How To Create Fantasy Football Projections Video Series
Ever wonder how winners prepare for the fantasy football draft? Do you want to learn how to create your own projection and cheat sheets for the draft? Bookmark this page because it's a growing library of videos and articles centered around teaching you how to create fantasy football projections and in turn use those projections to create a cheat sheet to dominate your draft. You'll learn how to prepare for the fantasy football draft and instinctively make logical calls when it's your turn to pick.
What You'll Learn in This Timeless Guide:
- How to create projections
- How to turn projections into ratings based on consistency
- How to use those ratings to create positional cheat sheets and tiers
- How to find sleepers
Okay, are you ready to learn how to draft a fantasy football team like the guru champions you read about in the Magazines?
- Buy our draft guide and save time.
Here we will introduce you to a few broad fantasy football draft concepts aimed at teaching you how to draft a fantasy football team like the professionals. I’ll dedicate extended time to each idea in later videos, so don’t get overwhelmed if you feel like you miss something.
The first concept we’ll cover in this video series is how to create projections. Projections—in terms of your exact stat predictions for each player—are really the foundation of your draft strategy. Essential to the accuracy of your projections is the idea of consistency—which stats carry over from year to year and which don’t. Consistency—or a lack of it—is one reason most experienced owners wait until the last rounds to select kickers and defenses. It doesn’t matter how many points players at a particular position will score if you can’t accurately predict those performances during the draft.
Related to the concept of consistency is regression toward the mean. One of the biggest mistakes fantasy owners make is believing that a player’s stats over the course of a season or even a few seasons are a reflection of what that player really is in some objective sense. The truth is that fantasy football is a game of percentages and randomness plays more of a role in outcomes than most of us want to believe. By understanding which stats are likely to regress toward the mean and how they’ll do so, you can create more accurate projections, providing exceptional leverage during your fantasy football draft.
Once our projections are complete, we’ll need to turn them into rankings. A lot of fantasy owners will simply use their projections themselves as rankings, drafting players who they think will score the most points. This isn’t how to draft a winning fantasy football team.
Just like a stock broker, the best fantasy owners have a keen understanding of how risk and potential reward affect value. Depending on a variety of factors—such as where you are in the draft and your current roster—you may very well want to bypass your best available player—or the player you expect to score the most points—for one with greater upside. Other times, it’s best to play it safe and stick with the sure thing.
Another factor that makes drafting more complicated than simply selecting your highest-projected players is scarcity. In almost all situations, you should draft the players who are outliers at their particular positions. By utilizing scarcity—or how players’ projections compare to others within their position—you can create tiered rankings that will help you make the right choices on draft day.
Finally, we’ll use our rankings to create separate big boards for each position. I advocate this approach because the value of each player in relation to other players at different positions is relative; it changes as the draft unfolds, and that limits the practicality of one overarching big board.
No matter how you construct your final fantasy football draft board, you need to be flexible with it on draft day. I’ll show you exactly how to use your draft board, when it is smart to choose your top-rated player and when you should actually bypass him.
Step 1: How to create projections
When we’re projecting stats, it’s important to realize that a player’s stats in the past might not be a good representation of his future production. In addition to things like aging and coaching changes, sometimes players simply get really lucky or unlucky in a given season. To make accurate predictions, you need to determine how much of a player’s production was due to skill, and repeatable, and how much was simply the result of luck, and thus not likely to repeat itself.
Consistency Matters In the Draft
In short, you’re looking for how consistent each player’s stats will remain from year to year. This consistency is useful. If a particular stat theoretically had 100 percent consistency, you could count on it to be the same every year. In contrast, a stat with zero consistency also has zero predictability. Kickers, for example, have almost no consistency from year to year. That’s why it’s useless to draft one before the last round; even if you knew one kicker would score 1,000 points, it wouldn’t matter if you couldn’t predict which one.
Each stat has a different level of randomness associated with it, but in general, things that tend to occur less often are more random and difficult to predict. It’s far easier to project a quarterback’s passing yards than his interceptions, for example. It’s not that interceptions don’t matter in fantasy football, but because they’re a low-frequency event, they’re really difficult to predict. Because of that, we need to weight them less heavily in our projections.
