I had the same approach to my first tri back in June (olympic-distance). I went reasonably hard on the bike -- after all, it's a RACE!! -- but there were certainly points where I deliberately held back...in particular on hills where I knew that hammering would give me some burn in my legs. My bike split was 1:10:10, about 21.4mph...about what I do in training. On the 10k I felt great and laid down a time of 44:40, which was only about 30 sec slower than my open 10k PB.

My conclusion was that I didn't bike hard enough.

In my second tri (also an Olympic distance) my goal was to bike so hard that I couldn't run at the end. Literally I wanted to be forced to walk the last couple of miles, so I could "find my limit" and better understand the interplay between bike intensity and ability to do the run. This race was truly an experiment ahead of my big "A" race (the Chicago triathlon).

I executed my plan on the bike and I turned in an average bike speed of 22.4mph (on a hillier course, although one that the race organizers conceded was only 38km and not 40km). The first mile of the run was absolutely brutal. It felt like I was running incredibly slowly, and like many more people passed me than in the first race. But strangely, by the second mile my running legs got going and I settled into a decent pace despite a run course with some pretty decent hills. My Garmin was telling me I was doing okay. At mile 5 when it was time to "empty the tank" I was able to pick up the pace, although my quads tightened up a bit. Overall my 10k run time was 45:08 -- I gave up 30sec on the run to gain 1mph (3:06) on the bike.

The big lesson for me here was that I didn't take a major hit on the run by going all-out on the bike.

So my advice is to experiment a bit and find your bike effort limit. I am still experimenting to find mine.