Sample report

A real Regulation M-B ladder game, analyzed at production settings. Player names removed. Notes on routine turns are trimmed to their first sentence; every negative verdict keeps its full reasoning.

← Back to Charmeleon · Join the beta to analyze your own games

85 / 100
side p1
best4 ok5 inaccuracy1 mistake2 blunder0
Precision
75%
EV-loss
4.5%
Turns analyzed
12/17
Time
1653.6s

Win probability

0%25%50%75%100%opening: indicative onlyT1T2T3T4T5T6T7T8T9T10T11T12Go to turn 1Go to turn 2Go to turn 3Go to turn 4Go to turn 5Go to turn 6Go to turn 7Go to turn 8Go to turn 9Go to turn 10Go to turn 11Go to turn 12T1: okT2: okT3: bestT4: bestT5: mistakeT6: mistakeT7: bestT8: okT9: bestT10: inaccuracyT11: ok

Squares under the curve are per-turn verdicts — click one to jump to that turn.

Turns 1–3 are shaded because the win% there is indicative only: measured over 205 real games, the model's opening win% does not rank outcomes better than chance (AUC 0.51; from turn 4 on, 0.69–0.77). Comparing the alternatives within a single turn is still informative.

Key moments

Turns

★ Team preview: not analyzed — enable it on upload (+3-7 min)
T1 ok -5.5%
You played
[Aerodactyl]: Tailwind
[Garchomp]: Stomping Tantrum
CI 90% [0.0%, 16.6%]
Engine best
[Aerodactyl]: rockslide
[Garchomp]: dragonclaw
Equilibrium mix
[Aerodactyl]: rockslide
[Garchomp]: dragonclaw
100%
Charizard: heatwave in 3/8 worldsCharizard: solarbeam in 3/8 worldsGrimmsnarl: PartingShot in 8/8 worldsGrimmsnarl: LightScreen in 6/8 worlds

Reasonable play.

T2 ok -1.0%
You played
[Garchomp]: -> Sylveon
[Aerodactyl-Mega]: Rock Slide
CI 90% [0.0%, 3.1%]
Engine best
[Aerodactyl-Mega]: protect
[Garchomp]: rockslide
Equilibrium mix
[Aerodactyl-Mega]: protect
[Garchomp]: rockslide
98%
[Aerodactyl-Mega]: rockslide
[Garchomp]: -> Sylveon
2%
[Aerodactyl-Mega]: tailwind
[Garchomp]: stompingtantrum
0%
Grimmsnarl: LightScreen in 8/8 worldsGrimmsnarl: PartingShot in 8/8 worldsArchaludon: DragonPulse in 8/8 worldsArchaludon: ElectroShot in 8/8 worlds

Your play was inside the Nash mix (1.7%).

T3 best ~0
You played
[Aerodactyl-Mega]: Rock Slide
[Sylveon]: Hyper Voice
CI 90% [0.0%, 0.4%]
Engine best
[Aerodactyl-Mega]: rockslide
[Sylveon]: hypervoice
Equilibrium mix
[Aerodactyl-Mega]: rockslide
[Sylveon]: hypervoice
92%
[Aerodactyl-Mega]: -> Garchomp
[Sylveon]: -> Kingambit
5%
[Aerodactyl-Mega]: protect
[Sylveon]: quickattack
3%
Grimmsnarl: LightScreen in 8/8 worldsGrimmsnarl: PartingShot in 8/8 worldsArchaludon: DragonPulse in 8/8 worldsArchaludon: ElectroShot in 8/8 worlds

The best possible action.

T4 best ~0
You played
[Aerodactyl-Mega]: Rock Slide
[Sylveon]: Hyper Voice
CI 90% [0.0%, 0.7%]
Engine best
[Aerodactyl-Mega]: rockslide
[Sylveon]: hypervoice
Equilibrium mix
[Aerodactyl-Mega]: rockslide
[Sylveon]: hypervoice
86%
[Aerodactyl-Mega]: -> Garchomp
[Sylveon]: -> Kingambit
13%
[Aerodactyl-Mega]: dualwingbeat
[Sylveon]: detect
1%
Grimmsnarl: PartingShot in 8/8 worldsArchaludon: DragonPulse in 8/8 worldsArchaludon: ElectroShot in 8/8 worlds

The best possible action.

T5 mistake -22.9%
You played
[Aerodactyl-Mega]: -> Garchomp
[Sylveon]: Yawn
CI 90% [12.8%, 30.7%]
Engine best
[Aerodactyl-Mega]: protect
[Sylveon]: -> Garchomp
Equilibrium mix
[Aerodactyl-Mega]: protect
[Sylveon]: -> Garchomp
69%
[Aerodactyl-Mega]: tailwind
[Sylveon]: hypervoice
28%
[Aerodactyl-Mega]: rockslide
[Sylveon]: detect
3%
Archaludon: DragonPulse in 8/8 worldsArchaludon: ElectroShot in 8/8 worldsPelipper: Hurricane in 8/8 worldsPelipper: Tailwind in 8/8 worlds

Significant error. You lost 22.9% of win% (90% CI: 12.8%-30.7%). Label softened because of belief noise. You should have attacked with protect / ->Garchomp. [Uncertain verdict: the analysis invented part of the opponent's set (no OTS, no pool); treat it as a hypothesis, not a fact.] [The model assumes: Archaludon: DragonPulse in 8/8 worlds; Archaludon: ElectroShot in 8/8 worlds; Pelipper: Hurricane in 8/8 worlds; Pelipper: Tailwind in 8/8 worlds.] Calibrated real cost: ~10pp win probability.

