Dynasty, in Theory: Winning the Informational Arms Race

How to overcome fixed roster limits and increase the odds that you're holding tomorrow's breakout player today.

Pop quiz: let’s say you play in a dynasty league with 24-man rosters. How many players can you roster? Here’s a hint: the answer isn’t 24.

All that roster limit means is that you can only roster 24 players at a time. But “time” is the key word, here. Managing your roster isn’t just about who you have, but also about when you have them.

Let’s back up a bit. I wrote this offseason about how we have historically measured player value on a per-season basis, but a better unit of measurement would be per-week. The reason is simple: start/sit decisions are made on a week-to-week basis, not a season-by-season basis.

If your league runs for 16 weeks and has nine starting spots, then you don’t have nine starters. Instead, you have 9 * 16 or 144 player-starts to dole out. Some guys, like Jamaal Charles or Aaron Rodgers, are really good. You’re probably going to want to give those guys the maximum possible 15 starts. Other guys, like Stevie Johnson or C.J. Spiller, are more of a matchup-play; they might only get a half-dozen starts during the year.

While the case for considering starts on a per-week basis is clear, the case for considering roster spots on a per-week basis is a little bit murkier. Why is the week the fundamental unit of measurement instead of the season? Why not measure by the day, instead?

The best way to make the case is to consider the players who are available on waivers in your league right now. They’re only available because none of the teams in your league thinks they’re worth rostering. If you add them, then you’re hoping that everyone else in your league was wrong about them.

In order for that player to be a good add, there needs to be some sort of positive value shock that bumps him up from “clearly not worth rostering” territory into “clearly worth rostering” territory. What might such a value shock look like?

During the offseason, we’ll regularly see value shocks in the form of positive media reports. But during the season, you don’t really see major media outlets raving about how a guy looks in practice anymore.

Instead, most positive value shocks are going to come in the form of either an injury to someone ahead on the depth chart, or in the form of impressive play on the football field. And since most injuries happen during football games, each player by and large only has one chance a week to receive such a positive value shock.

So, to recap… adding free agents off the street is an exercise in betting everyone is wrong about them. In order for that bet to pay off, we need to gain some new information that casts their value in a more favorable light. Such information usually arrives once a week. So the last few roster spots on your team— the spots you typically devote to taking fliers on players— should be considered on a week-to-week basis instead of a season-by-season basis.

Providing a Real-World Example

What does this mean? Let’s look at Brock Osweiler. Osweiler was drafted in the 2nd round by a franchise with very high hopes for his future. He looks like the potential long-term starter in an offense with two phenomenal wide receivers, an offense that has produced top-10 quarterback play in the past.

Based on the fundamentals, Brock Osweiler seems like a valuable fantasy prospect. But if you’d drafted him in 2012 and held him to this point, you would have burned a roster spot for three full seasons and still have little new information to show for it. Has anything happened in the intervening time to cause us to revise our opinions of Osweiler upwards?

Let’s imagine, instead, that you ignored Osweiler and devoted that roster spot to quarterbacks who you perhaps didn’t think were quite as good, but who were providing much more new information from week to week. You probably would have churned through some terrible players.

Maybe you grabbed Ryan Lindley when he earned the starting job in 2012. Lindley was a 6th round pick on a terrible franchise, and we had no reason to expect he would be as valuable as Osweiler. But remember, all of the players at the end of your roster are probably useless, and you’re just grabbing them in the hopes that they’ll prove you wrong.

Lindley did not prove you wrong. He threw seven interceptions against no touchdowns and looked like one of the worst, most overmatched rookie quarterbacks in NFL history. But the bright side is that that information came in relatively quickly. You only had to hold him for a few weeks before confirming he wasn’t worth owning. If Osweiler winds up being terrible, he will have cost his owners several years worth of a valuable roster spot.

Churning through lesser talents is a minefield. In addition to Lindley, owners might have rostered Mike Glennon or Case Keenum. Maybe they held Michael Vick over an offseason hoping he’d land somewhere to be the starter. The hit rate on roster churning is abysmal.

But remember, most of the guys you might consider with those spots are going to be terrible. We’re betting on the chance we’re wrong, and the odds are long. At least with these quarterbacks, we found out quickly and moved on. And who knows, maybe the churner landed on Nick Foles or Colin Kaepernick when they earned the starting job and lucked into a starting quarterback for their fantasy squad.

In the end, maybe Brock Osweiler winds up being better than all of these guys. But if you ask me whether I’d rather bet on Osweiler being good, or make a dozen different bets with individually worse odds over the same time span, I’m going to pick the latter. The dominant strategy when playing the lottery is simply to have more tickets than the other guy.

An Urgent Consideration

When filling out the final few spots on your roster, it’s best to prioritize players for whom we will be gaining new information very quickly. I like to call this a player’s “urgency”— the time-pressure we have to add a player before his value spikes.

Jimmy Garoppolo has very low urgency. We do not anticipate gaining any new information on him soon. Someone like Garrett Grayson, on the other hand, has slightly higher urgency; with starter Drew Brees ailing, there’s a slightly higher chance that Grayson sees the field in the next few weeks. With Jay Cutler being ruled out for this week, though, Jimmy Clausen has the most urgency of all. We will unequivocally be gaining new information about him as a player this weekend.

In terms of underlying fundamentals, Garoppolo is probably the most valuable of that trio. He was drafted highly into a great system. Grayson wasn’t drafted quite as high, and New Orleans’ offense looks like a bit of a mess. Clausen, on the other hand, has been in the league for years, and he’s looked terrible at every opportunity. If I had to hold one of those three for the next ten years, I would handily pick Garoppolo.

But we don’t have to hold for the next ten years. We can make roster decisions on a week-by-week basis. While Clausen’s chances of proving us wrong are the lowest, we only have to devote a single roster spot for a single week to re-evaluate that opinion. The cost of entry is by far the lowest.

If Garoppolo sits for two more years, he will cost us an addition 30 “spot-weeks” to find out what we have. That’s an expensive lottery ticket. I’d much rather buy a dozen tickets that cost 5 “spot-weeks” each, even if the odds are lower.

The team that rosters Clausen and the team that rosters Garoppolo might have the same number of roster spots, but the team with Clausen will be able to effectively roster many more players. Remember, the name of the game is being the guy holding the player when the music stops.

By paying attention to the concept of “urgency”, owners can overcome the fixed number of roster spots and still compile an overwhelming “informational advantage” over their peers.


More articles from Adam Harstad

See all

More articles on: Dynasty

See all

More articles on: Strategy

See all