I know that you're here to read and learn about fantasy football, and we'll get to that, but I'd like to begin first with a little thought experiment. I want you to imagine that you are playing a board game with a buddy - we'll call him Steve - when he turns to you and says "I bet you $100 that my next roll will be a six". You know that the die is fair because you've been using it all night. It's been a while since you took basic Algebra, but you're pretty sure that the odds of a six-sided die coming up on any specific number are just one in six. With very little effort, you calculate that this offer is a great deal for you; you have a 5/6 chance of winning $100, vs. a 1/6 chance of losing $100. Without hesitating, you take your numerically-challenged friend up on his offer.
Hopefully we can all agree that this thought experiment doesn't need to go any further. We have all the information we need to know whether this was a good choice or a bad choice. It doesn't matter what Steve actually rolls at this point. If he rolls a four, then great, you're $100 richer. If he rolls a six, then boo, you're $100 poorer. Either way, your calculations were flawless and your decision was correct. If the die does come up six, you wouldn't sit around and say "Oh man, apparently I was wrong, and I would never take that deal ever again." It's very easy for us to recognize that, when it comes to dice, the accuracy of the calculations is not dependent on the outcome of the roll. So why is it that we all have such a hard time recognizing the same thing is true in football?
Every decision that we make in fantasy football is the result of some sort of process. Processes can be highly sophisticated; some people base their projections on multivariate analysis and historical modeling. Processes can be quick and crude; some people decide who to select based on the flip of a coin. Sometimes processes are individual, such as creating our own personal rankings. Sometimes processes are broad, such as drafting from a magazine cheat sheet. Regardless of what they look like, there are always processes in place that lead to whatever decision we ultimately make. Those processes which are more likely to result in desirable outcomes are "good processes". Those processes which are less likely to result in desirable outcomes are "bad processes". Over time, the very nature of what is good process and what is bad process changes. There once was a time where showing up to a fantasy draft with a magazine from June gave you a strong advantage over everyone else in your league. Today, with the rise of the internet and up-to-the-minute information, that same process invites disaster. The baseline for what makes a process good is dependent on the competition and the quality of their own processes. Regardless of what our process is, however, it is imperative that we judge our fantasy skill not by the individual outcomes our decisions produce but by the quality and consistency of the processes we use.
To illustrate this concept in hypothetical terms, let's imagine for a second that all processes could be simplified as either "good process" or "bad process". Let's say that a good process is anything that is twice as likely to result in a positive outcome as a negative one, while a bad process is anything that is twice as likely to result in a negative outcome as a positive outcome. Let's also say that an average owner is someone who uses exactly as many good processes as bad processes. From these definitions, we can calculate a few things. Half of an average owner's decisions will be based on good processes, and two-thirds of those will result in positive outcomes. This category of outcomes can therefore be described as "good process, good outcome", and will represent 33% of the decisions that this owner made. At the same time, a third of those "good processes" will still result in bad outcomes, which means "good process, bad outcome" accounts for 17% of the owner's decisions. Bad processes drive half of the owners decisions, with a third of those bad processes still producing a desirable result, ("bad process, good outcome", which accounts for 17% of decisions), and two thirds producing an undesirable result ("bad process, bad outcome", accounting for 33% of decisions). Let's dive a little deeper into the significance of those categories, with examples of real-world situations that might qualify.
Good Process / Good Outcome
As I mentioned, this category accounts for 33% of all outcomes, and 66% of all positive outcomes. This is the brass ring of fantasy football, the category in which we always hope to land. It means we used a good process, and more importantly, that good process was rewarded with a good outcome. This reinforces the process, and makes us more likely to continue to use it in future seasons. Listening closely to camp reports to see which under-the-radar players are making noise is a great way to pick up sleepers before they enter the mainstream consciousness. If you had been listening carefully in 2005, you might have heard about an undrafted college backup named Willie Parker who was turning heads in Pittsburgh's camp. Those who took a flier on him in the last few rounds of their draft were rewarded with the 15th best fantasy running back that year. That's a textbook example of a good process (listening for buzz to identify early sleepers) that created a good outcome (a strong RB2 in the 20th round or later).
Good Process / Bad Outcome
This category, which accounts for 17% of all decisions, is still a pretty good category to be in. Many people see the bad outcome and figure otherwise, but as Steve's wager demonstrates, sometimes you can do everything right and still lose. In the long run, as long as your processes are good, you'll continue to have plenty more hits than misses. The key here is to recognize that the underlying process was sound, and to refrain from letting the negative outcome dissuade you from repeating that process in future years, which can often be difficult. As an example, we've already established that listening to camp reports to see who is generating buzz is a strong process. If you'd been listening to Tampa Bay's camps in 2010, you would have heard of a little-known running back named Kareem Huggins who was generating some real excitement. If you'd acted on that information and spent a 20th round pick on Huggins, you would have wasted a pick. Huggins only touched the ball five times in his entire career. Still, just because in this particular instance listening to camp buzz led you astray does not mean that you should avoid listening to camp buzz in future seasons.
Bad Process / Good Outcome
This category, accounting for 17% of decisions overall and 33% of positive decisions, is by far the most dangerous category to fantasy owners. Getting a good outcome is always nice, and it will improve your chances of winning a championship this year, but it can become a problem if it rewards bad processes and therefore reinforces bad habits. As an example, if in 2006 you'd decided to wait on RBs for too long until, with all of the starters off the board, you frantically started grabbing backups and hoping for a lucky break, that would have been a bad process. If the backups you'd grabbed, though, included Ladell Betts and Marion Barber III, that would have been a fabulous outcome; Betts finished 10th and Barber finished 14th that year. As strong as the rest of your team would have likely been, those two easily could have carried you to a championship, which would have been great… unless you extrapolated from that experience that waiting far too long on RBs and drafting backups late was a good strategy going forward, in which case that bad process would have left you with a cellar-dweller in the years to come.
Bad Process / Bad Outcome
In many ways, this category is not nearly as bad as it seems. Another phrase for events in this category is "learning experiences". I'm sure we can all think of dumb things we've done which became formative experiences. I can't count how many fantasy football rookies I've seen spend a first round pick on a quarterback, then follow it up by drafting a kicker and a defense way too early to "round out the starting lineup". Typically, this results in them taking their lumps for a year or two and then learning that the cheap availability of quality players at those three positions means you can afford to wait while loading up on the scarcer running backs and wide receivers.
I would encourage all owners to start thinking of their actions in terms of these four categories. Often, it's an incredibly difficult change in mindset. We naturally believe that every good result merely verifies how great our initial processes were, and every bad result invalidates all the reasoning that led to that outcome. It is difficult to accept that as many as a third of our right calls were made for entirely the wrong reason. If we don't confront that uncomfortable reality, though, we cannot grow and improve as owners. Only by looking back on our history honestly and drawing a distinction between when we were good and when we were lucky can we begin the long process of winnowing out our bad ideas and doubling down on our good ideas. And, honestly, that process is the best process of all.
More from Adam Harstad:
Dynasty, in Practice: Valuing in the Face of Uncertainty - September 19
Dynasty, in Theory: A Paean to Uncommon Sense - September 16
Dynasty, in Practice: Early-Season Overperformers - September 12
Dynasty, in Theory: Thinking Like a Bayesian - September 9
Dynasty, in Practice: Keeping a Fantasy Journal - September 5
Dynasty, in Theory: Musings on Confirmation Bias - September 2
A Narrative History of Fantasy Football - August 28
Diversification 101 - August 19
Dynasty, in Theory: The Components of Player Value - August 14
The Opportunity Cost of Top Tight Ends - August 11