01
Final Standings
#
Engine
Score / 700
Score %
Seed Elo
S-B
1
Stockfish 18
x64 (AVX2)
440.0
62.9%
2980
146.553
2
Stockfish 17.1
x64 (AVX2)
412.0
58.9%
2975
137.909
3
Reckless 0.9.0
x64 (AVX2)
386.0
55.1%
2970
129.648
4
Reckless 0.8.0
x64 (AVX2)
340.5
48.6%
2965
116.720
5
PlentyChess 7.0.0
x64 (BMI2)
338.5
48.4%
2963
115.857
6
Obsidian 16.0
x64 (BMI2)
320.5
45.8%
2950
109.692
7
Alexandria 9.0.0
x64 (BMI2)
315.5
45.1%
2945
108.763
8
Komodo Dragon 3.3
x64 (AVX2)
247.0
35.3%
2930
88.723
02
Cross-Result Table
Each cell shows Row Engine score vs Column Engine (100 games). Cells marked *-* = same-family (counted as ½-½, not played). · = self.
| Engine | SF 18 | SF 17.1 | RK 9.0 | RK 8.0 | PC 7.0 | OB 16 | AL 9.0 | KD 3.3 | Total |
|---|---|---|---|---|---|---|---|---|---|
| Stockfish 18 | · | *-* | 60.5–39.5 | 63.5–36.5 | 62.5–37.5 | 69.0–31.0 | 65.5–34.5 | 69.0–31.0 | 440.0 |
| Stockfish 17.1 | *-* | · | 49.5–50.5 | 57.0–43.0 | 56.0–44.0 | 65.0–35.0 | 64.0–36.0 | 70.5–29.5 | 412.0 |
| Reckless 0.9.0 | 39.5–60.5 | 50.5–49.5 | · | *-* | 59.5–40.5 | 57.0–43.0 | 59.0–41.0 | 70.5–29.5 | 386.0 |
| Reckless 0.8.0 | 36.5–63.5 | 43.0–57.0 | *-* | · | 48.5–51.5 | 48.0–52.0 | 52.0–48.0 | 62.5–37.5 | 340.5 |
| PlentyChess 7.0.0 | 37.5–62.5 | 44.0–56.0 | 40.5–59.5 | 51.5–48.5 | · | 50.0–50.0 | 53.0–47.0 | 62.0–38.0 | 338.5 |
| Obsidian 16.0 | 31.0–69.0 | 35.0–65.0 | 43.0–57.0 | 52.0–48.0 | 50.0–50.0 | · | 49.0–51.0 | 60.5–39.5 | 320.5 |
| Alexandria 9.0.0 | 34.5–65.5 | 36.0–64.0 | 41.0–59.0 | 48.0–52.0 | 47.0–53.0 | 51.0–49.0 | · | 58.0–42.0 | 315.5 |
| Komodo Dragon 3.3 | 31.0–69.0 | 29.5–70.5 | 29.5–70.5 | 37.5–62.5 | 38.0–62.0 | 39.5–60.5 | 42.0–58.0 | · | 247.0 |
self / diagonal
same family (*-*)
total score
green text = winning score
red text = losing score
03
Head-to-Head Results
100 games per matchup. Bar shows proportion: green=wins · blue=draws · red=losses (white-engine perspective).
| White Engine | Black Engine | W | D | L | Score % | Result bar |
|---|
04
Deep Analysis
4.1 Dominance Tier
Champion
440.0 / 700
Stockfish 18 — Undisputed Champion
28 points clear of SF17.1. Most emphatic result: 95.24% vs Obsidian (40W–2L–58D) — the highest score in the tournament. Scored above 74% against every non-Stockfish engine. Its 63 draws vs PlentyChess (most in the field) reflects elite solidity at blitz.
Runner-up
412.0 / 700
Stockfish 17.1 — Flawless vs Lower Half
91.84% vs Komodo Dragon (45W–4L–51D) — tied for the tournament's best per-matchup score. Its Achilles heel: Reckless 0.9.0 held it to 48.84% — the only cross-family matchup where a lower-seeded engine beat a Stockfish version.
4.2 The Reckless Phenomenon
Upset Machine
51.16% vs SF17.1
Reckless 0.9.0 — The Tournament's Biggest Story
Rank 3 with 386.0/700. Achieved virtual parity against Stockfish 17.1 — effectively beating a 2975 Elo engine while seeded at 2970. Tied for best matchup score (91.84% vs Komodo Dragon). Its tactical aggression creates structural problems for defensive Stockfish tendencies.
