Theory: The worse your draft goes, the better your chance to win. Discuss. (1 Viewer)

[dlcb1]

Around six or seven years In my big league, I had a draft where I loved just about every player I chose. I was quite excited going into the season, but the rest of the league was unconvinced, and they were right. It was a brutal first half of the season, and I barely missed the playoffs.

Last year draft was kind of the opposite. Stuck at the end of the draft, I took the falling players, and was uncertain about my squad going into the season. I led the league in points and won the championship. It was a very satisfying year.

I do think you're onto something here. One of the problems with the "great draft" is something similar to the winner's curse; we get the player that we want because we overpay for them. Many of the best picks I've ever had are players that I've settled for. Many players in general think they are incredible talent evaluators, but they are probably not. They're not any better than the rest of us. And they become so tied to their players and their opinions, that they miss out on what's in front of them. Overconfidence is the enemy; it's the struggle that makes us stronger.

I left my draft a little disappointed this year; many of the players that I wanted didn't fall to me. I've worked the WW; I've made a trade. And the more I look at this team, the more I am starting to like it. It's not exactly the team I wanted, but it's a team that I think can perform well. We'll see how it goes.

 

[fightingillini]

Ssolid response, and one I largely ageee with. 

something several folks have eluded to but not quite landed on is the fact that “record” could be entirely meaningless to the quality of team.

I’ve seen bad teams go 4-0 and good teams go 0-4… It all depends on who scores what during that stretch. I’ve had a team that won the 10% of the pot “points total” prize go 4-10. Every week I ran into a buzz-saw. While an extreme example, it does happen. 

so to the point of the topic, that 4-0 team that’s sitting back feeling good about their roster & not working as hard on the wire or trying to make trades might be a paper tiger who’s owner really shouldn’t be in love with their team, but is. 

Again; if there’s a better word than complacency, I’m all ears, but that seems to fit the definition. 

Record is the worst indicator.  I look at points scored and the All Play record as to strength of team.

 

[Long Ball Larry]

I think that what you are really talking about is bayesian prior probability and failing to update your prior probability distributions of players once the season has begun.

It's not really about loving your draft or not loving it or even complacency, but it's about assuming that the the pre-draft evaluation was accurate.  Like say Dalvin Cook starts off with 4 weak games.  Do you just say, well I know that Cook is good and at the beginning of the year it looked like he had a great setup, so let's wait it out?  No, you have to take the new information in those 4 games and add it to what you knew before and come up with a new projection. 

And you have to do that with all players, maintaining your previous opinions and letting them be newly shaped by additional information.

It is probably easier to do this when you have a bad draft because you are on higher alert and I agree that I have fallen victim to the wait and trust approach more than I would like.

 

[Hot Sauce Guy]

Also if your roster is too good top to bottom, you have too many bench points. Great insurance but I want tier upgrades. 

A great point - that’s why I mentioned working the trades too, but you really nailed it. Depth is great to a point. 

I was super happy with my recent draft. But I’ve already dealt a WR5 for a starting DB in my IDP league. That DB will upgrade me from a ~6 point starter to a ~14 point starter, a +8 improvement to my weekly lineup for a WR who’s unlikely to ever be in my starting lineup. 

so if that hypothetical 4-0 team truly is a stacked roster bottom to top, you *still* shouldn’t be complacent - always be looking for ways to improve. 

Great point. 

 

[fightingillini]

I think that what you are really talking about is bayesian prior probability and failing to update your prior probability distributions of players once the season has begun.

It's not really about loving your draft or not loving it or even complacency, but it's about assuming that the the pre-draft evaluation was accurate.  Like say Dalvin Cook starts off with 4 weak games.  Do you just say, well I know that Cook is good and at the beginning of the year it looked like he had a great setup, so let's wait it out?  No, you have to take the new information in those 4 games and add it to what you knew before and come up with a new projection. 

And you have to do that with all players, maintaining your previous opinions and letting them be newly shaped by additional information.

It is probably easier to do this when you have a bad draft because you are on higher alert and I agree that I have fallen victim to the wait and trust approach more than I would like.

This is an excellent post.

And the quicker you can make that reprojection, the better you will be.  Usually 4 games is enough, but starting 0-4 because you thought Dalvin Cook would shake out of his slump doesn't do you much good.   You need to begin looking at all your players after each game.

Context is so important.  Yesterday, Tyreek Hill had a somewhat quiet game.  Why was that?  Game flow and HOU defending against the big play.  KC wanting to show off their new toy in CEH.  HOU offense sucked and KC controlled the clock.  Does this game change my opinion about Hill?  No.  But say the next game is close and KC gives CEH 25 carries again and Mahomes only throws for less than 250 yds again, then....well maybe KC is going to change their offensive philosophy and will be more balanced on offense, and I may have to project Hill's stats down (as well as Mahomes), and project CEH's up.

But what I said here is easier said than done, since FF performance is volatile and so dependent on other players and game flow.  Also players can be playing hurt and not 100% and it affects their performance as well.  You want to do what Long Ball Larry says in his post, but in reality it's challenging to do.

 

[Hot Sauce Guy]

I think that what you are really talking about is bayesian prior probability and failing to update your prior probability distributions of players once the season has begun.

It's not really about loving your draft or not loving it or even complacency, but it's about assuming that the the pre-draft evaluation was accurate.  Like say Dalvin Cook starts off with 4 weak games.  Do you just say, well I know that Cook is good and at the beginning of the year it looked like he had a great setup, so let's wait it out?  No, you have to take the new information in those 4 games and add it to what you knew before and come up with a new projection. 

And you have to do that with all players, maintaining your previous opinions and letting them be newly shaped by additional information.

It is probably easier to do this when you have a bad draft because you are on higher alert and I agree that I have fallen victim to the wait and trust approach more than I would like.

Well said. And I agree. 

What makes this especially hard is a version of sample size that I can’t think of the definition for at the moment - where you’ve spent 3-4-5 months evaluating players so you get it really cemented in your head that Player A is worth X and Player B is worth Y, which is why you picked one over the other on draft day.

Now you’ve spent all that time convincing yourself that Player X is the man - but you get a slow start, and now you’ve got to overcome all that prejudice to accept that Player X isn’t what you thought he was, only based on a much smaller sample size than “3-4-5 months of evaluation”. 

Theres probably a term for it, I just can’t think of it at the moment. 

We all miss on some players - we all hit on other players. But it’s a lot harder to admit to ourselves we were wrong about a player, so we convince ourselves to wait just a little too long for them to prove us right when they may never do so.  After all, we loved them enough to draft them. They’re our guy, sometimes our ride or die. For me it was Montgomery last year. His college tape looked great. Was going to be a perfect 5th round target after going WR-WR-WR. 

Often the latter happens, like Montgomery. We know we need to give players more than a game. Can’t be the guy who panics and gives up on a 5th round pick. But how long of a leash do we give a given player?

Its hard to know. But when you have a player who’s struggling it’s a lot easier to eyeball potential replacements on the wire.

Like a dude who’s g/f isn’t giving him any love - he’s gonna go lookin for love in other places.

Dude who’s got the girl who cooks him dinners and whispers sweet nothings is often gonna he happy with what he’s got. Unless he’s a wolf. 

All FF managers should be the wolf, all the time.