TENNISEDGE.GG

R. Jodar vs K. Nishikori — Prediction & Analysis

Washington · · July 29, 2026 · 14:00 UTC

TennisEdge's AI model finds no betting value in R. Jodar vs K. Nishikori (Washington on 2026-07-29) — the market is efficient, with R. Jodar favoured at 1.22.

Surface hard Odds R. Jodar 1.22 · K. Nishikori 4.56

Verdict: no bet

The model found no value at current prices — the market has this one priced about right. That verdict is part of the public record too.
Profiles: R. Jodar · K. Nishikori · head-to-head record · Washington

Full analysis

R. JODAR vs K. NISHIKORI — ATP Washington (hard)
═══════════════════════════════════════════════════════

MODELS vs MARKET (Jodar = A)
  ELO         66.7%   |  ML  72.7%  |  Blend 74.0%  |  Sim 75.7%
  Book        1.22 / 4.56   (implied A 78.9%)
  Betfair ML  1.18 / 5.50   (normalised A 82.3%, B 17.7%)  <- executable
  All four models sit BELOW market on Jodar, i.e. all four call
  Nishikori @5.50 underpriced. Raw blend edge on B = 26.0% - 19.0%
  (comm-adj 5.275) = +7.0%. ELO-only edge on B = +14.3%.

WHY THE CONSENSUS IS ONE CORRELATED ERROR, NOT FOUR VOTES
  resimulate baseline: Nishikori spw 0.607 / rpw 0.418 / hold 77.6%
  vs Jodar spw 0.639 / rpw 0.438 / hold 74.7% — the sim has the
  36-year-old holding BETTER than the man who beat Fils yesterday.
  matches_analyzed = 46 for Nishikori, but he has exactly ONE match
  in the last 90 days and is ranked 238. That 46-match base is
  former-top-4 Nishikori, not 2026 comeback Nishikori. ELO does not
  decay on a layoff (ranking points do — hence 238), so ELO, ML,
  blend and sim all inherit the SAME stale input. Convergence here is
  a shared upstream error, not independent confirmation.

  Bounded re-run (Step 4b): spw_b 0.607 -> 0.570 (-0.037, within
  +/-0.05, evidence = 1 match/90d, age 36, rank 238 vs 46-match base)
  => Nishikori 24.3% -> 11.7%, margin 3.3 -> 4.9 games. Extreme
  input sensitivity. Honest band for Nishikori = 11.7%–24.3%; the
  exchange's 17.7% sits almost exactly mid-band. Coherent number
  taken forward: Nishikori 19% / Jodar 81%.

RECENT FORM (max 5 shown; all fetched matches used for counts)
  Jodar (19, #165) — SHARPENING
    07-28  W  Washington   A. Fils (1954, Elite)          7-6 6-3
    07-03  L  Wimbledon    S. Mochizuki (1757)            grass, 4 sets
    07-02  W  Wimbledon    P. Carreno-Busta (1734)        5 sets
    06-29  W  Wimbledon    F. Gill (1381)                 6-3 6-3 7-5
    06-02  L  French Open  A. Zverev (2035, Elite)        QF, narrow 1st set
    Shape: over-performing his bracket — straight-sets scalp of an
    Elite yesterday, plus an FO quarter-final beating CB (1854),
    Michelsen, Duckworth, Kovacevic. Rank 165 lags his real level.
  Nishikori (36, #238) — INSUFFICIENT (n=1)
    07-27  W  Washington   J. Shang (1723, Strong)        6-7 6-3 6-4
    Single GRIND win over a Strong opponent. Genuine and positive —
    and exactly what the market already priced when it moved him to
    5.50. Nothing here the exchange does not know.

PATH / SHARPNESS
  Both 1 match deep at Washington, both match-sharp on the surface.
  Jodar had a 25-day gap (Wimbledon 07-03 -> 07-28) and Nishikori a
  90d+ layoff; both rust flags are already discharged by a win.

FATIGUE (Step 2 calibration)
  Jodar: 99 min, straight sets, yesterday (07-28). Age 19, ≤22 ->
    recovers fast, softened to roughly equal.
  Nishikori: 154 min, 3 sets, 07-27 — TWO days ago, not yesterday,
    so the measured -4.5pp 3-set-hangover (a next-day effect) does
    NOT apply. Age 36, ≥30 -> hardened one tier: slight edge Jodar.
  Net: slight, and it points the SAME way as the market. Cumulative
  load is context only — nothing to harvest.

NARRATIVE
  Let-down: None (no Final in either window). Venue: Jodar 1-0,
  Nishikori 1-0 at Washington. Title recency: none.
  Surface transition: Jodar off grass, Nishikori off a layoff — both
  already resolved by a completed hard-court win.

ELO vs RANKING SANITY
  Jodar: ELO-favourite at #165 — but the rating was earned vs Elite
  and Strong opposition at Roland Garros and Wimbledon, NOT farmed at
  weak events. A 19-year-old's ranking lags a fast rise. ELO credible,
  arguably understated. No discount.
  Nishikori: #238 with a stat base implying he out-holds Jodar. ELO
  and sim OVERSTATE him — the stale-layoff failure mode. Discount.

GATES
  G1 edge-origin: FAIL — the pre-narrative edge on B exists ONLY off
    stale-input models; at the coherent corrected probability (19%)
    edge = 19.0% - 19.0% ≈ 0, and min_odds = 1.05/0.19 = 5.53 >
    executable 5.50. Automatic PASS under the min-odds rule.
  G2 consensus:   n/a (tier-only, never FAIL) — would cap at 0.5-1.0u
    anyway: contaminated signal (46-match stats base vs 1 match/90d).
  G3 fatal-risk:  n/a — ML market, G3 never blocks ML.

VERDICT — PASS
  A four-model "consensus" on the dog that traces to one shared stale
  input is not model-vs-market disagreement, it is model error. The
  qualitative read runs AGAINST the models, not with them: Jodar is
  SHARPENING off an Elite scalp, Nishikori is 36 with one match in
  90 days and 154 min in his legs. No independent qualitative
  validation exists, so the Step-4 conjunctive bar fails.
  No side prices up either: B needs 5.53 and offers 5.50; A at 1.18
  (effective 1.171, implied 85.4%) is -4.4pp against my 81%.
  Both sides PASS. Nothing to do here.