Field note
A personal record is the fastest you have ever covered a distance. A recent best is the fastest you have covered it lately. Same stopwatch, two windows. Neither one is a race forecast.
Both sides are runs you actually did. An estimate that invents a time you never ran is a different kind of claim.
If you covered the distance recently, your recent best is your fastest stretch inside that window. It is measured the same way the all-time record is, over a shorter period. How a record stretch is found inside a longer run lives in What are personal record distances?
| Read | Plain meaning |
|---|---|
| Record | All-time fastest exact stretch of that distance |
| Recent best | Fastest stretch inside a recent window (often about a month) |
| Gap | How far lately sits off the gold mark, as a share of the record |
Distances you have timed lately sit next to both numbers. Distances with no recent run usually show only the record. Filling that gap with a model is optional and should be labeled as an estimate, not as a measured best.
Each distance is read on its own. A fading 5K and a rising marathon can sit together honestly. Holding today's recent best next to the same read from about 30 days ago names whether lately got quicker or slower. Both sides are still runs you did, not a model of race day.
This is not a race forecast. Hills, heat, taper, and how you feel this morning stay out. A flat-course assumption hides terrain. Scaling one distance into another with a formula is a separate step from reporting a best you actually ran.
Riegel (1977; 1981) and Daniels (2014) give common ways to relate times across distances. Those formulas assume endurance scales the same for everyone. For many people they overstate the marathon if you do not train that far. GPS times are not chip times. Treat any scaled number as a read of recent form, not a promise.
Smyth and Muniz-Pumares (2020) showed you can pull critical speed from raw training data in recreational marathoners. That still rests on runs that happened. It does not invent a race you skipped.
How you spread effort across a single outing is a different question again. That lives in What is pacing? Equal pieces that show where pace drifted live in What are running splits?
Calling a recent best "close" to a record is a product choice about what is worth noticing, not a number from the research. GPS noise is worse on short marks than on long ones. A second or two on a kilometer is often a tie worth re-running.
Recent best versus record is a form board. It shows where you stand against yourself. It never promises what you will run on race day.
I write these because I build Aera, an iPhone app that reads Apple Health against your own baseline and puts the research and the error bars underneath every number it shows you, including this one. You do not need it to use anything above.