Marathoning + data, practice and perseverance (part 1)

From https://ico.org.uk

This post has been long in the making. It started in October 2012, when I started to wonder how much preparation and training help to reach a marathon goal. Common sense would refuse such a question, but my Fall marathon in 2012 was such a disappointment, I started to doubt:

year	month	place	time	90d/km	180d/km
2011	October	Kosice	03:48:50	363	450.2
2012	April	Bratislava	03:42:23	254	601.7
2012	October	Kosice	03:52:17	465	644.0

I intended to finish at 3:30, but hit a crisis at 31st kilometre, had to walk for a few, and then finished ignominiously at 3:52. However, I did improve my personal best next Spring, when I ran the marathon at 03:39:44 after running 463, respectively 480 km in the 180 and 90 days preceding the run.

I guess my belief in training very hard took another hit, when I managed to finish Two Oceans 56-ultra at 6:00:48 with some (clocked) 160 km in the preceding 6 months. I played soccer regularly, and ran infrequently, so that time looked far too good. 

I put the "training does not help" theory to the test in Fall 2014. I finished at 4:25, and felt I should quit running. Instead, I put the data together:

year	month	place	time	pace/min/km	90d/km	180d/km
2011	October	Kosice	03:48:50	05:25	363	450.2
2012	April	Bratislava	03:42:23	05:16	254	601.7
2012	October	Kosice	03:52:17	05:30	465	644.0
2013	March	Bratislava	03:39:44	05:12	465	480.0
2014	April	Cape Town	06:00:48	06:27	125	163.0
2014	October	Kosice	04:25:35	06:18	49	69.6

It clearly struck me my pace correlated negatively amount of training I put in: the corr. coefficient for 90d was -0.83 and -0.89 for 180d. Perhaps training does help, after all?

I managed to put in 170 km in the 90-day lead up to Kosice 2015, and barely made my way to the finish line at 5:36/km (3:55:57). Much better, but I felt 3:50 should have been met easily. Then came a similarly lukewarm Bratislava 2016: 03:51:39 at 5:29/km after 242, respectively 286 km.  I did feel I trained enough to finish at 3:45 (or better), but lost my way after half-way mark. (Does it help or does it not?)

I put in some training later in the spring, had a usually quiet (lazy) summer, an okay run-up to the race. I started wondering  about my goal pace the before the race. I had all this data:

year	month	place	time	pace/min/km	90d/km	180d/km
2011	October	Kosice	03:48:50	05:25	363	450.2
2012	April	Bratislava	03:42:23	05:16	254	601.7
2012	October	Kosice	03:52:17	05:30	465	644.0
2013	March	Bratislava	03:39:44	05:12	465	480.0
2014	April	Cape Town	06:00:48	06:27	125	163.0
2014	October	Kosice	04:25:35	06:18	49	69.6
2015	October	Kosice	03:55:57	05:36	186	285.5
2016	April	Bratislava	03:51:39	05:29	242	286.4
2016	October	Kosice	?	?	210	520.0

The 90-day aggregate was nothing special (it was better in April), but the 180-day seemed quite decent. I also recalled I had been able to put in a 32k long run at 5:33/km, and had felt good at the end of it. In the end, I went for 5:19/km, which should result in 3:45, my third best result.

In spite of some difficulties (part 2), I finished with my personal best 3:38:30, at 5:11/km, and negative(!) splits. (This also confirmed that 180-d aggregate is a much better predictor of race pace.)

To be continued.