Are 3 Year Trend Statistics Relevant?



With so many indicators, which one provides us with the most reliable information at a glance?

Let's start with a little background for our new many of our readers already know, we created the Fantistics concept through a mathematical model, in fact our name is actually a derivative of two endearing words “FANTasy” & “statISTICS” (if I had a dime for every time someone called us FantAstics, subscriptions would be complimentary). The stimulus for the research came from my experience in solving human behavioral scenarios through mathematical means. Using regression analysis we were able to define the variables that fit a pattern of relevancy, thus was born the Fantistics Projections Model. A model that has a solid record through our 7 published seasons and through 56 years of back testing. It has never failed to provide us with less than 70% accuracy in correctly forecasting the direction of player values (up or down) the following season, excluding injuries.

One of the leading indicators or variables that we use to compute our projections is an index that we created called the Fantasy Production Indicator (FPI) or fantasy points per plate appearance. Since it adds walks into the denominator, it should really be called fantasy points per plate appearance. What makes this indicator valuable is that it has a higher correlation to fantasy values than some of the more common indicators (see chart below). However this isn’t rocket science.....rather it considers some of the most available statistics to produce a numeric value which is highly correlated to fantasy production. In his recent best seller, MoneyBall, Michael Lewis describes the importance of the Walk as explained by guru GM Billy Beane. Similarly we've been touting the importance of the "walk" in Fantasy Baseball since 1999, as walks lead to many other relevant fantasy categories such as runs, stolen bases, and ultimately every other offensive player category (as opposing pitchers start making better pitches to those players who have a keen eye).

Fantasy Production Indicator (FPI) gives us the player's quality per plate appearance. Stat Categories within this indicator include and are valued accordingly:


AB -.5

S   +3

2B  +4

3B  +5

HR  +6

BB +1.5

SB  +2

CS  -1

These categories are then divided by the players at bats & walks (PA)

Important Roto categories such as RBI’s & Runs scored are not included in this calculation because of their dependence on outside influences (teammates getting on base, and teammates driving them in).

Comparison: Finding the Best Fantasy Value Indictor













Top 10%






Top 20%






20%- 40%






40%- 60%






60%- 80%












The analysis in the chart above considers all batters who had a minimum of 400 at bats during the season. This analysis chart provides the results of the 5 leading indicators of fantasy value (assumes a typical 5x5 league) . The higher the values in the upper ranges (Top 10%, Top 20%, 20-40%) the more direct correlation that the indictor has to fantasy value.  

Below is a synopsis of the leading indicators used within the baseball community in comparison to our indicator Fantasy Production Indicator (FPI):

#5. Strike-out per plate appearance (K/PA):   Strikeouts per plate appearance are a sign of a hitters’ ability to make contact. Although the chart above shows that there is very little correlation between the ratio and fantasy value (The top 10% of hitters who had a low K/PA only produced on average $14 in Fantasy Value), it does play an important role in accessing the ability of young hitters when combined with other stats or when determining an individual trend. However it has very little if any relevance to fantasy success.

Regression Statistics

Refresher on the Use of Regression Analysis

Multiple R 0.0558598 R Squared tells us how well the estimated equation
R Square (LOW) 0.0031203 fits the data, or the "goodness of the fit".
Adjusted R Square -0.0009156 As a rule of thumb, the higher the R Squared, the
Standard Error 0.0549236 more confidence you have in the correlation between
Observations 249 the two variables (in this case the indicator and the
  Fantasy Value).


#4. Walk per plate appearance (BB/PA):   Walks per plate appearance is an often ignored stat, and probably one of the most underrated statistics in Baseball. MLB Hitters that are frequent "walkers" are often disciplined hitters who wait for strike-zone pitches or pitches that are hittable. However taking Walks is not enough of a factor on it's own to produce a strong correlation with fantasy value.

Regression Statistics
Multiple R 0.2466601
R Square (LOW) 0.0608412
Adjusted R Square 0.0570389
Standard Error 0.0336639
Observations 249

#3. EYE: Walks/Strikeouts (BB/K): The term “eye” or “batting eye” was developed by sabermetricians years before my entry into the sports forecasting industry, and has served as a very important forecasting tool. However according to our analysis it’s usefulness is marginally better than the BB/PA (Walks per plate appearance). In fact our 2005 data shows that it actually fared a little worse, and the R Squared regression below confirms our findings.

Regression Statistics
Multiple R 0.2233144
R Square (LOW) 0.0498693
Adjusted R Square 0.0460226
Standard Error 0.3164427
Observations 249

#2. OPS: On Base Percentage + Slugging Percentage (OPS): One of the most telling indicators of fantasy excellence. However it will distort some  players, such as Barry Bonds whose 2003 OPS produced a 16% gap between himself and the next highest OPS players, Sammy Sosa and Jason Giambi. The problem was that their fantasy values were much closer than the stat indicates.

Regression Statistics
Multiple R 0.6846581
R Square (High) 0.4687568
Adjusted R Square 0.466606
Standard Error 0.0679085
Observations 249

#1. FPI: Fantasy Production Indicator (FP/PA): This is the Fantistics Value Indicator, which has the best correlation to fantasy values according to our analysis. Again, it's not rocket science as it should produce a strong correlation to fantasy values (since many of the same components are included in the calculation.) This indicator is a very important one because it not influenced by events such as injuries & managerial decisions that affect player playing time. Which assists us in expanding on a player’s partial season stats and potential.

Regression Statistics
Multiple R 0.7014556
R Square (Highest) 0.4920399
Adjusted R Square 0.4899834
Standard Error 0.0750323
Observations 249

Using FPI,  we can assign labels to certain types of players:



Player Example

> .90+

Fantasy God

Albert Pujols


Super Star

Hanley Ramirez


Fantasy Star

Vlad Guerrero


Above Average

Jermaine Dye



Brian Giles


Slightly below Average

Dmtri Young


Below Average

Jose Lopez

< 0.35

Fantasy Irrelevant

Adam Everett

Veteran players do not deviate greatly from their historical FPI's, infact they usually fall between 5-10% of their base FPI every season. It's the young up and coming players that frequently trend upward and provide us with the most value on draft day. Below is an early preseason list of young players that are headed for more playing time in 2008 and could produce some meaty numbers. They are sorted by their projected value (FPI):

The Subscriber Section and Projections Software contains the 2009 FPI movers


Please feel free to inquire about any of our products: 

Anthony A. Perri is the founder and the resident "Stats Nerd" here at Fantistics. Anthony is the designer of the Fantistics Projections, Grading, & the VAM drafting strategy models. His fantasy expertise has been published in several national publications, including being featured as a guest expert on Major League Baseball's official website. Anthony has worked as a He can be seen hanging around the MLB spring training facilities (wishing they let him play) during the months of February and March. Having won a "trophy room full" of Fantasy Sports Championships over the last 13 years, he hopes to continue to lead you in the same direction. 



Copyright 2009 Fantistics Baseball, all rights reserved


Copyright 1999- 2007 Fantistic Technologies
All Rights Reserved.