Well I'd have to work that all through for a while but a couple of things jump out:
The degree of success is not necessarily the barometer in horse race predicting.
Erm, huh?
All selections are made utilizing only a horse's number, morning-line odds and the numbers derived from the random deal of the cards.
Well, firstly, I love the way the actual horses odds is kind of tossed in as an afterthought.
It's almost like a dowser saying "I can detect underground pipes using only two beechwood twigs, a specially blessed 'holding' cloth, a water authority map, and my SACRED CAP of DOWSING."
Secondly, of course, why could he not also be using knowledge of the turf condition, horses' previous form, weather conditions, jockey information etc. etc. Like, say, a professional gambler? How could the JREF possibly ensure his isolation from all this gambling information?
The cards are agreat littl prop, but, ultimately, I would imagine a knowledgeable and organised, disciplined professional gambler could achieve these same results.
Doesn't sound too paranormal to me.
See if he can do it with coin tosses - that would be more convincing, and he only has 2 choices each time. AND he already knows the odds.
Or can we add "It only works with horses" to our burgeoning list of claim excuses?