I'd be the first to confess that I'm no stats guru, but those figures do seem a little strange to me, I wonder if there hasn't been some obscuration of the detail in order to make a clear presentation? Feel free to correct my interpretation but:
If you accept the figures that 95% of Grads and Ex Rankers get through finally, that would suggest that only 1 in 20 Graduate applicants to OASC completely fails to get in. That seems very much higher than the mythology of OASC usually presents.
Could it be that the second set of figures represents the pass rate of personnel returning for a second plus time, but ommitting the fact that very many first time failures might not return to be counted again? In figures, what I am suggesting is that for every 100 Grad applicants 75 get through first time. But then of the remaining 25, only 10 decide to try again, at which point 9 (and a half!) get though. So in fact, only 84% of all who ever tried, made it through.
Also, it says pass rate, is that the same as 'being offered a place on IOT' rate? There is a very big difference to the candidate between passing OASC and being offered a place - I know, I spent 3 months waiting between the two stages!
Any other thoughts/corrections to my analysis?