I didn't say that they aren't worthwile training, but you can't really think that you can hop on a gym exercise bike and that equals real distance ridden on the road. It is exercise, and it is a workout, so it is not worthless. My main point of contention is your desire to see how your "speed" compares to a real TT, and the issue is that a mile on a trainer loses its context and is therefore impossible to compare to real life. A computrainer is probably better than most but certainly a gym's exercise bike
(which is what you said you were using) is completely bogus. The numbers they put out on those are designed to make the exerciser feel good about what they just did.

Now, the second part of my post was that you should be comparing POWER not SPEED. Assuming that the power reading (watts) is accurate, you should be able to get a good idea of how you stack up against various categories. You can do tests of varying durations to determine your Functional Thresholds, and compare those. I've never done it, but a teammate has been using a PowerTap for a little while now and did tests for his FT and found that he was at the high-to-mid cat 3 level for most durations.

Btw, sorry I didn't reply back to you sooner, I was out riding real miles on a real road.

Edit, oh yeah, my best time for a TT is just over 25 min (23.6mph) for a 10 mile TT. The record on the course is just under 22 minutes, set by a cat 3 (with points to upgrade to a 2) while training specifically for TTs. 329W and 27.23mph for a time of 21:48. Look up pro/1/2 times for 40K TTs (close to 20 miles) and you will find your time doesn't stack up very well even if it were real.