Pacing and negative splits

I've heard of negative splits.  I've even run negative splits in training runs.  But negative splits in a race?  Forget it!  Usually when I run in a race, I try to hold back at the start, to hold some energy in reserves for later in the race, but it's never enough.  Negative splits constantly elude me.

The official results page from White Rock lists the times and rank for 10K, 1/2 way, 20 miles, and finish.  I don't feel like figuring the paces for each of these, but it's instructive to look at the relative rankings of the winners.  The following is the rank at 10K, 13.1 miles, 20 miles, and at the finish for the top 10 males:
9, 5, 1, 1
7, 6, 2, 2
5, 7, 3, 3
10, 10, 6, 4
14, 13, 8, 5
4, 3, 4, 6
1, 1, 7, 7
2, 4, 10, 8
13, 14, 12, 9
15, 15, 15, 10
From this you can't tell how their pace changed throughout the race, but you can sure see that the winners were holding back.  The seventh and eighth finishers must have had some confidence at first--or maybe they knew those Kenyans (1-5 were all Kenyans) were just a few paces back and they would inevitably close in. . . .

In my case, I'm guilty of going way too fast out of the gate.  I felt like I was doing great.  But check out my places, again at 10K, 13.1 miles, 20 miles, and finish: 403, 477, 666, 1241.  My pace was steady from 10K to 13.1 miles, but others were gaining on me.  I lost some ground to mile 20, then a lot of ground the last 6 miles.  Compare these results to Mark's: 1193, 1139, 1048, 1027.  I'm not sure he ran negative splits, either, but at least he gained position throughout the race instead of losing.  To further illustrate, according to the White Rock data, Mark passed 129 people in the last quarter of the race, and 32 passed him.  Me?  During the last quarter, 436 people passed me, I passed 1.  Ouch.  As I look forward to another race, I definitely need to work on pacing and finishing strong.