Dynasty, in Theory: How Does It Work in Practice?

There's a lot of really strong dynasty analysis out there, especially when compared to five or ten years ago. But most of it is so dang practical-- Player X is undervalued, Player Y's workload is troubling, the market at this position is irrational, and take this specific action to win your league. Dynasty, in Theory is meant as a corrective, offering insights and takeaways into the strategic and structural nature of the game that might not lead to an immediate benefit but which should help us become better players over time.

Theory Describes Reality

It won't come as a surprise that I'm a big fan of theory. I love any conclusion that can be reached from first principles using reason alone.

For instance, managers tend to be higher than consensus on the players on their own team. This is why those players wound up on that team. Given this, here's a favorite observation of mine: if you asked a manager in August to estimate each team's chances of winning the title, those chances will sum to 100%. If you asked all 12 managers in your league to repeat the exercise, you'd wind up with twelve sets of odds that all summed to 100%.

But if you took each manager's estimate for their own odds and summed up all of those, this would total more than 100%. And why shouldn't it? After all, each manager is higher on his or her own team than the rest of the league is. As a result, your leaguemates collectively overestimate their own title odds. (Most likely you do, too.) Given this, future first-round picks must collectively be underpriced in August because everyone thinks they're going to be drafting later than they actually are (on net).

The great thing about observations like this is they're either true or not. As a writer, my job is to persuade you that my observation is correct, and walking through the steps from premise to conclusion is a powerful way to do that.

But the thing about theory is eventually, it has to survive contact with reality. There's an old quip by economist Ely Devons: "If economists wished to study the horse, they wouldn’t go and look at horses. They’d sit in their studies and say to themselves, ‘What would I do if I were a horse?'" We can sit around and ponder life as a horse as long as we want, but if we then go out and observe a horse and discover it behaves far differently than we'd reasoned it must, then we must discard our theories and create new ones.

Before the season, I wrote that for any measure of dynasty value that corresponds even weakly with future production, if you manage to consistently increase your team's value over time, you must necessarily increase your team's production. And if you increase your team's production, you must necessarily win games. As a result, my focus as a dynasty manager was typically on increasing the value of my team rather than winning any one individual game, assuming that eventually wins would be the necessary byproduct of my efforts. I wrote:

"The entire instinct to sacrifice value to increase your odds of winning in the short term stems from, in my opinion, a fundamental misunderstanding of fantasy football. It imagines a much more deterministic universe than actually exists, one where the best teams win and the worst teams lose."

Last week, a reader approached me because he was in the finals in a league that didn't have a trade deadline, and he wanted some advice on a potential deal. I offered my opinion, and then after the games he followed up to let me know how things turned out. With his permission, I wanted to share his story.

Win-Now Trades: A Case Study

This manager, as I mentioned, was in his league's championship game, but Footballguys' tools projected him as a slight underdog. He had the opportunity to trade a young player who wasn't starting for him to acquire an aging running back who was projected to be a significant improvement to his starting lineup-- significant enough for Footballguys' tools to project him as the new favorite.

I told him that I thought the prospect he was giving up was more valuable than the running back he was acquiring, and then I walked through a probability matrix for all win-now type trades. In these types of deals, there are four possible outcomes.

1. You win the league but still would have won it without the trade.
2. You win the league and would NOT have won it without the trade.
3. You lose the league but WOULD have won it without the trade.
4. You lose the league, and you would still have lost it without the trade.

I explained how, no matter what the two teams look like and no matter what player(s) are being acquired, the two most likely outcomes are #1 and #4. (Which outcome is more likely depends on whether you're currently favored or not, but one of those outcomes is always the most likely and the other one is always second.) In practice, there's almost always a 90+% chance that you will ultimately wind up in bucket #1 or #4. In this case, you sacrificed value and changed nothing.

Outcome #2, on the other hand, is the brass ring of "win-now" trades, the situation where making the trade pushes you from a loss to a win. In this case, almost no matter how much value you had to give up, the trade was worth it. If you're looking at a very big upgrade to your starting lineup (with no commensurate downgrade elsewhere from players you have to trade away), then maybe the odds of #2 are as high as 6-7%.

Many managers will stop here and say just how much they'd be willing to pay for a 6% boost to their odds of winning a title. But the analysis doesn't stop here, because we still haven't discussed possibility #3. Sometimes the player you acquire has an especially bad game or the player you're replacing has an especially good one and the supposed win-now trade prevents you from winning a title you otherwise would have claimed.

