It's odd for a wireless bike computer to tell you that you're going faster (or longer) than actual speed and distance. It's much more common for it to miss a few rotations of the wheel and report that you are going slower. I agree that the race distance was likely rounded off. You can map the course on a site like gmaps pedometer to see what the distance was.

One other possibility is that you have the wrong wheel size programmed into your bike computer. That would also screw up your readings. Check the instructions to make sure.

My computers are usually spot on, but I use the roll out method to determine the wheel size. It's more accurate than the wheel sizes they give you (close but not perfect) or from a GPS, which can be off as much as 2 to 4%
The other thing to consider is that races are not always accurate. I recently went to a workshop on course measuring. Wheel roll out with a Jones Wheel or a bicycle set up by the roll out method is still the accepted mode of measuring, not GPS since they can be inaccurate. Many races, recently measured by GPS tend to be off, sometimes by quite a bit.
Case in point. There is a local trail race here, and the last few years, at the end of the race folks were complaining to the RD that the course distance was wrong. None of the six complaints were the same! The RD measured the course with a wheel, twice, walking the course with a friend who had GPS. The GPS distance was different both times and off by almost a half mile.

If you're interested in how they work, read up on electromagnetic induction.

@ JonO

I would agree if the bottom of the wheel stoped at the very spot that you placed your magnet. If you set your TOTAL wheel circumference in the computer then every revolution will add up the TOTAL wheel circumference, not the positioning of the magnet.

I agree with TriSooner.

thanks for the info. The Rpm thing is correct. An rpm is an rpm so that should convert speeds and distances accurately. Looking back at the race, they must have figured the speed based on the 18 miles and the time recorded for the race. I did the calculations it figures to be exactly what the race results indicate. Conclusion, the race must have been off for their distance. the race results cheated me a little speed (as well as the rest of the pack-so it evens out) :)
Just to indulge the calculus talk, not that it's important
velocity=arc length/delta time.
Translation, the outside of the wheel has a larger arc (more distance) to travel than the inner wheel and must travel at a higher velocity than the inner wheel. Regardless of position, the number of times the magnet crosses the sensor will be the same and the computer reads that as the entire wheel circumference so distance is the same.

C=R(theta)...

Not quite right, unless (theta) = 2pi. Otherwise, you merely get a length of the segment bound by theta. Presuming your computer does not require you to add multiple magnets and then specify the segment degrees, it will only know the wheel made one complete rotation (2pi). So, naturally, your wheel circumference comes to 2pi*r.

Now... the only variable in that is r, which can be determined (albeit very poorly, due to variations in tire width) by a simple input of 700c, 650c, 26in, etc. I have actually owned very cheap cycling computers that have done the calc this way. They were never accurate.

So... we have a simple calculation that determines distance from 2 variables: count - the number of times the magnet passes the sensor, and r - the radius (size) of the wheel, so total distance = 2pi*r*count. Simple. (Note: r = radius of the wheel, not the distance from axle to magnet, and as such is fixed for a given wheel size.)

However... a computer which allows you to input wheel circumference (c) directly has an even simpler calculation: total distance = count * c.

Now... if were are wanting the computer to give us any form of speed data, and I'm certain we do, we need this to somehow be able to divide a distance by a time interval. Only one way to do that if we're dealing with simple circuits. It needs an internal clock. Speed = dx/dt. No simpler, or more accurate way around it for our purposes.

Conclusion... regardless of the calculation method employed by the computer, distance merely boils down to 1 variable: count. As such, the placement of the magnet on the spoke is irrelevant for such a simple system. If we were measuring forces, the distance from axle to magnet would play a role. But we're not.

Dude. You’re a genius! I am going to go do that right after work. And here I’ve been working my tail off just trying to gain a few more MPH’s.