T6 ok ~0
You played
[Sylveon]: Hyper Voice
CI 90% [0.0%, 6.6%]
Engine best
[Garchomp]: protect
[Sylveon]: quickattack
Equilibrium mix
[Garchomp]: protect
[Sylveon]: quickattack
48%
[Garchomp]: -> Aerodactyl-Mega
[Sylveon]: -> Kingambit
38%
[Garchomp]: stompingtantrum
[Sylveon]: hypervoice
12%
[Garchomp]: rockslide
[Sylveon]: detect
2%
Archaludon: DragonPulse in 8/8 worldsArchaludon: FlashCannon in 8/8 worldsPelipper: Hurricane in 8/8 worldsPelipper: WeatherBall in 7/8 worlds

Your play was inside the Nash mix (11.5%).

T6 mistake -13.3%
You played
[?]: -> Kingambit
CI 90% [9.6%, 17.4%]
Engine best
[Garchomp]: -> Aerodactyl-Mega
[Sylveon]: pass
Equilibrium mix
[Garchomp]: -> Aerodactyl-Mega
[Sylveon]: pass
100%
[Garchomp]: -> Kingambit
[Sylveon]: pass
0%
Archaludon: FlashCannon in 8/8 worldsPelipper: Hurricane in 8/8 worldsPelipper: WideGuard in 8/8 worlds

Significant error. You lost 13.3% of win% (90% CI: 9.6%-17.4%). The Nash mix favours ->Aerodactyl-Mega / pass (99.9%). [Uncertain verdict: the analysis invented part of the opponent's set (no OTS, no pool); treat it as a hypothesis, not a fact.] [The model assumes: Archaludon: FlashCannon in 8/8 worlds; Pelipper: Hurricane in 8/8 worlds; Pelipper: WideGuard in 8/8 worlds.] Calibrated real cost: ~3pp win probability.

T7 best ~0
You played
[Sylveon]: -> Aerodactyl-Mega
[Kingambit]: Sucker Punch
CI 90% [0.0%, 1.8%]
Engine best
[Kingambit]: suckerpunch
[Sylveon]: -> Aerodactyl-Mega
Equilibrium mix
[Kingambit]: suckerpunch
[Sylveon]: -> Aerodactyl-Mega
73%
[Kingambit]: -> Aerodactyl-Mega
[Sylveon]: quickattack
24%
[Kingambit]: kowtowcleave
[Sylveon]: hypervoice
3%
Archaludon: FlashCannon in 8/8 worldsPelipper: Hurricane in 8/8 worldsPelipper: WideGuard in 6/8 worlds

The best possible action.

T7 skipped
You played
-

Opponent replacement — not your decision, not scored.

T8 ok -1.4%
You played
[Aerodactyl-Mega]: Tailwind
[Kingambit]: Low Kick
CI 90% [0.0%, 4.6%]
Engine best
[Kingambit]: suckerpunch
[Aerodactyl-Mega]: rockslide
Equilibrium mix
[Kingambit]: suckerpunch
[Aerodactyl-Mega]: rockslide
99%
[Kingambit]: suckerpunch
[Aerodactyl-Mega]: -> Sylveon
1%
[Kingambit]: lowkick
[Aerodactyl-Mega]: tailwind
0%
Grimmsnarl: PartingShot in 8/8 worldsArchaludon: FlashCannon in 8/8 worlds

Your play was inside the Nash mix (0.3%).

T9 best ~0
You played
[Aerodactyl-Mega]: Rock Slide
[Kingambit]: Low Kick
CI 90% [0.0%, 0.0%]
Engine best
[Kingambit]: lowkick
[Aerodactyl-Mega]: rockslide
Equilibrium mix
[Kingambit]: lowkick
[Aerodactyl-Mega]: rockslide
100%
Archaludon: FlashCannon in 8/8 worlds

You played the best possible action.

T9 recommendation
You played
->Sylveon

Replacement: only one switch available (no real choice).

T10 inaccuracy -9.6%
You played
[Sylveon]: Detect
[Kingambit]: Low Kick
CI 90% [4.3%, 16.4%]
Engine best
[Kingambit]: suckerpunch
[Sylveon]: hypervoice
Equilibrium mix
[Kingambit]: suckerpunch
[Sylveon]: hypervoice
100%

Minor error. You lost 9.6% of win% (90% CI: 4.4%-16.4%). Label softened because of belief noise. The Nash mix favours suckerpunch / hypervoice (100.0%). [Uncertain verdict: the analysis invented part of the opponent's set (no OTS, no pool); treat it as a hypothesis, not a fact.] Calibrated real cost: ~10pp win probability.

T11 ok ~0
You played
[Sylveon]: Hyper Voice
[Kingambit]: Low Kick
CI 90% [0.0%, 0.0%]
Engine best
[Kingambit]: suckerpunch
[Sylveon]: hypervoice
Equilibrium mix
[Kingambit]: suckerpunch
[Sylveon]: hypervoice
50%
[Kingambit]: lowkick
[Sylveon]: hypervoice
50%

Your play was inside the Nash mix (49.9%).

T11 skipped
You played
-

Opponent replacement — not your decision, not scored.

T12 skipped
no_opponent_otsown_bring_inferred
NOTE: ONE-SIDED (PESSIMISTIC) ANALYSIS. Win% values are upper bounds because the opponent is over-informed. The real score may be ~5-8% better. Labels are indicative, not absolute.