Agreed, I would bet $5k on a 1700 player in a heartbeat vs. curtains with rook odds. If you honestly think he's a favorite in that game, I don't think you understand rating strengths (1700 being 1700 USCF not some other unknown rating system).
I haven't played regular time control chess for a good 15 years, but back when I did a friend of mine who was a chess tutor said he'd put me at about 1800, when I'm not being an idiot. I've also played a bunch of blitz check in the last 15 years and when not being an idiot was at about 1800 (I understand there's some ratings inflation there and so that was probably equivalent to 1600 or so USCF).
So while you guys (at least curtains, Lalu) have a much better sense than I do of what a 1700 USCF could do, I didn't think I was completely talking out of my ass, especially if this game would be a blitz game. I don't think that while I was around my best that I could take someone of curtain's caliber down 75% of the time, as one person put it, or 95% of the time, as I believe Lalu put it.
At blitz, I think the 2600 would be a favorite (if not a huge one), though that might not be true for a longer game. Even with a longer control, though, I really can't see the 1700 player straight-out WINNING 95% of the time, as was claimed.
So (and now, I *am* pulling some #s out of my ass), I'd expect a the score in a 10-game blitz match to be something like 5:5 or 6:4 in favor of the 2600, and the score to be something like 3:7 in favor of the 1700 if there was a longer control.
your friend is probably being too nice to you.