[deadmansshoes]

All well and good, but I'm not sure doing "research at a reputable university laboratory" and having "a good recommendation from a well-known professor" can really be classed as bypassing the academic admissions system.

[scott_s]

It's not, but it's certainly outside of the "standard" sequence. The useful lesson to extract is if anyone uses one metric (say, grades) as a proxy for another (say, ability to work and flourish in a university environment), and it's possible to score points on the second metric, do so. Even if there's no official channel for it, that's what people really care about and they'll immediately recognize it as such.

The reason that this sort of thing is so rare, I think, is that we don't know what metrics matter until we're well past the point of applying to something. Even if we're told, it's hard to overcome the expectation and pressure of doing well on the proxy metrics. For example, in college, I had no concept of what mattered for grad school: research. And even if someone sat me down and told me, I'm not sure if I would really understand it. It's a rare person that realizes it's even possible to obtain points in the metric that really matters.

[hga]

That said, it was MIT's experience back when the SAT meant something (pre-1994) that class rank (i.e. grades) were one of the two best predictors of subsequent success.