A race result is three separate updates
Crossing a finish line changes more than one thing in your record, and the parts change independently. There is the result itself. There is possibly a new personal best. There is possibly a new threshold, because a hard race is one of the better tests you will do all season. And there is a goal that is now finished.
Every training platform has a place to put those. A conversation does not. If you raced on Sunday and open a chat on Monday, nothing you did on Sunday is in it.
The old number keeps producing paces until you replace it
This is where it costs you something. Threshold pace is what run prescriptions are derived from, so an assistant working from a pre-race figure will keep handing you the paces that matched the athlete you were before the race.
Take a 10K that goes from 41:14 to 39:40. That is a meaningful jump, and it moves threshold pace with it. Ask for a tempo session the following week without saying so and the target comes back derived from the old number, comfortably slower than it should be. The session is not wrong in structure. It is prescribed for someone slightly slower than you now are, and it will feel easy in a way that is easy to mistake for good form.
The same mechanism runs the other way after a disappointing race or a spell of illness, where an optimistic old figure keeps prescribing work you cannot currently hold. Either way the fix is a value that carries its own date, which is what Athmex stores. What a stale anchor does to everything derived from it is worked through on the cycling side.
Not every fast time is a personal best
Before updating anything it is worth being strict about what the result actually was, because a number recorded loosely is worse than no number. A 5K split inside a half marathon is not a 5K personal best. Neither is a solo time trial down a hill, a point-to-point with a tailwind, or a treadmill effort.
The question to answer is what the effort was: a race, a standalone effort, or a split inside something longer, and whether the distance actually matched. Recorded honestly, the figure stays useful for years. Recorded loosely, it raises the bar on every future comparison for no good reason.
What is worth updating, and when
Within a few days of the race, while you still remember the detail:
- The result against the event: finish time, and position if it matters to you.
- A new personal best, but only if the effort genuinely qualifies as one.
- A revised threshold, if the race was long enough and hard enough to be a fair test, with the date of the race rather than the date you got round to typing it.
- The goal itself, closed rather than deleted, so the record of having done it survives.
- The next race, if you already know what it is, which is what turns the following weeks into a block rather than a series of sessions.
What not to do is overwrite the old threshold in place. The previous value and its date are what let you or an assistant answer whether this block worked, and an overwritten number cannot be compared with anything.
Where Athmex fits
Athmex is built around exactly this update. A result is recorded against the event it belongs to, and a genuine new value creates a new dated version rather than replacing the old one, so the threshold from March stays visible next to the one from September.
Corrections are handled separately from new values, which matters more than it sounds: fixing a typo should not create a fake improvement in your history. Goals are closed as completed rather than removed, and personal bests are stored for running distances with the date each was set.
Because every value carries the date it applied, an assistant reading your context also receives how current each one is, and is told to say so rather than reason silently from a figure from last season. Preparing for the next race is a different job, covered separately.
Common questions
Should I update my threshold after every race?
No. A short race or a bad day is poor evidence for a threshold change. That is a training judgement rather than a software one, and the practical rule is narrower: whenever you do change it, record the date it applied so whatever reads it later can weigh it.
Does my training platform not handle this?
Platforms handle their own copy of it well. The gap is that your assistant is not reading your platform's threshold field unless a connector is supplying it, and even then a derived value is not the same claim as one you set deliberately.
What about a race I did badly?
Still worth recording as a result. A slower time on a hot day is context an assistant can use, and it is a fact about the season rather than a failure to hide.
Does Athmex work out my new paces?
No. Athmex stores the values you confirm and their dates. Claude or ChatGPT does the calculating, and it does it from numbers you have checked rather than inferred.