I bet you knew I'd bite Duritz. :-)

I'm working on NSW races at the moment using 1994 to 2002 to work out the system. On those races (33361) I get 28.63 percent winners, 18.70 percent seconds and 13.80 percent thirds. That's on races covered by the NSW Tab and it only produces 1.87% profit if you bet on every race which has greater than half the runners having had a start before.
Using that system to predict results for 2003 to present.
9409 races, for 27.86 percent winners, 18.81% seconds and 14.17% thirds for a 0.8% loss on turnover.

That's not exactly answering your question but I don't actually keep a record of what my second, third, etc choice does. I used to but they were depressingly low compared to the first choice and were just taking up cpu time to process.

BTW Good to see you back in circulation. You've been quiet of late.

KV