Scouting Summary
Madison Keys's current window is too thin or skewed for a stable public style read. Only 9 charted matches sit in this window, so the read is still directional. Use the metric cards as directional hints, not a settled profile.
Public WTA coverage-qualified profile built from charted matches. Start with the summary, then follow one reliable trait into the wider tour.
Scouting Summary
Madison Keys's current window is too thin or skewed for a stable public style read. Only 9 charted matches sit in this window, so the read is still directional. Use the metric cards as directional hints, not a settled profile.
The top line stays simple. Expand any card when you want the raw sample, uncertainty, baseline, rank, and nearby names in the leaderboard.
Profile Shape
Missing sample on: Short service points.
Read the page like a scout
Deep return rate
36.5%
Higher values usually mean the return is buying time, pinning the server back, and moving the point off easy first-ball offense.
Adjusted value: 36.5%
Raw value: 39.3%
Baseline: 33.4%
Vs baseline: +3.1 pts
Confidence: Medium confidence
90% interval: 32.6-40.4
Effective sample size: 409.3
Sample: 412 returnable returns · weighted 467
Threshold: 300+ returnable returns
Forehand attack rate
34.5%
Higher values point to a forehand that steps on the point instead of simply recycling crosscourt.
Adjusted value: 34.5%
Raw value: 37.9%
Baseline: 31.6%
Vs baseline: +2.9 pts
Confidence: Medium confidence
90% interval: 31.2-37.8
Effective sample size: 555.3
Sample: 559 charted forehand directions · weighted 633
Threshold: 500+ charted forehand directions
Serve variety
99 / 100
Higher values usually mean the first-ball read is harder because serve patterns are not concentrated in one lane.
Adjusted value: 99 / 100
Raw value: 97 / 100
Baseline: 99 / 100
Vs baseline: -0.9 pts
Confidence: High confidence
90% interval: 97.9-99.1
Effective sample size: 565.4
Sample: 569 charted serves · weighted 645
Threshold: 300+ charted serves
Net approach rate
8.5%
Higher values point to a profile that uses the front court as a recurring pressure tool rather than an emergency option.
Adjusted value: 8.5%
Raw value: 9.1%
Baseline: 7.8%
Vs baseline: +0.8 pts
Confidence: High confidence
90% interval: 7.2-9.9
Effective sample size: 1158.7
Sample: 106 net points across 1166 total points · weighted 1320
Threshold: 30+ net points and 600+ total points
Net conversion
72.6%
Higher values point to cleaner front-court execution once the approach or transition decision is made.
Adjusted value: 72.6%
Raw value: 74.5%
Baseline: 68.6%
Vs baseline: +4.0 pts
Confidence: Low confidence
90% interval: 65.5-79.8
Effective sample size: 105.4
Sample: 106 net points · weighted 119
Threshold: 30+ net points
Nearby names
No nearby leaderboard context is available here.
Break-point resistance
52.8%
Higher values point to a service profile that stays stable when the score tightens.
Adjusted value: 52.8%
Raw value: 50.0%
Baseline: 57.0%
Vs baseline: -4.2 pts
Confidence: Low confidence
90% interval: 41.8-63.8
Effective sample size: 55.6
Sample: 56 break points faced on serve · weighted 65
Threshold: 20+ break points faced on serve
Nearby names
No nearby leaderboard context is available here.
Break-point attack
44.4%
Higher values point to a return profile that leans forward at the moment the service game can swing.
Adjusted value: 44.4%
Raw value: 42.9%
Baseline: 46.8%
Vs baseline: -2.3 pts
Confidence: Low confidence
90% interval: 34.6-54.2
Effective sample size: 69.5
Sample: 70 break points on return · weighted 79
Threshold: 20+ break points on return
Nearby names
No nearby leaderboard context is available here.
Short service points
Sample too thin
Needs 300+ charted service points.
Higher values point to service games that stay short and front-footed. Lower values usually mean a profile that lives in longer points.
Adjusted value: Unavailable
Confidence: Confidence unavailable
Sample: 117 charted service points · weighted 121
Threshold: 300+ charted service points
Nearby names
No nearby leaderboard context is available here.
Surface cards only show up when the selected season window has at least 10 charted matches on that surface.
No surface split clears the threshold here.
This season view does not have 10 charted matches on a single surface for Madison Keys.
Methodology
The methodology page explains thresholds, weighting, uncertainty, reliability grades, and why some players or surfaces disappear when the sample gets too thin.
Private workflow
Tour Radar is the public proof layer. Private work starts when your team brings a live coaching question and sends team footage, tracking exports, or official reports or exports it can share.