In reply to Alan James - Rockfax:
An attempt to introduce some (possibly false) logic here from someone whose sport climbing experience is admittedly limited - I much prefer trad.
Isn't it possible to rationalise both arguments put forward above: A climb should only have one grade, even though there are two possible grading approaches - grading for "on-sight" or for "red-point".
Take an established climb graded 7a (for example - any grade would do). It will almost certainly be easier to red-point than to on-sight (ie more people could red-point it than could on-sight it). It's still 7a either way, but an individual climber will find it tougher to on-sight than to red-point, and won't find that surprising.
Now take a new route. The first ascensionist will normally be red-pointing it. So he or she grades it by comparing it to other climbs they have red-pointed of about the same difficulty. A second ascensionist comes along who hasn't any beta and on-sights it. He or she grades it by comparing it to other climbs they have on-sighted of about the same difficulty. Other things being equal, they will come up with the same grade, even though they used a "different" method.
Sometimes other things are not equal. The climb might be more than averagely difficult to "read" for the on-sighter who might give it a harder grade as a result. Or it might be more than averagely easy to read as an on-sight and might therefore be given an easier grade by the on-sighter, but this should average out over a large number of climbs. We shouldn't systematically find that the grade arrived at by a red-pointer comparing with other red-points is higher or lower than the grade arrived at by an on-sighter comparing with other on-sights.
So it should be possible to apply the "RockFax" system (which Alan is at some pains to disown) without any problem at the boundary. Grade up to, say, 6c, by comparing on-sight with on-sight (because that's the way these grades are approached by most climbers) and grade from 7a upwards by comparing red-point with red-point. No serious discontinuity should occur because a climb graded, say, 6c+ will, on average, be given 6c+ whether graded by either method.
Martin
Post edited at 10:55