Should KenoRND() Be Used for Serious Applications?

I don't think KenoRND() should be used for any serious application. There will never be enough data from the same machine or same set of balls to determine how much the bias effects things. Also KenoRND() is really slow at generating numbers. Only about 3 numbers a minute are generated, and that simply is not enough numbers to be used for something productive. Finally, this site is on a self-imposed 15 minute delay, so it is easy to predict what the next random number generated will be.

If you are committed to the concept, then you could purchase a fully automated Keno system for under $20,000 and generate a random number every two minutes. A test can be run every 8,000 games or so to see if it is producing random numbers well, and if it is, then I say use it in production. Of course I don't know how long the balls last for, so if the are only good for 8,000 games, then by the time you calibrate them, it is too late and you have to buy new balls (which costs $400).