RACE LOG — raw.xrk

Track🏎 2T 125cc9 laps · 8 valid✓ AiM hi-fi telemetry · RPM

● Fastest lap

0:37.36

Lap 6 · V max 97.1 km/h

Patch-best · near-term

0:37.24 −0.12s

→

Physical ceiling · motivational

−1.48s

Patch-best = stitch each segment’s own best (proven doable); physical ceiling = every corner at the grip limit (still to train). The latter already includes the former — not additive; the true ceiling ≈ 1.48s.

Lap consistency · CV

1.2%STABLE

Steady rhythm — focus improvement on corners

GPS quality CLEAN · 9964 frames · 8 despiked · 25.0 Hz

★Fix first today replicate your best—

You are already near the grip limit in every corner — focus on replicating your best lap each time (see L2 “Consistency”).

How to read this screen: the fastest lap is today’s result; the patch-best tells you how much is left by just stitching each segment’s best. The real headroom is the line work — scroll down to L1 to see exactly which corners lose time.

📷 Recording note — data range

Recording started while driver was already moving on track (start speed 28km/h). First 50s (about 739m) is an incomplete segment heading toward the start line, doesn't form a complete lap, and is excluded from analysis.

If you want complete recording from Lap 1, recommend starting recording while car is stopped in the pit, then start the timer after accelerating out of pit.

L1Where time is lostLocate which segment / corner. Not why yet.

LAP TIMES & PHASERACING laps only join optimization

Lap	Time (s)	V max	V min	V avg	Arc (m)	Phase
L1	38.44+1.08	92.1	39.1	68.1	728.2	RACING
L2	38.56+1.20	96.0	43.5	68.6	734.8	RACING
L3	37.72+0.36	97.4	44.2	69.4	727.2	RACING
L4	38.00+0.64	95.5	43.5	69.0	728.9	RACING
L5	37.44+0.08	96.1	39.9	70.1	730.8	RACING
L6	37.36BEST	97.1	39.9	70.1	728.1	RACING
L7	37.48+0.12	97.1	36.8	70.0	729.7	RACING
L8	37.68+0.32	95.5	40.9	69.6	729.1	RACING
L9	42.52+5.16	94.6	24.3	58.4	688.3	COOLDOWN

📐 What does Phase mean

Every lap is classified into one of six categories, so you can immediately tell "which laps are actually racing":

RACING: real racing laps, included in all optimization advice and comparisons.
WARMUP: opening laps, speed hasn't built up yet.
COOLDOWN: closing laps, already winding down.
INCIDENT: laps with a mid-lap stop (vmin near 0) — could be a crash, off-track, yielding, red flag, or marshal stop. Safety first — review what happened.
GPS NOISE: laps with many GPS position spikes (e.g. red lines flying off the track on the map). Trajectory and distance are unreliable, excluded from analysis.
NON-KART: not driving data at all (walking in pit / handling equipment).

The last five categories don't participate in lap-time comparisons — they're only shown for session completeness, avoiding polluting the analysis.

📉 Lap evolutionwarmup / fade / mistakes

Lap time vs lap number — spot tyre warmup, fatigue fade and mistake laps. Cooldown laps are excluded to keep the axis readable.

🏁 Sector composition

Split the lap into 3 sectors to locate which one drags the lap; each flags the corner most worth practicing.

Corners: T1T2T3

0–32% · len 231m

11.92s

BEST @ L5

Corners: T4T5T6T7

32–64% · len 238m

▸ Focus: T6 · T7

12.56s

BEST @ L7

Corners: T8T9T10T11

64–100% · len 260m

▸ Focus: T9

12.76s

BEST @ L6

👻 Ghost Lap · realistic best

Stitch each lap’s fastest S1/S2/S3 into one lap = a target you can actually hit.

—realistic best

📐 What does Ghost Lap mean

Your actually achievable target lap time.

The idea is simple: you've already driven the fastest S1, fastest S2, fastest S3 in different laps — stitching those three best segments into one lap gives you the Ghost Lap. It's not fantasy; it's capability you've already demonstrated.

The other number, Sum of Corner Gaps, is the physical upper bound "if every corner were driven perfectly" — reference value, not realistic, because corners constrain each other.

Treat Ghost Lap as your target, Sum of Corner Gaps as motivational ceiling.

🛰 Track map · mini-sector

Lap split into 24 equal-distance segments; each colored by how much slower it is than that segment's own fastest across all laps. Purple = segment-best, green→yellow→red = bigger loss. Pick a lap below to recolor; toggle the satellite view for real imagery.

Pick a lap

segment-best

bigger loss →

🏁 S/F● corner⊙ sector◆ top spd

📐 What does Track Map show

Overhead view of the track, with every lap's actual trajectory drawn on a satellite basemap.

