I don’t know any scientists who have a success rate of their experiments greater than about one in ten. Ninety percent, total flops, learn nothing, something was dreadfully wrong, just wasted time. Of the ten percent that actually succeed, maybe ten percent of those actually contribute in some way to new knowledge, and the rest just sort of confirmed something that was already known. So if one is going to do this — and don’t get me wrong, it’s the most exhilarating thing in the world to do if you decide that this is your calling — that one has to expect failure. If you’re doing experiments that work all the time, you’re not working on anything very interesting. You’re not really at the cutting edge, you’re just sort of replicating things that were already known. And that’s a hard thing at first to get used to.