Form score of 69.82 is only 1.66 points below the Tactico Score of 71.48, indicating a broadly stable trajectory with no meaningful decline. Score confidence is high at 0.92, and the sample is backed by 2553 domestic minutes in Ligue 1, suggesting a reasonably deep sample rather than a small-window read. Sub-scores show real spread (finishing 19.19 vs creation 47.10 vs progression 31.01 vs defense 39.09 vs physical duel 45.48), but this variance is expected given the defensive midfielder role profile rather than pure volatility — finishing is naturally low for this position. Data completeness sits at 83.3%, with missing fields limited to positionLabel and nationalityName — non-form-critical metadata rather than performance data, so no injury/availability signal here. Competition context shows a single non-continental competition (Ligue 1, CDI 0.93) with no European fixtures, so there is no added fatigue/congestion risk to flag, and no continental-driven dip to reinterpret.

Evidence is entirely from Ligue 1 (cdi 0.93, blendedDifficulty 0.93), a moderately-below-average difficulty league with no continental minutes recorded — no European testing evidence to justify an upward adjustment. Sample is substantial: 2553 minutes in a single competition suggests a regular starter with low rotation risk, supporting a reasonably confident read. Role clarity is decent but not fully confirmed: defensive_midfielder label is only partially corroborated by physical signals — clearances (2.64/90) and duels won (5.04/90, 57% success) fit a defensive midfield profile, but defenseScore (39.09) is modest relative to creationScore (47.1), and bigChancesCreatedPer90 (0.28) is low, so the profile leans slightly more balanced/hybrid than a pure defensive anchor. Data completeness is 83% with missing positionLabel and nationalityName — minor gaps that don't materially affect environmental read but limit full role verification. No club-level goal/xG diff context provided (player-level packet). Single-competition data means no cross-context comparison is possible, capping confidence.