I came up with my own training philosophy, which I freely admit I've rarely followed, after repeatedly failing in my attempts to run a marathon in 2:30. I ran 10K in 32 minutes and was regularly beating guys who had broken 2:30, so I figured I just needed to train as they did; their training universally had more mileage and longer long runs. So, I'd run long frequently, get in as many miles as I could and continued the 10K races as speedwork (I enjoyed them and I was good at it), though my times slowed into the mid 33's, which didn't alarm me, as I was training for a marathon and not a 10K. Each time I tried the marathon, however, I ended up running closer to 2:45 after "blowing up" somewhere on the course.
If I ran 2:45 when I trained to run 2:30, what did I need to do to run 2:30? Train to run 2:15???
Years later, it dawned on me that I shouldn't have thought of myself as a 2:30 marathoner who had a bad day and finished slow. Instead, I was a 2:45 marathoner who overtrained and went out too fast. The answer was to train better to run what I already could do. If I trained to run 2:45 and paced myself, it was likely I'd have energy left toward the end to speed up and run a personal best. [I'm aware of the irony that I denigrated negative splits in an earlier post in this series. So I contradict myself. It's a different mind-set.] There are some "natural" points at which one can decide whether one should increase speed; they come at 3 hours, 75 minutes and 10 minutes from the end of the race (which of course requires one to be able to predict a finish time); for a 5 hour marathoner, that first decision point at 3 hours comes at the same point on the course as the 75 minute point for a world-class runner.
There's a sort of Catch-22 inherent in this method: you have to race before you race. What if you've never run a marathon or it's been a long time since your last race? You can race right now, if only a 5K (assuming you're healthy, you can always finish a 5K, though it might not be pretty); if there's no race available, as is often the case here in Minnesota in the winter, measure out your own course - it'll be accurate enough.
The next "rule" is that one can race well over a 2-fold range of distances, and at least finish over a 4-fold range. Thus, if one races a 5k, one then trains to do that 5K, but at the end of the training cycle, instead of another 5K, one could run a 10K. One would then train to run 10K and race a 1/2 marathon (admittedly just out of range, but 20K races are rare). Then one trains to run a 1/2 marathon and races a marathon. Finally, one trains to race the marathon. This method is very slow, but it allows one to build mileage slowly as well.
The amount of time between races should be adequate for recovery. One should never race more than 10% of one's miles and 5% over the course of a year is preferred. For every mile of a race, one should have a day before the next race, preferably two. These two rules become the same at 70 miles per week.
I believe that there's a point where mileage becomes counter-productive. I've covered it before, but I've found that at 5K, ideally one averages 50 minutes per day, with long runs of 90 minutes; at 10K, it's 60 and 120; at 1/2 marathon 65-70 and 135-150; at marathon 75 and 2:45; at 50 miles 80 and 3:15; at 100 miles it's 100 and 4:15. For difficult trail races, one adds 5 and 15 minutes to these (for very hard races, like the Hardrock 100: add 10 and 30).
Training for any race consists of four components (repeating myself), most easily described as 1) how far you can run, 2) how far you can run at a given pace, 3) how fast you can run a given distance and 4) how fast you can run. Each of these should be included in any training schedule, but, of course, the emphasis shifts depending upon the length of the race.
My preferred method of peaking is to start with alternating short fast runs with long runs, then switch to alternating short easy runs with long runs that have speedwork at the beginning, then switch to alternating short easy runs with long runs that have speedwork at the end.
Eventually, one reaches a plateau in performance. At that point, one can try training more than once per day, but the second session should be an exercise other than running. For trail runners, hiking and biking (especially mountain biking, as it requires shifting balance and rapid changes in speed and direction) are probably best; speed-walking is also an excellent choice, especially for multi-day athletes.
Going up the country
3 days ago