Sorry, I have no idea which projection system is the best. There have been
some studies made – not sure how valid they are, but they rarely show one projection system to be much better than the
others – and the winner changes from year to year. In fact, by the time the studies come out, they are obsolete, because
many of the projection systems get tweaked every year.
The Bill James projections come out in FanGraphs first, so I start with his. A few other projection systems come aboard FanGraphs later – around late January to lend
an alternative or confirming opinion. FanGraphs used to carry ZIPS and CHONE, but you have to find those projections elsewhere
Some well respected contributors to scoresheet-talk are unimpressed with Bill James’ projections these days. You can find about all of the rest of free ones useable for
Scoresheet including ZIPS, ESPN, and CBS from one site: Fantasy Rundown (2012). If you chose the RotoChamp projections from there, you can select one of several other useable projections not shown on
the home page by clicking the drop down button at the end of the Projection System Used selection box. One of those is a Composite
- presumably of the others available from that selection box.
Do not waste time with too many projection systems
in order to hone an ideal ranking. Especially now with web drafts, where we don’t have to blindly submit a long draft
list and live at the mercy of Scoresheet’s computer draft, it is more important to make your picks based on how the
player squares with who is already picked for your team and who is left. Group players into echelons, so you know how many
players are remaining in the upcoming echelon still on the table. Most projections are not going to differ so wildly as to
misplace many players in the wrong echelon.
Projections systems leave out much else that is important to consider. They
don’t finely consider a player’s injury history with expected return-to-health dates or sophisticated odds of
re-injury. They are generally ignorant of the specific competition a player may have towards a regular role on his team. With
35 – 45 man rosters and assuming not more than a half dozen prospects, you have the player depth to select those with
greater upside rather than safe picks. Projection systems don’t announce which players come with greater upsides. Each
of these you have to factor in yourself – and are more important than .030 or .040 points of OPS.
If you are a Baseball Prospectus subscriber, PECOTA’s methodology is quite different from any other systems’ in that it is based on comparable
players rather than age and regression massaged individual stats. Its uniqueness makes it worth merging, if you are using
more than one projection system.
This year (2012) Baseball Prospectus (combining the considerable brain power
of Rob McQuown, Ben Murphy, and Scoresheet’s Dave Barton) has developed a Scoresheet Draft Aid that creates an on-line Scoresheet drafting
spreadsheet where you can rank players by position with a single qualitative rating. It includes their actual range factors
and projected playing time and calibrates their VORP to the actual Scoresheet AAA replacement players. Here is the introductory BP blog on it. Here again, it may be dangerous to rely on a computer system to do all of your thinking. By projected playing time,
you should know each player individually: whether their partial playing time is based on time in the minors, injury history,
or limited use during the week. A starter with a 5.00 ERA projected to pitch 160 innings might be rated higher than a 3.50
reliever at 60 innings, because he is compared to Scoresheet’s dreaded Pitcher AAA, but I would rather have the reliever,
thank you – as long as he is expected to be on the big club all year.
On this site’s home page you will find contact information for Ken Warren. You may know Ken from his sometimes arrogant sometimes amusing scoresheet-talk contributions from a few years back. However, from my analysis, his projection methods are completely sound. His method was
included in a 2004 BP study and fared better than all of PECOTA’s competition at a time when they were considered state-of-the-art. And Ken’s
projections are tailored for Scoresheet. In his efforts to get SWARP available for most teams’ drafts, Ken does not
bother with many players who are near or below replacement level. Unless they are good prospects, they probably aren’t
worth drafting, anyway. Ken didn’t have time to make his projections this year, but hopes to resume in 2013.
So, remember: injuries, playing time, and place in and shape of his career
arc should be part of your projections. Platoon advantages matter, too, especially when you are drafting your back-up outfielders
and firstbasemen. How a player arrives at his playing time matters as well. Will he be released in August? Will he be called
up in July? Will he be hurt? Will he be platooned? Will he be sent down after 5 bad starts? These are more difficult to project
than OPS and ERA, but are just as important to consider. No system can devise a ranking or rating that encompasses all of
that, because their significance depends on who else is already on your team. That’s a good thing, because Scoresheet
wouldn’t be so much fun, if you let computer systems do all the thinking for you.