Some thoughts on marathon training:

My next marathon is the Marine Corps Marathon in 22 weeks, on October 28, and I've been thinking lately about how to train for it.

A bit of background: this will be my fifth marathon. I started running seriously in the Summer of 2009 (at age 34), after jogging occasionally for years. (In a previous life, I ran cross-country for one year in high school, my junior year, but wasn't any good. I also did some triathlons as a teenager, but running was by far my worst triathlon event, perhaps because I hardly trained for it. I was a good swimmer and a tolerable biker in those days, before bulking up from weight lifting in college.) My first race (in modern history) was a half-marathon in the Fall of 2009, which I ran in 1:42. My first marathon was the 2010 MCM, in which I ran just under an 8:00/mi. pace for about 17 miles, then hit the wall hard, walked and jogged the rest of the way, and stumbled across the finish line in 3:53. A little over two months later, I ran another 3:53 at the Disney Marathon, this time running the whole way at a consistent pace with a friend, not trying to go faster. Three and a half months after that, in March 2011, I ran a 3:29 at the National Marathon. This was the first marathon in which I ran hard the whole way without exploding, though I felt like it was an off day for me and I could have run faster. The last marathon I ran, my fourth (in one year!), was the 2011 MCM in 3:19. I was on 3:15 pace (about 7:25/mi.) for about 20 miles before slipping to 8:00/mi. for the final 10k.

I have never trained appropriately for a marathon. For my first one I did the typical beginner's approach: gradually increasing my weekly mileage and long run until both peaked 3-4 weeks out (at just over 50 miles for the week and a long run of 20 miles), and then tapering. After recovering from that debacle, I kept my weekly mileage around 50 and did 18-22 mile long runs nearly every week, which led to my improvements in the Spring of 2011. But I did no speedwork. So last Spring I started running short (5k-10k) races frequently as a form of speed training. In the Summer and Fall of 2011, I ran a short race nearly every weekend and did almost no other running faster than easy/long pace, except for a few marathon-pace runs leading up to the MCM. In the 18 weeks leading up to the 2011 MCM, my average weekly mileage was 53 miles, with my heaviest week in the high 60's, and on average I took one day off each week. But those off days were not distributed evenly - they clustered around periods when I was sick, nursing some minor injury, or tapering.

After the 2011 MCM, I decided not to run another marathon for a year and instead to focus on getting faster at shorter distances. So over the Winter I started doing regular speedwork for the first time (not including races), loosely following Jack Daniels' 5k-15k training plan. Two track workouts a week led initially to some improved times in short races but then, in January, to a mild injury that sidelined me for a few weeks. In February I started following a more scaled-down version of Daniels' plan, which has led me to much faster times in races this Spring from 5k to 10 miles.

I'm inclined to rely on Daniels for my marathon training as well, in part because his plan for shorter distances has worked well for me, and also because it focuses heavily on the key element that I think my previous marathon training has lacked: tempo (or threshold) runs, especially embedded in long runs. Another book I've been looking closely at lately, Advanced Marathoning by Pete Pfitzinger and Scott Douglas, claims that "[a] high lactate threshold is the most important physiological variable for endurance athletes," and that "it's your long runs and tempo runs that have the most relevance to your performance on marathon day" (5, 146). I suspect that is true, but I have never included tempo runs in my marathon training at all, and even this year when focusing on shorter races I have underemphasized tempo runs in favor of faster intervals. So this time around I want to make tempo runs the core of my marathon training plan.

Pfitzinger and Douglas, in spite of the passages I just quoted, do not make tempo runs central to their plans. They start you off with tempo runs every other week, and on the weeks in between they have you run marathon pace (starting at 8 miles and increasing in later weeks) during long runs (starting at 17 miles and later increasing). About halfway through their plan, the tempo runs disappear (although the increasingly difficult marathon-pace long runs continue) and are replaced by 5k-pace intervals and short (8k-15k) tune-up races.

By contrast, Daniels' marathon training plan has a lot more tempo running in it, and it puts the 5k-pace intervals at the beginning instead of at the end of the cycle. For the first six weeks, you do a tempo run or tempo intervals each week along with 5k-pace intervals. After that, you do tempo runs twice most weeks. Each week includes either a long run or a long workout mixing easy- and tempo-pace running, and a shorter workout of tempo intervals. A few, progressively longer marathon-pace runs replace the long runs about once a month.

Let me use some numbers to illustrate why I prefer Daniels' approach. Most of my tempo runs lately have been at Daniels' VDOT 55 or 56 paces, which are 6:15-20/mi. tempo pace and which (he says) correspond to a marathon pace of 6:37-43/mi. That's a 2:54-5 marathon, which is way faster than I've ever run. How should I train in order to run that fast, or at least faster than I have run? In the past I've done lots of long runs, so it seems that the missing element is tempo-paced running. Notice that tempo pace is about 25 seconds faster than marathon pace at my current level, according to Daniels. Marathon pace runs are important for getting used to how that pace feels. But if the physiological benefits I've been missing come from tempo-pace running, then I should emphasize that instead of marathon-pace running. It's tempo pace (under 6:20) that I should eventually mix into my long runs, in the form of tempo or (as Daniels says) cruise intervals, rather than marathon-pace running (around 6:40), though a few tune-up marathon-pace runs are still necessary. That's my main reason for preferring Daniels. Another reason is that I don't see the point of introducing 5k-pace intervals late in the marathon build-up, as Pfitzinger does. You need to do some running faster than marathon pace, but tempo pace fits that bill too.

Another reason I'm turned off by Pfitzinger is that I don't think I could handle 8 miles at 6:40 pace at the end of a 17 mile run right at the beginning of my 18-week marathon training plan (a few weeks from now). Either Pfitzinger thinks that I should already be in that kind of shape before starting his plan, in which case I wonder what his plan is for; or he thinks that my marathon pace should be much slower than Daniels does. But I think Daniels has my goal pace right, or at least he has the pace right that I should train towards, though of course a few weeks before the actual race I should decide on a realistic pace based on how my training and tune-up races have gone. Two months ago I ran 10 miles at 6:27 pace and felt confident that I could have held it for a half-marathon, so it's not crazy to think that with the right training I might run 6:4? pace for a marathon five months from now. The question is: what is the right training?

Still another difference between Daniels and Pfitzinger is that while both effectively have you do long runs every week, Daniels sprinkles faster running into your long runs more often than Pfitzinger does. So Daniels has you do fewer long runs simply at easy/long pace. Last year I did slow long runs most weeks, and I think I got about as much out of that as I would have from doing them every 2-3 weeks. I would be better off redirecting the energy I'd save from doing those less often to long workouts with a lot of tempo-pace running. But it will take some time to work up to a point where my body can handle that.

Rule number one in running is, of course, don't get injured. I know from experience that Daniels' training plans are very difficult, but I have gotten good results from following a scaled-back version of his 5k-15k plan. If I can manage to avoid getting injured while following a (slightly less?) scaled-back version of his marathon training plan, then maybe I can run closer to my potential in the 2012 MCM than I have in the past. As for mileage, I think it's reasonable, given the trajectory I've been on, to aim at averaging around 65 miles per week in the 18 weeks before the race, with 2-3 weeks over 80. I don't want to do doubles, so I'll have to keep my easy runs longish and take as few days off as possible. All the more reason to take it easy the next few weeks.