Understand pace, variation and confidence
Know what came from a result, what Driver OS calculated and what is still unknown.
Recorded, calculated and explained
Recorded evidence comes from the supplied result: laps, timestamps, ranking information and score snapshots where available. Missing values stay missing.
Calculated measures use that evidence under stated rules: pace summaries, variation and comparisons. They do not establish a cause for a faster or slower lap.
AI explanations are separately labeled model output, with sources and limitations. A fluent explanation cannot make weak evidence stronger.
Pace and variation
Best lap is one recorded lap. Representative pace summarizes the eligible retained laps; variation describes their spread. Read the measure's label and the laps behind it instead of treating these as interchangeable.
Retained and excluded laps remain distinguishable. A single comparable visit gives you a baseline. More visits can reveal a pattern, while missing layout, class, fleet or purpose can still limit the comparison.
Position and K1 score
Fastest-lap rank is not finishing position. Finishing position alone does not show overtakes; positions gained needs a known start and classified finish. Unknown ranking basis remains unknown.
US K1RS score snapshots are separate from pace. Annual resets mean cross-year values should not be joined into one continuous improvement line. Driver OS does not invent field strength when source ratings are missing.
Dates, venues and chart context
Use the venue-local date and time for session and kart review. Missing source time is not midnight. Earlier years are shown explicitly when needed so different seasons are not confused.
Home’s observed history includes all session types at one selected venue. The average line is a median of session retained-lap means; the fastest line preserves the minimum from all valid retained observations, including shorter sessions. Average and variation require at least five retained laps per session. The fastest-recorded headline uses the minimum retained lap across full venue history, including shorter recorded sessions, and stays independent of chart range. Purple dashed baselines mark the first displayed eligible period only; no later period substitutes for a missing first period. Variation colors describe the displayed range, not performance targets. Monthly points give visits equal weight and preserve the fastest valid retained lap. These observations are not like-for-like improvement claims. Progress and kart comparisons still use their selected matching contexts. Ten-row pagination limits the visible list, not the analytical evidence. Public venue times and events are dated reference snapshots, not live feeds.
Sensor estimates are a different kind of evidence
Phone and Watch signals are not official timing or calibrated vehicle telemetry. Delivery, storage, matching and analytical usefulness are separate. A saved source or moving lap illustration does not prove a measured driving line.
Review a session · See your progress · Read the advanced sensor guide
Previous interface referencePreserved guidance · 2026-09-28
Notebook evidence and public references
All recorded laps remain visible alongside the retained subset used for pace and variation. Statistical exclusions describe a calculation decision, not an observed cause. Personal comparisons require matching owner, venue, layout version, class, fleet and purpose; visit summaries do not pool unrelated laps.
Field averages require the supplied comparable metric across the complete field. They average driver summaries, not every driver's pooled laps. US K1RS/RPM snapshots stay within their scoring year; missing scores and expected finishing positions are not reconstructed.
Public K1 tables show their checked date, exact scope and observed period. A rolling “Best of Month” table need not be a calendar month; an undated track record is labelled separately. These references cannot grant an official leaderboard position to your result. Lap-duration animation is illustrative uniform motion, not measured location or speed.
The broader calculations and optional research tools below remain available through More tools.
Official result
An imported result keeps official laps and available standings context. It remains separate from any public benchmark.
Published benchmark
A static track record and rolling leaderboard are different references. Each needs an exact track configuration, class, source link and retrieval time before comparison.
Driver OS calculation
Representative pace, lap-time variation, confidence and comparisons are deterministic calculations from saved results. Raw laps remain visible when an excluded lap is omitted from representative pace.
In All driving, Driver OS first calculates a representative pace for each race day, then expresses it as a percentage from the best representative pace recorded at that same track and in the same racer class. A smaller percentage is closer to that track's best. Separate track lines share the percentage scale but are never connected.
Overall repeatability expresses representative lap-time variation as a percentage of representative pace at each track. Choose one track in Driving View when you want the original seconds. A track with only one usable visit is labeled as a baseline instead of being called improvement.
Experimental sensor estimate
Route, lap, braking and motion labels describe available device evidence. They are not official timing, calibrated vehicle telemetry or a proven cause of performance.
Race Day Review
Each review keeps one exact venue, recorded layout, class and layout version together, using venue-local dates and official timing. Long gaps or unresolved timing can split a visit or require review. Different layouts never share a lap-time calculation.
The day’s representative pace is the middle value of its available per-race competitive averages. Typical lap variation is the middle of the retained per-race variation values. The longer personal baseline uses the middle of prior comparable visit values, so one unusually long visit does not outweigh every other visit. Lower seconds mean faster pace and tighter variation.
A lap is called clean or affected only when individual-lap evidence says so. Excluding a slow lap from representative pace does not prove traffic, a yellow flag or a driver error. Full-lap spread includes all retained official laps and appears under its own disclosure. Cloud-saved conditions are contextual; device-local tags and notes are not part of this review or its AI packet.
Earlier journal results with unresolved official provenance can preserve a faster personal best, but cannot establish a new official best or fill a missing comparable visit. The review explains when this limits the history comparison.
The usual best lap is the middle of the best laps from earlier comparable visits; it is separate from your all-time personal best. Podium and win trends show the share of known finishes in each visit. Field-strength trends average the observed field ratings, giving each rated race equal weight. Fastest-lap trends appear when verified comparisons exist, include ties and leave missing comparisons as gaps. A P1 finish does not by itself prove the fastest lap; that label and a zero lap-time gap require an explicit field-best time.
Highlighted slowdowns show both the actual lap time and how much longer it was than a usual lap. Recovery shows the following lap and the reduction in time. These changes do not establish what caused them. The evidence notes are intended for every driver, including those new to racing.