As long as the projections you're using are decent (and Footballguys' projections are more than decent), the chances of landing in outcome #2 are always higher than the chances of landing in outcome #3. But not that much higher. Fantasy football is a very random mistress. The amount of variance in a single week is huge.

"The entire instinct to sacrifice value to increase your odds of winning in the short term stems from... a fundamental misunderstanding of fantasy football. It imagines a much more deterministic universe than actually exists, one where the best teams win and the worst teams lose."

The odds a win-now trade helps you are not the same as the odds you wind up in Outcome #2. They're the odds you wind up in Outcome #2 minus the odds you wind up in Outcome #3. If there's a 6.66% chance you wind up in Outcome #2 and a 3.33% chance you wind up in Outcome #3, then the net odds of a win-now trade working out in your favor are just 3.33%.

Having discussed all this (and pointing him to a few sources for player values that I find credible), the manager opted not to trade for the running back. That running back was Alvin Kamara, who was hurt in the first half of his game and finished as the #43-ranked running back in PPR. Kamara scored fewer points than the back this manager was hoping to replace in his starting lineup, so trading for Kamara would have cost him points.

Now I suspect you think I'm going to reveal that this manager was in Outcome #3 territory and not making the trade was the thing that won him the championship. But he was not. He was in Outcome #4. He wouldn't have had enough points to win it with Kamara, but he didn't have enough points to win it without Kamara, either. The only benefit of not making the trade is he got to hold on to that extra value for next season.

Or at least that's what I'd be saying if there wasn't a second side to this story that I hadn't mentioned yet.

Because the manager's rival in the championship was also interested in acquiring Kamara to shore up his championship odds and these two managers had gotten into a bit of a bidding war. After discussing with me and bowing out of the running, his rival acquired Kamara and slotted him into the starting lineup. And Kamara underperformed the back he was replacing for that manager, too. And this time, that difference was greater than the final margin of victory. Meaning the other team was in Outcome #3 territory.

If my reader had traded for Kamara he would have lost. If neither manager had traded for Kamara, he would have lost. The only path to victory for him was (rather ironically) for his opponent to make a win-now trade. It probably didn't feel great when that trade went through and this manager saw his road to victory narrowing, but his opponent had unknowingly sown the seeds of his own destruction.

The moral of the story isn't "never make win-now trades". It's not even that the other manager was wrong. The trade did improve his title odds in expectation, but the way it played out was bad luck. But as with everything in this hobby, it comes down to a question of costs vs. benefits. And as always, the best way to answer that question is to accurately estimate both the costs and the benefits.

Most managers don't realize that the two most likely outcomes both mean a win-now trade doesn't change anything. Or they don't consider that sometimes win-now trades lead to losses. (Not as often as they lead to wins, but not much less often, either.) As a result, the estimated benefits are wrong. As a result, the cost they're willing to pay is often too high.

Through theory alone, we can infer that Outcome #3 is possible. Through theory alone, we can even estimate how probable it is. But, 'lest we spend so much time imagining ourselves to be horses that we begin to hanker for hay, it's good to sometimes go outside and touch grass (so to speak) to confirm that reality does, in fact, look like we'd expect it to.

I suspect if you look around, you can find examples from your leagues of presumed upgrades leading to losses. The Top 12 running backs last week in PPR scoring were Kyren Williams, Isiah Pacheco, Breece Hall, James Conner, Jerome Ford, Travis Etienne, Najee Harris, De'Von Achane, Justice Hill, Khalil Herbert, Jaylen Warren, and Joe Mixon; many of these are players who teams were actively looking to upgrade from. Christian McCaffrey and Alvin Kamara both left early with injuries, Raheem Mostert never played, James Cook finished as RB51; many of these are players who teams were actively looking to upgrade to.

(This isn't surprising. Delving back into the realm of theory, I've written about how if you look at any individual week, "weird" results like this are the most normal thing imaginable. Crazy deviations like this should be the expectation, which is why it's so risky to pin your hopes on a short-term upgrade to help you in a one-week sample.)

If you can find examples from your leagues, that's good. It is good to confirm that the horses are behaving as we would if we were horses. If reality conforms to our theory, it give us an excuse to continue theorizing.