[nearmiss]

All of us who have been around for awhile and flown online realize that people do have patterns of behavior that occurs in air combat regularly.

There are just things you learn from your experience that can be programmed into AI performance. Probability hasn't been discussed much, but human behavior can be predicted in many air combat situations with high probability.

An example might be you are low and see a fast moving enemy dropping on you from above onto your 90. You know when he starts to pull up out of this dive your best course of action may be to nose down pulling towards him, but definitely not to pull up going away. These kinds of behavior can be programmed, but a probability of what the low pilot will actually do should be considered. Panic or pressure of the engagement may actually convert to the absolute worst thing to do or the smart thing to do.

I realize most AI performance is based on pilot experience, but still the best most experienced pilots made mistakes. Probability calculations should be part of the AI performance.

Programming AI is difficult. It cannot be ignored... just like in baseball, the probability statistics do mean something. Psychologists have studied proability as it relates to human behavior for years, and it is an ongoing study. No, the extremes and radicalized parts of behavior may not be determined by probability... but for most of us grunts in the world the statistics are pretty well on target.

I'm not trying to throw something new into the mix. The AI performance for a COD will take time and testing to arrive at a passable performance, that most of us can accept.

As it is now... the arcade is open