In the past, the "other guys" were the driving force behind the standards movement, so they could essentially write the standards as they built their product. They also had reason not to like MS very much (since they were involved in anti-trust litigation at the time), so they did things that in hindsight appear to be simply out of spite.

An example of this would be the CSS Box Model, which had reference implementations in both IE and NN at the time the spec was written. The spec changed how width was calculated, leaving every existing browser incompatible, and leading to silly things like Netscape.com not rendering correctly in the first released version of Netscape 6.

Nowadays, it's a mystery why MS gets left out of the loop. The standards committees are still non-MS dominated, so conspiricy-theorists could conclude that there's still some collusion between the Mozilla & WebKit teams and the W3. Or it could simply be Microsoft not wanting to get burned yet again.

The knee-jerk answer is that the IE team's ideas often seem to come from a crack pipe. Not to say that every single one of their ideas has been bad, or that everyone else's ideas always turn out great, but IE has a lousy track record. There are more than a few instances where the IE version is either internally inconsistent (many of the IE6 CSS bugs) or provably worse than the competition (JS event handling springs to mind).