If you've ever trained for a marathon, you've probably hit the moment where you have to pick the goal time before training really starts. It's one of the more consequential decisions of the whole cycle, since every interval, tempo, and long run over the next four months ends up paced off that number. Pick something too aggressive and you spend the build grinding paces you can't actually hold. Pick something too conservative and you train comfortably for a race that ends up not really mattering.
Here's the approach I've landed on, which seems to work pretty decently.
Start from a recent shorter race
The best input I've found is a 10K or half-marathon from the last two or three months. Shorter races give you a pretty clean read on current fitness without all the distortions a marathon picks up. A marathon time gets tangled with the long taper, a bad weather day, spotty fueling, and the way the mile-22 number you remember isn't really the runner you are now. Easier to work from a recent race that doesn't carry all that baggage.
Run that race through a VDOT table or a Riegel calculator and you get a marathon-equivalent fitness time. A 40:00 10K lands at roughly 3:05. A 1:30 half-marathon lands at roughly 3:09. These aren't predictions of what you'll actually run on race day though. They're closer to a ceiling: the marathon a runner with your current fitness could run if they were marathon-trained, executed well, and got reasonable conditions.
If your last marathon was more than a year ago, I'd just throw the time out and train off a recent shorter race instead. Marathon-specific fitness decays pretty fast when you're not in a marathon block.
How big a jump is realistic
Goal-driven training is built on the idea that you train at the paces of the runner you're trying to be, not the one you are now. That works decently well when the goal is close enough to actually reach by race day. The trap, though, is picking a goal so far above current fitness that every workout becomes a survival exercise.
For a well-trained runner, a 3–5% time improvement in a cycle is in normal range. That's a 3:10 marathoner going for 3:00, or a 4:00 marathoner going for 3:48. Bigger jumps definitely happen, but they usually come from a structural change like adding 20 miles a week or going from one quality day to two, not from picking a more ambitious number out of the air.
If you've never run a marathon, take your most recent reliable 10K, run it through an equivalency calculator, and then add a fade factor of 5–10% slower per mile if you have a deep aerobic base, or 10–15% if you don't. A 10K really only reflects threshold fitness; a marathon also asks for the durability to hold pace at mile 22, which is what the fade factor is trying to account for.
Quick sanity check I'd run: if your current weekly mileage isn't already close to what your goal calls for, the goal probably isn't realistic for this cycle. You can fix that with a base block before the build starts, but you can't really fix it inside the build itself.
Course, weather, and the rest of your life
The course is more of a variable than most people give it credit for. A flat certified course (Chicago, Berlin, CIM, Indianapolis Monumental) is a different goal than a hilly one (Boston, NYC). The difference between flat and moderately rolling is often 2–4%. That's seven to ten minutes on a three-hour marathon, which isn't trivial.
Weather is the same kind of variable. Spring marathons in the U.S. coin-flip between perfect and 70°F race-day temps. Fall marathons can hand you a heat year. So what should you actually do with this? Mostly, just be honest with yourself about what you're aiming for. A goal that only works in ideal conditions isn't really a target so much as a best-case ceiling.
The rest of your life is on the plan whether you write it down or not. A 3:00 goal on six hours of sleep is a different race than the same goal on eight. A training cycle that lands on top of a job change, a new baby, or a heavy lifting block isn't going to feel like one that doesn't. I don't think the right move is to lower the goal preemptively though. It's more about watching what's actually happening week to week and being willing to adjust pretty quickly when something changes.
Where the goal time shows up
Once it's set, the goal time really does show up everywhere in your training. Easy pace ends up as a percentage of it. Long-run pace stays within 30–60 seconds of it. Tempo, threshold, interval, and rep paces all derive from it, either directly through a VDOT table or indirectly through heart-rate zones anchored to it.
Which is why neither extreme tends to work out. Too aggressive and your easy days run too fast, your tempos creep up into threshold territory, and you accumulate fatigue you can't really sleep off. Too soft and your tempos won't build the threshold tolerance you need on race day, and the whole build feels weirdly easy while you're actually training for a harder race than the paces suggested.
I built Wave to take this pretty literally: the goal time is the input you give the app, and the whole pace ladder regenerates whenever you change it.
Updating it mid-cycle
A tune-up race three to five weeks out is probably the cheapest check you'll get on whether the goal still makes sense. A faster-than-expected tune-up means the goal can move up. A slower one means it should probably move down. Most cycles I've seen break do so because the runner refused to move the goal down when they should have.
Big disruptions mid-build, like an illness that takes a week off, easing back from a minor injury, or a work spike that ate two long runs, should pull the goal toward what you've actually trained for, not what you hoped you'd train for. The goal you pick at week one doesn't have to be the one you chase at week sixteen. Race day is the only part that really counts.