Back in the early 1990’s after I played a few seasons of Fantasy Baseball, I realized that I did not always agree with the projections being made for certain players.  At the time, I noticed a lot of magazines were using 3-year averages as the basis of their projections or they were assigning arbitrary numbers to players without enough history.  Like many others, I used the projections as a guide to subjectively rank players in an order list.  I usually ranked players within a position first, then overall.  This allowed me to put my imprint onto how I drafted each of my fantasy baseball teams.

As I put more time into making projections, I tried to develop a system that would automate my rankings based upon the projected statistics provided to the formulas that eventually became RotoRank.  After years of tweaking and research, I realized that no one set of projections is really much better than basic trending.  No matter how hard you try to accurately predict statistics, there are too many unknowns.  On average all stat services are likely 60% to 70% accurate in a given year.  See the links at the end of this article for other perspectives on fantasy baseball projection accuracy.

In the end, I settled on building draft guides that would best predict Draft Value versus Statistical Accuracy.  The goal of my projections is to help order players into a draft guide according to their projected fantasy baseball value as compared to the rest of the fantasy baseball league’s player pool. 

Rather than trying to accurately project 5 statistical categories for each of the 1200+ baseball players that will perform in 2010, RotoRank focus on building player clusters.  Approximately 10% of all player projections will be wiped out by unforeseen injuries (RotoRank identified over 150 player projections that were off due to injury in 2009).  Another 10% of player projections will be off to compensate for the injured players they replaced.  So, even if you were able to build a forecasting model that was amazingly accurate, the best you could do is being 80% accurate.  Unfortunately, models cannot predict trades, changes to MLB rosters, changes to equipment, changes in rules, affects of new ballparks (see NYY and NYM in 2009), etc.   In the end, you’ll find that most statistical projections will hover around 65%.

When you rank players by fantasy value within the context of the league scoring system, you are more likely to produce a better draft guide.  When looking at fantasy value, a player projected to have 15 HR and 15 SB, may be equivalent to a player projected to have 25 HR and 5 SB.  Therefore, RotoRank doesn’t need to be as accurate in projecting specific stats; the goal is to group similarly valued players into a Grade Category.

Over the next few weeks, I will post the reasoning behind each phase of my 2010 statistical projections.  As we get closer to the start of the 2010 MLB Season Opener, players will continually move up and down the RotoRank fantasy baseball draft guide grading chart.

This past week, I focused my projections on adjusting the stats for players who only played a small portion of 2009 due to injury or limited major league playing time.  I posted all of these projections on my Twitter account (http://twitter.com/RotoRank).  The impact of this first step is immediately reflected in the Top 10 player rankings for 2010.

Top 20 Players Based Solely on 2009 Actual Statistics

Top 20 Players After Step 1 of RotoRank 2010 Projections

As you can see in the above tables, there doesn’t seem to be much of an impact on the Top 20 players, but some shifting occurred and Alex Rodriguez moved into the Top 6.  The overall Top 200 shows a more dramatic shift in rankings.

Over the next week, I will be reviewing last year’s top performers and deciding if they can reproduce their numbers in 2010. Look for a few of the Starting Pitchers and Hitters over 35 to slide down the rankings.

If you have any feedback, send me a tweet or post a comment. 

If you want to know more about RotoRank’s SHARB/SKEWP Ratings, check out the following link:

http://www.rotorank.com/FantasyBaseball/2009+Fantasy+Baseball+Articles/246.aspx

If you want to read more about the accuracy of fantasy baseball from other top fantasy baseball sites, check out the following links:

http://www.baseballhq.com/books/myths.shtml

http://www.baseballprospectus.com/article.php?articleid=2515

http://www.insiderbaseball.com/VamProjections.htm