 2013
Votes: +0

PuckPastor said:

PuckPastor
RE:Number54 Really good points from 54; but I think what you are actually doing is providing contexts (waiver wire / deep benches) where the Lindy effect no longer applies rather than contradicting it as a beneficial principal. I'm guessing many of the readers are in formats where they can't as easily minimize the downsides of volatility and are actually vulnerable if they get stuck with a sleeper who ends up getting 25 points.

At any rate, the modified Lindy effect serves to remind us that:
- our tendency is for 'new' and 'breakthrough'
- what is new has more volatility.
- volatility means more upside (hooray) but it also means more downside (we are often blind to this)
- in some situations you are wiser to minimize the downside than hope for the upside

Thanks for the welcome, comments, and insights it's all given me lots of fodder for further articles.

August 21, 2013 | url
Votes: +0

Spleentastic said:

Spleentastic
... While the article is mostly correct, I think number54's approach is a better measure of value than simply choosing proven vs. unproven. It's called expected value, and it's a simple economic concept that most investors will use when trying to balance the risk and return of a particular investment. It's actually a really nice way to rank your players in an objective fashion. If you maximize the expected value of your team, then more often than not you will come out ahead.

However, the problem is whether or not we can be objective in those kinds of measurements. The most valid point Weselake makes is that we have a tendency to overrate young players for a variety of reasons, which means that if you want to use expected value as a tool, then you need to be brutally honest and dispassionate about every player, something the vast majority of us will struggle with. Either we overrate their upside or, more likely, we overrate the likelihood that they will ever reach the upside we project for them.
August 21, 2013
Votes: +0

ktox said:

ktox
Agree with Number54 Number54 gets it right here. When drafting players in the 50 point range upside is critical. The reality is that waivers is full of guys chugging along at 50 points. Franzen in this example is almost guaranteed to be a marginal player on your squad. And similarly it is almost guaranteed that his impact on your team is minimal, i.e. picks like Franzen do not win you your league.

My advice would be to draft Silfverberg, if he fails which is very likely, then drop him and pick a Franzen off waivers.

You won't find the upside (70 points) of Silvferberg on waivers.
August 21, 2013
Votes: +0

number54 said:

number54
You're missing an important point here This article has two glaring holes in it: it ignores the relative value of success for each player; it ignores the relative cost of failure for each player. First, you're treating successes for both players as equally valuable outcomes (or at least, you're not going out of your way to stipulate otherwise). In fact, even in your own example, I'd be leaning towards Silfverberg: he needs about 25% sleeper chance to be square with Franzen: F-100%(53) = 53; S-75%(47) + 25%(70) = 52.75. Even a Silfverberg failure is close enough to Franzen's point totals (IMO) that I'd take a gamble on this pick. If you would advise against gambling on low-percentage wins, then you would essentially be advising against participation in fantasy sports to begin with.

This brings me to the second point: Let's say I have 5 bench spots on my team, and I can pick either 5 Franzens with my last 5 picks or 5 Silfverbergs. Assuming that I only need 1 or 2 of those 5 guys to count towards my final point tally, having high variability is a GOOD thing. It means you get to keep the successes at full value without losing value for the failures. Meanwhile, if I were to take several Franzens and all of them succeeded, half of them wouldn't count on my roster, and I'd LOSE value even with a string of successes.
August 20, 2013
Votes: +1

angus said:

angus
... Great debut!

One of my golden rules of fantasy hockey - 90% of the time, go for proven over unproven. You may lose out once in a while on a stud rookie or sophomore, but over the long haul you will come out batting over .500.
August 20, 2013
Votes: +1

goose191919 said:

goose191919
... Yeah but chicks dig the longball....
August 20, 2013
Votes: +0

Doctor Robz said:

Doctor Robz
... Every fantasy GM DREAM about getting a Giroux or a Stamkpos right before his breakthrough season. That's why a lotta GM will reach for potential over fading veteran. This article is great and it gives great advices, but all depend on if you have a good core before this pick. If you get a grab on stamkos , Giroux and karlsson on the early rounds , I would go for the veteran guy. If I draft last and I get guys like getzlaf , vanek and chara , I might be temping to go for the homerun guy as wining the pool with a somewhat weak core is less probable.
August 20, 2013
Votes: +1

jer_33 said:

jer_33
Young vs. Old Consistancy is incredibly important, especially in one-year H2H leagues. For that reason alone, picking a vet is better than picking a player based on a "break-out". Purely from a draft angle, you won't often find much value in a break-though player - he is almost always drafted too early. These players are worth owning (and can win you a league), but too many "hail-mary" picks will usually fail to place.
August 20, 2013
Votes: +0

back check said:

back check
you're porbably right Good piece and advice I should follow more often, as you are probably right. However the last line about playing only the odds takes some of the fun out of fantasy hockey for me. Winning is great (and frankly is the point in a competative structure), but if its just about playing the odds without any risks its not realy like real life at all.
August 20, 2013
Votes: +0

robmyatt said:

robmyatt
agreed well said.
Great writing style.
Never really heard of the lindy affect before but many on the board often preach the addage "the biggest mistake you can make is assuming the young gun will reach his potential this year"...or something to that effect.
August 20, 2013
Votes: +0

Kofax said:

Kofax
... Unless of course it's a late mean nothing pick that won't crack your lineup unless they put up 60-70 points. Then the choice when all else is equal is Silfverberg.

Otherwise, I completely agree with all you say and it took me 3 years of fantasy sports to learn this.
August 20, 2013
Votes: +0
You must be logged in to post a comment. Please register if you do not have an account yet.

busy