Randomness and Regression
When something is random, it eventually regresses toward the mean. For example, if you flip a quarter 10 times and it comes up heads all 10 times, that doesn’t make it more or less likely to come up heads in the next 10 flips. Since flipping a coin produces a theoretically random result, the coin is likely to regress toward the average number of heads you’d typically get in any series of 10 flips, which is five.
We know the long-term average of flipping a quarter is 50 percent heads and 50 percent tails, but what about in football? How can we tell Adrian Peterson’s long-term “average” YPC, for example? One way to do it is to infer that average from past events. The average of players’ rate stats, such as YPC, YPR, and YPA, over the past three seasons is typically superior to that from the past year alone. A three-season sample is less susceptible to lucky play than just one year. If a quarterback throws 5, 6, and then 20 interceptions over a three-year period, he’s probably more likely to come closer to the lower numbers than the 20 picks, even if he threw those 20 interceptions in the most recent season.
In this way, note that regression isn’t always bad. If a running back has a down year and posts only 3.9 YPC after two seasons above 5.0 YPC, he’s likely to “regress” upward toward his three-year mean, leading to a positive result.
How To Draft Quarterbacks
In the next video, I’m going to show you exactly how to project stats for quarterbacks, the first step in learning how to draft a winning fantasy football team. I’ll use passing yards as an example, detailing how consistency and regression should factor into your predictions. Again, I’m going to walk you through every step so you don’t get overwhelmed.
Step 2: How to turn projections into ratings based on consistency
We’ve created stat projections, and now we’re going to use those projections to create a power rating for each player. The power rating incorporates year-to-year consistency into our rankings, allowing for better predictions.
How to turn stat projections into point projections
The first thing we’ll do is turn our actual stat projections into point projections. To do this, you’ll need to know your league’s scoring system, but the process is pretty straightforward. If you project a running back to rush for 1,100 yards, for example, his points projection is 1,100 divided by 10, assuming your league awards one point for every 10 rushing yards, for a total of 110 projected points. Kurt wrote more about how to do this in excel.
How to determine Predictability / Consistency
Now, we’re going to alter that point projection using year-to-year consistency. Each stat has a certain level of consistency from one season to the next. If you multiply your point projections by that level of consistency, you can obtain a power rating that will be more useful than your actual projections for each player. Don’t worry about the consistency correlations, because I’ll provide each one in later videos. Once you add together the revised projected points for each player, you’ll have an overall power rating that can be compared to others at his position.
Let’s use Packers quarterback Aaron Rodgers as an example. Suppose we’ve gone through the projection methodology and decided to project Rodgers at 40 passing touchdowns in 2013. In a league that awards four points per passing touchdown, that’s 160 points. Most people would stop right there, but that number isn’t much use without knowing how accurately we can predict passing touchdowns. If they were totally random, for example, our projections would be completely useless.
The year-to-year correlation for passing touchdowns is around 0.6, or 60 percent, meaning about 60 percent of a quarterback’s passing touchdowns carry over from season to season and the rest regress toward a league mean. By multiplying our original projected points for Rodgers’ passing touchdowns—160—by the consistency correlation—0.6—we’ve calculated Rodgers’ passing touchdown rating. Once we perform the same task on each of Rodgers’ stat projections and add up the results, we’ll have Rodgers’ overall power rating.
It’s important to note that the rating is no longer a projection of how many points Rodgers will score, but it’s actually more useful than projections.
The reason is that projections are worthless without some understanding of predictability. Traditional projections and rankings are all counted the same, failing to account for risk. My power ratings have consistency and risk already factored into the mix, meaning they’re more accurate than traditional projections.
On top of that, they also lead to a better representation of scarcity within each position. Scarcity is something I’ll discuss later, but it refers to how much different players deviate from one another within the same position. Understanding scarcity is essential on draft day, but it’s not accurately represented in traditional projections and rankings.
In the next video, I’ll show you the rest of the quarterback consistency correlations and we’ll create an entire quarterback power rating.
Step 3: How to use those ratings to create positional cheat sheets and tiers
How to Find Sleepers.. (coming soon)
We are going to continue to update this "how to draft" in fantasy football series forever. Yes I said forever! Check back soon for updates or you can subscribe to our mailing list and follow us on G+, Twitter or Facebook for updates.