Different colors = different laps, easy to compare racing lines.
Orange glowing dots = coast hesitation, marking segments where you neither braked nor accelerated — these are usually places where you're losing lap time for no reason.

After other modules pinpoint a problem, come back to this map to see exactly which corner it happens at.

🔥 Sector × Lap time deltagreen=fastest · redder=slower

Rows = sectors, columns = laps; cell = how much slower than that sector’s own fastest lap. ★ = fastest.

📐 What does this heatmap show

Helps you answer: "A specific lap was slow — slow in which sector?"

X axis = laps, Y axis = sectors. Greener = better performance in that sector (★ = sector's fastest); redder = more behind.

For example, if one lap is 0.4s slow overall and the heatmap shows S2 alone is 0.3s slow, then improvement should focus on corners in S2, not generic "drive faster".

🎯 Corner improvementsorted by savable time

How much faster the apex still can be (speed ceiling). Continuous corners (e.g. T2←T1) improve at the *previous* corner.

Potential is estimated from GPS-fitted corner radius (medium confidence) — focus on *which* corners to practice; absolute seconds are indicative.

L2WhyPer-theme deep dive: data + matched advice.

📈 Speed + Δ time

Where did speed not keep up, and how much time piles up there?

97.1 km/h

Top speed

▼

Two stacked plots: top = speed, bottom = Δtime vs the fastest lap. Shared lap legend — click a lap to toggle both.

Speed (km/h)

Δ time vs fastest (s)

📐 What does this chart show · Reading the Δ Time subplot

Every lap's speed at the same track position is overlaid, so you can see "at a specific spot on the track, exactly where is the gap between the fastest lap and your lap".

X axis is relative track position (lap start to lap end), not time — so laps of different speeds can still be directly overlaid at the same corner apex.

Top — Speed (km/h): each line is a lap's speed curve.
  • Bold green ★ = best lap
  • Normal brightness = the next 3 fastest laps
  • Semi-transparent background = other valid laps (still shows the "spread band" width without stealing focus)
  • Click legend to toggle any lap on / off

Bottom — Δ Time vs Best (s): standard "where am I losing time" view used by MoTeC i2 / FastF1.
  • Y axis = cumulative time this lap is behind best lap at this position (green dashed = 0 = best itself)
  • Slope going up (more positive) = this segment you're slower than best, steeper = losing more
  • Flat / down slope = this segment you match or beat best
  • The steepest ramp = where you're losing time on this lap (usually a specific corner)

Clicking the legend toggles a lap — both subplots show / hide together (shared legendgroup).

Gold S1 / S2 / S3 labels and faint bands at the top map to the three sectors above — so you can directly locate a curve segment to a specific sector and its corners.

🏁 Racing line · best vs worst

Same corner — how do your best and worst lines differ? Apex early or late?

▼

Best (green) vs worst (red) line for each top corner, ★ = apex.

📐 What does this chart show

Picks the "3 corners with highest potential" and overlays your best lap and your worst lap on the same chart, so you can directly see "what line I took when I got it right, and what line when I got it wrong".

Green = best lap's trajectory at this corner, red = worst lap's; ★ = that lap's apex.

Seeing the difference between the two lines (entry point, apex, exit direction) is the most direct way to turn potential into result.

🔄 Grip · friction circle (G-G)

How much grip do brake / accel / cornering actually use?

AiM hardware IMU▼

G-G scatter: X = lateral G (left↔right), Y = longitudinal G (brake↑ accel↓). Outer points = more grip used.

📐 First time? 30-second guide

This chart tells you "how much of your tire's grip you actually used".

Principle (10 sec)
Total tire grip is fixed. You ask it to do 3 things: accelerate (throttle) / decelerate (brake) / turn. Do several at once, each gets less. The usable limit drawn out is a circle — that's the "friction circle". Good driving = always at the circle's edge (using all grip); conservative = clustered near center (wasted).

How to read (axes and forces)
• X axis = cornering G. Right = right corner, Left = left corner.
• Y axis = decel/accel G. Up = braking, Down = accelerating.
• Each dot = an instant on your best lap, color matches which corner (T1-T12, gray = straight).
• Red dashed circle = your session's grip limit.
• Gray thin dashes = 0.5g / 1g / 1.5g reference circles.

What problems can you spot (4 signal types)

Crowded center, sparse edge — grip underused, driving conservatively / not pushing hard.
Empty bottom (sparse lower half) — no trail braking, you "brake then turn", losing 0.1-0.3s/corner (most common issue).
Missing half of left/right semicircle — that direction's corners aren't using full grip (track-biased / you fear that direction).
Too flat vertically (narrow top-bottom) — brake/accel intensity both weak. Karting has smaller longitudinal G than lateral (engine torque limited) which is normal, but near-zero longitudinal spread needs diagnosis: check your hardest brake — under 0.4g for the whole session means car/tire/surface physical limit (drum brake / rental kart common), not your problem; 0.5g+ peak but persistent 0.2g means soft brake style, you can push harder.

