This is why I think "bad losses" is a rather overrated factor. Who cares? At least for the teams not at the top of the bracket, every NCAA tournament game is against a good team (though, obviously, the tournament has been partially watered-down, but still). Showing an ability to beat good teams should be the most important factor. I think that's largely a different issue all together. Best wins give a measure of how good a team can be at their best. Bad losses give a measure of how bad a team can be at their worst. Both are valid considerations for at-large worth. Of course, how applicable a bad loss or two in Vegas six months before tournament selection is to a team's at-large worth is a separate discussion... Yup, well its been said that the committee weighs the last 10 games more than the first ten, but I'm sure they put all of a team's 'quality wins' and especially 'bad losses' on an equal level. Now I'm almost certain they don't consider that anymore. It's a ridiculous metric to use with such wildly different schedules within a conference, let alone across conferences. I don't know exactly what is used, but I remember people talking about Michigan's tournament chances being better last year because they were hot at the end of the season and that's where I read that the record in the last 10 games is part of the process. But they could have been talking out of their asses. The argument was that if comparing teams on the bubble, they would favor the team that is playing really well at the end of the season, as they would hypothetically make a tougher team come tourney time. I do hope that's bs because like you pointed out, there are too many differences in schedules to use it as a tool to compare teams. Otherwise, the team that gets a healthy dose of Penn State and Nebraska at the end of their season gets a boost over a team that has Duke and UNC multiple times in the last 10 games.