I find it surprising that Heroku was the breakeven point. There have been other (what I consider) rather high-profile exits, such as Cloudkick, Reddit, Etherpad, 280 North, etc. These were good companies and were acquired for significant amounts; it seems like YC would have realized some significant gains on them.