How to improve (specific actions)
1. Main gap = bottom (insufficient trail brake): see "weak corners" at top. At these corners you brake-then-turn. Try: in the last 0.5-1s before entry, keep light brake while starting to turn (release brake a beat later than steer-in). This shifts weight to front tires, increases grip, allows higher entry speed.

2. Main gap = accel phase: hesitation between brake-release and full throttle on exit. Try: on exit go straight to full throttle, let car find its own limit.

3. Main gap = one side's corners: in that direction, push harder, steer to limit, see if car understeers (understeer = true limit).

Note: G values inferred from GPS speed + heading (no IMU). After 0.2s smoothing + extreme clipping, karting scenario error ±5%. Only A-type braking corners count for trail brake evaluation; K-type high-speed kinks don't brake and don't count.

🎯 Consistency · Apex

Can you reproduce your best lap's moves every lap?

▼

Unlike L1 (how much faster you could go), this asks: can you replicate your best every lap?

📐 What does this chart show · Reading the boxplot

One boxplot per corner, showing distribution of apex speed across all laps at that corner — the standard visualization used by FastF1 / PITWALL and other F1 lap-time analysis tools.

How to read:

Box (cyan rectangle) = Q1-Q3, middle 50% of lap distribution
Solid white line = median (typical level)
Dashed white line = mean
Top / bottom whiskers = extremes (excluding outliers)
Dots = each lap (hover to see which L#)
Gold ★ = best lap's apex speed at this corner (capability ceiling)

Note: worst lap not marked separately — whiskers already show the extreme range, double-marking would clutter.

How to use:

Narrow box: strong apex-speed repeatability at this corner (steady rhythm)
Wide box / long whiskers: unstable, big lap-to-lap variation → improvement focus
Gold ★ far above median: you've done it once — there's room to bring the slower laps up
Outliers (dots far from box): a specific lap had an incident at this corner (crash / yield / error) — hover to see which lap
Distribution biased high / low: high = most laps near best; low = most laps lagging

📊 Driver profile · radar

What kind of driver are you? Strengths and weak spots?

▼

💪Strength · Consistency97/100This dimension is your top strength.
🎯Improve · Late Braking59/100Brake point can be slightly later — push brake start 5-10m toward apex, straight end will gain a few more km/h.
🎯Improve · Apex Discipline68/100Each lap's apex position can be more consistent — treat brake → turn-in → apex as one fixed sequence, copy best lap's rhythm.

📐 What do the 6 dimensions mean

This session's "driving fingerprint", showing strengths and weaknesses at a glance:

  • Aggressiveness: how hard you brake and how aggressively you accelerate.
  • Smoothness: symmetry / fluidity of corner entry-exit, vs. abrupt stab-and-release.
  • Late Braking: how late you dare brake — closer to apex = more aggressive.
  • Apex Discipline: how consistently each lap hits the same apex (line discipline).
  • Exit Throttle: how quickly you can get back on throttle after apex (exit efficiency).
  • Consistency: whether lap times are stable, or fast-slow vary wildly.

Single-session sample is small and scores have noise — focus on the radar's overall shape and labels, not specific numeric scores.

L3Raw data · trustUnderlying data, methodology disclosure, full coach debrief.

Key Observations

8 valid laps; fastest L6=37.36s, slowest valid L2=38.56s, gap 1.20s.
Lap-time CV = 1.1% (<5% consistent pace, 5-10% improvable, >10% erratic).
L9=42.52s flagged as outlier (likely cooldown / pit-in / pit-out).
Most consistent corner: T8 (apex speed std=0.78 km/h, mean 53.1 km/h).
Biggest opportunity: T3 (apex speed spread 7.3 km/h, best L3=44.2 km/h vs worst L7=36.8 km/h).
Hardest braking: L8 at 0.7s, -0.86G (held 1.00s, 95.3 → 77.7 km/h).
Strongest acceleration: L1 at 15.8s, 0.67G.

Data quality · trustCLEAN

GPS jump rateHigh

8 / 9964 frames despiked · max jump 68×

0.1%

Sample rateAmple

25.0 Hz · enough for apex / brake-point fixing

25.0 Hz

G-force sourceMeasured

Calibrated (gyro + accel)

3-axis

Track extentMeasured

144 × 124 m · small circuit

398 s

Conclusion confidence

ℹ

Start/finish line is auto-detected (not manually calibrated).

🎙

Lao ZhangAI veteran debrief

_The engineer couldn't comment this time (LLM call failed), but all the data charts below work fine._

Share Card