Problem with benchmarks is that it cannot take account for 0 day malwares. That's why in my uneducated opinion, many quality antiviruses are slipping up and down on charts.
I'm not sure I understand what you mean, Paul, but if you're talking about the kind of benchmarks AV Comparatives and other organizations do, they can
(and, in at least the case of AV Comparatives, actually do
) take into account 0-day malware, i.e. "unknown" malware. In fact, the test I mentioned in my previous post is precisely that: a test to see how each AV performs against a number of previously unknown (to their signature databases) malware. This kind of tests allows us to see how well each program uses heuristics
to detect this 0-day malware.
I also wish there were proper benchmarks done to see how AVs really affect a machine, nothing I've seen is really comprehensive enough...
There have been tests done to see how various ISS's affected a machines boot time, but I can't find that particular site ATM - although it was last year, (or early this year), I think.
I think you're talking about the same article I remember, 4wd. Is it this one?: What Really Slows Windows Down
In any case, if it wasn't it, this article is well worth reading. It's a little outdated now (written in 2006) but very informative.