Pacing decisions
How to choose a marathon goal time
A useful goal brings together a recent race, repeatable training and the course you will actually run. A calculator supplies a starting estimate; it cannot decide how much uncertainty you should accept.
Why a prediction is not yet a goal
A short race measures performance over that distance on that day. Extending it to 42.195 km adds assumptions about fatigue, preparation and conditions. A neat result such as 3:59:30 can look more certain than it is. Round your planning decision sensibly and keep the exact checkpoint arithmetic for later.
The opposite mistake is treating one difficult workout as proof that the whole goal is wrong. Check whether that session was unusually hot, hilly or poorly recovered before drawing a conclusion.
Four inputs to review together
- Recent performance: use a representative race with a known distance. Note its date, terrain and how you paced it. An old half marathon is not automatically more informative than a recent 10K.
- Training continuity: review ordinary weeks, missed weeks and recovery, rather than your highest mileage week. There is no mileage cutoff that makes a particular finish time safe.
- Longer-run experience: consider how effort developed and how you recovered. One completed hard session cannot certify marathon readiness.
- Race circumstances: consider elevation, exposure, forecast, aid stations and your familiarity with the planned fuel. Revise the goal if those assumptions change.
Use the time predictor to explore estimates, then the pace calculator for exact splits.
Write outcome goals and decisions separately
An A/B/C plan can help you stay engaged when the original time becomes unrealistic. It is not a probability forecast. For example, A might be a time target supported by your build, B might be a slower revised time, and C might be finishing with controlled running if conditions allow. Stopping for a health problem remains a separate decision.
Write the action beside each goal: what will make you reduce effort, when will you reassess, and what will you stop chasing? Do not set A as an unsupported pace simply because you have a B plan.
What the numbers can—and cannot—tell you
For an illustrative 1:45 half marathon, the power-law conversion T₂ = T₁ × (D₂/D₁)k gives about 3:38:55 at k = 1.06 and 3:41:58 at k = 1.08. Those are two model scenarios, not the boundaries of your possible result. Neither exponent is assigned reliably by a weekly-mileage band.
My 2025 and 2026 Run Melbourne comparison shows why the shape of a race deserves attention too. In 2025 my recorded 5 km blocks became progressively slower; in 2026 the last comparable block was quicker than the first. I felt the difference as a stronger finish. It did not establish a guaranteed marathon pace.
My later Sydney Marathon review adds a second check: half-marathon fitness and a good individual workout did not deliver my sub-four goal.
Avoid these shortcuts
- Selecting the fastest result from several models while ignoring their shared assumptions.
- Using a peak week as if it represented the whole training block.
- Treating pain, illness or disrupted recovery as a reason to squeeze in one last proof workout.
- Choosing a fixed time first, then explaining away every piece of contrary evidence.
Turn the decision into a race plan
Record your chosen target, opening checkpoints, likely difficult course sections and a fallback action. Check that a deliberately slower opening is included in the total: extra seconds early must be recovered later if the finish target stays unchanged. See the negative-split arithmetic.
Review the readiness questions using sessions already in your plan. You do not need to add a demanding test to earn permission to choose a goal.
Questions about goal selection
Which race should I use?
Use the most representative recent result, with its distance and conditions understood. A half marathon is closer to marathon demands than a 5K, but recency and execution still matter.
Does high mileage justify the aggressive estimate?
Mileage provides context, not a validated switch between model settings. Consistency, recovery and individual history also matter.
Can I change the goal on race morning?
Yes. A changed forecast, illness or another material change can make the original target inappropriate.
Sources and limits
- Blythe and Király (2016), Prediction and quantification of individual athletic performance of runners
- Smyth (2018), Fast starters and slow finishers
- Halson (2014), Monitoring training load to understand fatigue in athletes
Worked examples are calculations for the situations described. Adapt them to your own training.