Rummaging around in the 1870s and 1880s, I've begun to notice a few more stats errors for famous horses. I stumbled upon a possible source for certain ones; namely, the indexes for what likely served as data mines for folks who compiled lifetime starts and winnings.
Wallace's furnished charts and indexes similar to DRF chart books. Veteran researchers are used to going to the backs of the books to make a simple count of DRF race index #s.
However, Wallace's used a unique protocol, different from DRF, Krik's, etc. Like Krik, their indexes were page numbers. Oddly, though, Wallace's used superscripted numbers following them as a shorthand to indicate that two (or more) PPs were contained on one page.
Case in point: Baden-Baden, 1877 KD winner. Credited with 10 lifetime starts (four in 1876 as 2yo, six as 3yo). However, I found 5 starts in 1876 by using the index that - superficially - would look like only four starts to an unsuspecting researcher. (Undoubtedly, someone performed the calculation maybe 100+ years ago, the work product was never challenged or re-visited.)
See for yourself.
http://books.google.com/books?id=899IAA ... en&f=false
Use the "GO" button on the left. Scroll to the bottom of the link to page 739 box. "Baden-Baden" has a superscripted "2" for page 542; and, indeed, if one enlarges p. 542 (top box of link), there are two races that Baden-Baden competed in; Sept 25 and Sept 27 at CD.
I just uncovered this yesterday, so I haven't had a chance to do delve into other horses very much. However, the 1877 version index seems to contain another uncredited race for Ten Broeck, although the page needed to confirm it was skipped in the Google Books copying process. Who knows how many more stars (if any) this impacts. There are a lot of superscripts in those indexes.
