It doesn't matter that the roll is reset after every cast, because that's how statistics work.
The nice thing about the game is that you can generate pretty large samples pretty quickly. As a caster, I'll cast upwards of 20-30 critable spells per fight. Given your fixed probability of 10%, you can simply base the prediction off the cumlative probability.
For more on dice statistics (which is what i gather we're assuming here):
http://www.tauonline.org/ViewContent.php?id=121
Code:
A pretty table:
Result Crit% (Probability) Cumulative Probability
1 10% 10%
2 10% 20%
3 10% 30%
4 10% 40%
5 10% 50%
6 10% 60%
7 10% 70%
8 10% 80%
9 10% 90%
10 10% 100%
So if you cast 100 shadowbolts, you have a 100% chance that 10 of them will crit. This of course is not perfectly accurate, but could give you a ballpark image. Something as simple as "unlikely to crit," "probably crit," and "you're damn unlucky if you don't crit."
|
With such an addon, you could make a lot of non-crits and the addon would tell that you'd have 100% chance to crit on your next attack.
|
And thus would be the point of the addon. If you have a LOT of non-crits, the odds are in your favor that you've got at least 1 coming your way (if not a whole long streak of them).
It would be nice if an actual mathematician would hop onboard, and perhaps enlighten us. I'm fairly certain there's more complex models you could use than the one I cited above.
And yes. This addon would probably require a decent bit of work. I myself have never written an addon (or worked with LUA for that matter), and am a mediocre coder at best, thus why I posted the idea rather than writing it myself.