Marathon and Predicted Results

I've already sort of settled in on 3:40 as my goal time for the Clarence DeMar Marathon. As a sanity check on the realism of this goal - and also because it was just fun - I decided to plug in my recent results into the estimated finish time calculator at Marathon Guide. What the calculator does is take your time from one distance and estimate your finish time at other distances. It does more than just a simple linear projection based on mile pace. Based on how runner pace typically varies between distances, it either discounts your pace estimating a longer distance or increases it for a shorter distance. My most recent result was the Seacoast 10k. Plugging in the 46:23 I ran at that race I got the following results:

Distance	Guide Estimate	PR	Delta (PR-Guide)	Delta %	PR date
5k	22:08	21:48	-00:20	-1.5	6/3/15
10k	--	46:23	--	--	8/22/15
10 mile	1:16:34	1:16:52	+00:18	+0.4	6/21/15
Half	1:42:51	1:44:15	+01:24	+1.3	5/3/15
Full	3:35:25	3:52:35	+17:10	+7.4	1/11/15

The second column is the estimate from Marathon Guide. Following that is my personal record (PR), the difference between my PR and the Guide estimate, that same difference as a percentage, and the date of my PR. It's impressive how close the estimates are on my 5k, 10 mile, and Half results - all within 1.5%. So the formula they are using seems pretty good.

Marathon Guide estimates me at a 3:35 marathon based on the 10k result. So that should give some confidence that I can get sub 3:40.  Assuming the "error bars" are the same on the marathon as they are on my other races, let's assume +/- 1.5% on the marathon time, which gives a range of times from 3:32:12 to 3:38:38.

Another way to do this is to plug the 3:40 marathon time into the estimator and see what it says I should be doing for other races:

Distance	Guide Estimate	PR	Delta (PR-Guide)	Delta %	PR date
5k	22:37	21:48	-00:49	-3.7	6/3/15
10k	47:22	46:23	-00:59	-2.1	8/22/15
10 mile	1:18:11	1:16:52	-01:19	-1.7	6/21/15
Half	1:45:02	1:44:15	-00:47	-0.8	5/3/15
Full	3:40:00	3:52:35	+12:35	+5.4	1/11/15

The interesting trend here is the way my percentage decrease on the estimated time gets higher the shorter the distance, -0.8 for the half all the way up to -3.7 for the 5k. This would seem to imply I have more speed vs endurance than the "formula runner" used in the calculation. A little surprising since I think of myself more as an endurance runner than a speed runner.

The advice the experts give for a marathon is to "let the race come to you", meaning don't start with an aggressive pace but rather one more comfortable, and if later in the race you are still feeling it, then the race has come to you and you can safely increase your pace. This is my preferred approach to racing in any case. I think I can safely start with a 3:40 goal time (about an 8:20 pace), and go with how I feel later in the race.