Well this experiment wasn't a success Although the star Opal ran fine at 22 Mhz, I actually could not obtain any fundamental overtone crystals beyond 22 Mhz in order to see if it could run faster than that. Of course, if you put in an overtone 32 Mhz crystal, for example, it is only going to run at 16 Mhz.

To make matters worse, the power consumption even running at 22 Mhz was horrendous. The Star Opal chewed through a brand new set of Lithium batteries in only 20 hours! Obviously such an overclocked Opal is only of any use when connected using an AC adapter.

But now comes the worst part of all. Both subjectively (i.e playing the unit myself) and also with games against other computers such as Cavalier, I did not notice any subjective improvement in playing strength with the program running almost three times as fast!

It was still making the same elementary blunders that it made at 8 Mhz. Infact in a 7 game match against Cavalier, it lost 6 games and only won 1 of them! And to make that result even the more baffling, a previous match played between the standard speed 8Mhz Star Opal saw the latter score 8.5/20 - a much better result!

So it seems that giving this program a lot more thinking time does not really help afterall It seems that if it makes a blunder through it's selective search, then extra thinking time won't allow it to see the mistake and instead make a better move.

I think there is something fundamentally wrong with these 16K Novag programs. They just seem to drop pieces and pawns through rather obvious and shallow combinations and using simple tactical themes such as pins and overworked pieces, etc. The much earlier 16K programs such as Constellation, VIP, Primo, etc were far, far stronger than these current models. I wish Novag would have just brought out an updated version of the VIP or Primo program. That would be a far stronger opponent than the current Star Opal / Carnelian II / Star Aquamarine.