After more than half a year, somebody else agreed with me that Matt Fitzgerald’s concept of a marathon simulation is a little unclear in 80/20 Running. Here’s an attempt to add some clarity: (although as this is the opinion of a Person On The Internet, you may take it with as much of a pinch of salt as you feel necessary. Or post enraged disagreement in the comments of this blog)
You need to be ready
You need to run your simulation when you’re race-fit; it should be a few weeks prior to the race – much further back and you won’t be peaking, much closer and it will wear you out too much before the race itself.
You need the same gear
You never want to do something in a race for the first time, rather than having tested it out while training. So the simulation run should include the same clothes, the same footwear, the same drinks, food, revolting gels, hair cut as you’re using in the actual race. Footwear is particularly important because the volumes you’ll do in training will tend to wear out shoes (and they’ll stink), but you probably shouldn’t be stepping into a box-fresh pair on the day of the race, only to find they aren’t quite what you were expecting.
You need the same course
Try to make sure that the terrain you’re running your simulation on is similar to what the race will be. For example, there’s a big hill at the 37th km in the Osaka marathon. The first time I trained for this, all my training was on the pancake-flat jogging paths of Singapore’s Park Connector network. Suffice it to say, walking up that hill while weeping wasn’t one of my happiest moments.
Depending on where you live and where the marathon is, that may be simple or difficult. Singapore has a few minor hills which makes practicing for hilly marathons more challenging. It also has ludicrous heat and humidity, but you can’t safely assume that if you can cope with the heat and the humidity, you automatically get a power-up when you run anywhere cooler and drier. (Eg when I went and ran at altitude in Boulder, I was just schooled in another form of physical incompetence.)
Oh, and if the race is at midnight, try and do a run around the same time. Fairly obvious, but may not be very practical.
You need less distance
Almost contradicting the previous point, you don’t simulate the marathon by running 26.2 miles. That isn’t a simulation. (I did read in Bowerman of training runs the athletes did of more than marathon distance, but that just seems mental.) My sense (and it needs testing) is that the best distance to run to shake things down properly, without wrecking yourself, is about a half marathon distance. But with the following condition:
You need to be similarly fatigued
The hard part of most marathons is the second half, when you’re knackered and trying to push through. So your simulation run should be on the Sunday, after you’ve done a fairly hard, long run on the Saturday to toughen yourself up. I’d recommend 10 miles on Saturday at race pace, and then 13.1 miles on Sunday at race pace. That should simulate things good and proper. If you can’t do 23.1 miles over two days, you’re probably not adequately prepared for the marathon. Adjust your pace if that’s the case, or otherwise change your expectations.
You need to run the same speed
And here is the hard bit that Fitzgerald didn’t specify – if you’re running your easy runs at 5:40 km pace, and your hard runs at 4:40, what should be your marathon pace?
You might have a rule of thumb like “race pace should be recovery pace – 30 seconds per kilometre” but then you could spend a lot of time worrying about whether that was fast enough. But we don’t need to worry – we have McMillan’s Race Pace Calculator. What you could then do is a 5k session as one of your hard sessions in a week close to the simulation, and then plug your 5k time in and see what that predicts for your marathon pace.
And then hope. You always have to hope…