I recall Dave giving exact measurements as well as an ambitious schedule of swapping chains to give maximum use of your chains and cassette
I've been doing the chain rotation usage that DaveS put forward in good detail for the last couple of years.
It worked on using 3nr chains across 1nr cassette with wear being put sequentially across the 3nr chains for around 1500km, then re-used for another sequence between 1500-3000km overall usage, and then again for another 3000-4500km usage.
The idea is that you prolong the longevity of the cassette buy always using a fresh chain & then using a suitably worn chain on a suitably worn cassette so that the chain & cassette meshed well enough to get good performance.
It does work really well - you get optimum wear out of your kit & you don't end up throwing out chains semi-worn or replacing cassettes early because they're slipping when new chains are fitted.
If you add in running an A bike & B training bike, you can always run the fresh & newest chains / cassettes on the A bike & then run out the older kit on the training bike & in bad weather when you might otherwise get annoyed that you were using your nice new kit in bad weather.
Here is the science bit
Any chain with a .5 inch pitch will mesh perfectly with a Chorus cassette. What causes the chain to become longer is wear between the pins and the bushings formed into the inner side plates. In that regard, Campy is better than any brand that I tested, by a large margin. There's more to chain wear than just elongation, however. If you use a precision 12" rule to measure the elongation, you may find very little, even after 4-6,000 miles, with a Campy chain, but that does not mean that it's not shot. Roller wear can also affect the cog profile and cause some of the most-used cogs to skip with a new chain, even if the chain has far less than .5% elongation. I found this out by using a Campy 10 chain for 6,000 miles with a new cassette. The chain showed a true elongation of only 1/16 over the entire length, or about .15% elongation, but the rollers were shot. The OD was about .005 inch smaller and the ID about .010 inch larger. The side clearance was also nearly twice that of a new chain. A new chain skipped on the most-used 19T cog.
If you really want maximum cassette life, alternate the use of 3-4 chains and change each one before the roller spacing increase from it's original .200 inch to .215-.220 inch. Shimano or KMC chains will measue in the .210-.215 inch range when new, but the .015-.020 inch increase in roller spacing still applies. Shimano or KMC chains may elongate up to four times faster than a Campy chain, so elongation should also be monitored. Alternate chains before reaching .25% elongation.