Prior Year Factor

It's one of many time proven variables used
in our Classic Cheat Sheet Ranking Process

Ask your local weatherman for tomorrow’s forecast and he may very likely leverage weather stats from today. Ask an economist for tomorrow’s closing stock price and you’ll often get an answer near today’s price.

How significant is player performance from last year when it comes to forecasting the upcoming season?

Let’s face it, NFL football is multifaceted and Fantasy Football is very situational. As such, should last year’s data be valued or ignored? If not ignored, how should last year’s data be used?

Take a look at last year’s Top 10 Fantasy Quarterbacks. Now, write this down: 60% of these quarterbacks will not make this year’s Top 10 list! Is this an extremely bold prediction? Actually, it’s not gutsy at all.

Examining numerous NFL seasons reveals the year-to-year turnover rate among Top 10 quarterbacks is about 60%. Reviewing year-to-year turnover rates for Running Backs, Wide Receivers and Tight Ends paints a similar picture.

Amazing but true, most of the Top 10 QBs from last year will likely not return this year. What makes this really sobering is the fact that having the 8th, 9th or 10thranked Fantasy QB barely helps you win in most leagues.

Given these facts, should one completely disregard last year’s statistics? The short answer is - - - Definitely Not!

After examining over 31,000 year-to-year NFL player performances and combining data from 1,500 NFL games with scientifically proven problem-solving techniques, several things become very clear.

Analysis Fact: As it turns outs, on average, properly leveraging last year’s data controls for almost 50% of the uncertainty in forecasting a given player’s output in the upcoming season. Statisticians would say 50% of the uncertainty is explained by examining last year’s data.

There are four points to be made here.

#1 Overall, last year’s performance makes for a decent leading indicator. As such, it simply can’t be ignored.

#2 Using last year’s performance data only provides part of the picture. We often see that last year’s data lacks proper circumstance. Certainly, a team change, a new offensive scheme, a new coach, a new contract, an off-season injury, free agent movement, off-season injury rehabilitation and other factors all have a direct impact on player output from year to year.

#3 Very important - note the word properly was clearly underlined above! Often, football averages and totals lack context. This really becomes an issue when you need to forecast player outcomes.

#4 Exercise great caution if you tally last year’s statistics for the purpose of using those totals and/or averages as the basis for this year’s projections! Before you use prior year stats, careful adjustments should be made.

The 4for4® Fantasy Draft forecast process uses player performance data from the prior year. That’s correct, warts and all, last year’s data is one of several factors that go into the forecasting process (see point #1 above). The key is how we use last year’s data. As mentioned, it requires much more than simply totaling statistics, by player, from all games last year.

A Classic Example. It's from the 2000 season, but it nicely illustrates several key points.

Dante Culpepper - Doug Flutie Case Study: During the 2000 season, thirty-eight year old Doug Flutie passed for 1,700 yards. That ranked him 29th among NFL quarterbacks. In isolation that ranking certainly could imply a sub-par performance for the aging signal caller in 2001. However, in 2000, Flutie only started 5 games for the Bills. In that light, suddenly 1,7000 yards isn’t that bad. In fact, when you rank QBs based on passing yards gained in games started, Flutie was among the NFL’s Top 10 in 2000. On a per game started basis, Flutie actually generated more yards than Dante Culpepper, arguably the best Fantasy quarterback from the 2000 season. So, was Flutie only the 29th best? When put into the starting lineup, was Flutie real Top-10 fantasy material? If given the same chance, would Flutie have outperformed Culpepper over the entire 2000 season?

This example illustrates some of the real problems with season totals and per-game averages. In the case of Doug Flutie, most would agree, his success in the 2000 season (albeit over a rather limited time) was very helpful for fantasy managers. However, it would be misguided to give the nod to Flutie (253yds/pgs) over Dante Culpepper (246yds/pgs) heading into 2001. In this situation, a player starting 16 games (like Culpepper) should receive more credit than a player sporadically starting 5 games (like Flutie). Right here, even if nothing changes from year-to-year, we can see PER/GAME and SEASON TOTALS both have major flaws when it comes to determining a player's fantasy worth. leverages last year’s data and carefully provides context to the statistics. Via systematic review and vigilant adjustments, our analysis of last year’s data considers:

A) Player Position and Relative Value - As such, you will notice scores displayed in the ‘Last Year’ column of our Classic-Style cheat sheets. The scores range from 0 to 100 for each player. Generally, only one player from each position earns a score of 100 (the best rating possible) based on last year’s performance. Again, this rating is not just adding up FF points from last year, we're way past that.

B) Player Production when given Opportunity - The Flutie-Culpepper example illustrates how prior year stats often ignore opportunity level and/or can be easily influenced by just a handful of games. As such, in preparing for the upcoming year, we review each week of the prior NFL season and only consider stats when a player was given a reasonable opportunity to produce.

C) Rewarding Sustained Effort and Upside Potential - We have illustrated that averages from last year can be highly influenced by just one or two games. As such, in preparing for the upcoming year, we carefully aggregate player performance from last year and objectively reward players that have demonstrated sustained excellence over consecutive weeks. While this may sound like the 4for4consistency factor, it is not. A brief example: Let’s assume Player A had five 100-yards games during the prior season; also assume Player B did the same. However, Player A was able to reel off 100 yards each week, for five consecutive weeks. Conversely, Player B randomly popped for 100 yards, five times over the entire season. Everything else being equal, who would you rather draft? If you said Player A, then you understand the value of sustained effort and our upside component.

After the process of correctly evaluating last season’s data is complete, individual player scores are computed, saved and then feed into the overall 4for4 Fantasy Draft ranking process.

Properly leveraging last season’s performance provides 4for4 with a concise, objective, measurable, unbiased, effective and consistently replicated variable that goes into your Classic-Style Cheat Sheets.

By itself, ‘Last Year’ data scores have modest value. However, when combined with the other 4for4 components, ‘Last Year’ data scores make up an integral component in improving overall forecasting accuracy.