I think the reason it is so rarely successful is simply because the 800 has zero margin for mistakes. Any one strategy tends not to work well in the 800 very often. KD tried to sit and kick today and it didn't work for him. Symmonds usually waits 'til 100m to go, but today he went earlier. Going out hard and hanging on didn't work for KD at the trials. And, as we saw today, even splits can work sometimes, and other times not.
It does seem, though, that running your very fastest for 800m requires you to positive split by 2-3 seconds. However, given that, an even split strategy works GREAT if you are dead-sure that the leaders went out TOO hard. cf. Dave Wottle in 1972. Like I said, there is zero margin for error. So if you're half a second too fast at 400m, you're going to be toast at 700m. If you're even faster, then you're gonna be toast earlier (like Jelimo: 55, 69--ouch).