Maybe I'm going about this the wrong way, but using the State monad for random numbers was always leaky for me. Any time I use it, I have to go through and strategically place "seq"s to counter the excessive laziness. IO and ST have worked well, and seem to be faster as well (though that's purely self-anecdotal evidence). I'm stuck in IO anyway for a lot of this (reading data from files to build my lookup tables), so ST didn't seem worth the extra trouble either. Admittedly it would have been good practice, though. What monad do most people use for random numbers? If performance is an issue, do you still use State? Generating an infinite list of randoms seems like a good approach if you just use them in one place, but beyond that you end up passing around the tail of the list at each step, so then you may as well wrap the generator in State, I think. Thanks, Chad

Chad Scherrer wrote:

...

I prefer the purely functional approach as well, but I've been bitten several times by laziness causing space leaks in this context. I'm on a bit of a time crunch for this, so I avoided the risk.

Well, space leaks won't magically disappear if you use IO a .

Regards, apfelmus