Underperformer
340.5 / 700
Reckless 0.8.0 — Exposed by Its Successor
45.5 points below Reckless 0.9.0 — the largest intra-family gap. Scored only 46.81% vs PlentyChess and 45.83% vs Obsidian, engines it should control. The 0.8→0.9 update represents a transformative leap in engine strength.
4.3 Mid-Field Battle
5th Place
338.5 / 700
PlentyChess 7.0.0 — Solid but Style-Limited
Exactly met its seed expectation. Strong vs Komodo Dragon (78.57%) but suffered a perfect 50.0–50.0 deadlock vs Obsidian — a genuine stylistic stalemate. Lost to both Reckless versions and both Stockfish versions by consistent margins.
6th Place
320.5 / 700
Obsidian 16.0 — The Anomaly Engine
Perfectly split 50.0–50.0 with PlentyChess yet was catastrophically outplayed by SF18 (4.76% — lowest score in the tournament). Extreme variance: competitive in the middle tier, helpless against the top. Highly style-sensitive.
7th Place
315.5 / 700
Alexandria 9.0.0 — Below Expectations
Beaten by every engine above it. Scored below 50% vs PlentyChess (42.86%). The narrow Obsidian result (47.92% vs 52.08%) confirms both sit in a tightly contested mid-tier. Its 65.38% vs Komodo Dragon was its clearest dominance display.
4.4 The Fallen Giant
Last Place · Systemic Crisis
247.0 / 700 · 35.3%
Komodo Dragon 3.3 — Critically Outclassed
68.5 points below 7th-place Alexandria. Conceded 45 losses to both SF17.1 and Reckless 0.9.0 in 100 games each. Scored above 30% only against Alexandria (34.62%). Once a world top-3 engine, Dragon 3.3 now shows the full cost of falling behind the NNUE revolution. Total white wins across all 600 games: only 61.
05
Performance vs Seed Elo
Performance Elo estimated via: Perf = AvgOpp + 400·log₁₀(S/(1–S)). Green = outperformed seed · Red = underperformed.
| Engine | Arch. | Seed Elo | Score % | Perf. Elo | +/- vs Seed |
|---|
06
Tournament Statistics
| Metric | Value | Context |
|---|---|---|
| Total games played | 2,600 | 100% completion rate — all rounds finished |
| White wins | 1,118 (43.00%) | Blitz time control slightly favours the first mover |
| Black wins | 61 (2.35%) | Extremely low — elite engines rarely lose as Black to weaker foes |
| Draws | 1,421 (54.65%) | High draw rate typical of elite-level blitz between near-matched engines |
| Best single matchup | 95.24% — SF18 vs Obsidian | 40W–2L–58D across 100 games; far above Elo-expected ~67% |
| Lowest single score | 4.76% — Obsidian vs SF18 | Mirror of above; 2 wins in 100 games — worst result in the tournament |
| Most decisive series | SF17.1 vs Komodo Dragon | 45W–4L–51D → 91.84%; tied with Reckless 0.9.0 vs Komodo Dragon |
| Most balanced (cross-family) | SF17.1 vs Reckless 0.9.0 | 49.5–50.5 → virtual parity; the biggest upset in the tournament |
| Biggest upset | Reckless 0.9.0 > 50% vs SF17.1 | Lower-seeded engine achieving parity against a Stockfish version |
| Most draws in a series | SF18 vs PlentyChess — 63 draws | Ultra-solid play; fewest decisive games in any 100-game match |
| Fewest draws in a series | Alexandria vs Komodo Dragon — 48 | Most decisive cross-tier match; 34W–18L–48D |
| Largest points gap | 68.5 pts (Alex 9.0 → Komodo 3.3) | A chasm between 7th and 8th place |
| Tightest gap | 2.0 pts (Reckless 0.8.0 → PlentyChess) | 340.5 vs 338.5 — separated by a single game's worth of points |
07
Conclusions
1
Stockfish 18 is the unchallenged champion. Its 28-point margin over SF17.1 and near-perfect record against mid-tier engines — including the extraordinary 95.24% vs Obsidian — demonstrate it operates in a class of its own under blitz conditions.
2
Reckless 0.9.0 is the tournament's defining story. Finishing 3rd and achieving parity against Stockfish 17.1 marks it as a genuine elite contender. Its tactical sharpness makes it uniquely dangerous at blitz time controls, and the version gap over 0.8.0 is the most dramatic engine improvement in this field.
3
Komodo Dragon 3.3 is critically outclassed. A last-place finish 68.5 points below 7th-place Alexandria, with only 61 total wins across 600 played games, confirms the engine is no longer competitive at the elite level. The NNUE revolution has left it behind.
4
The 54.65% draw rate reflects elite-level blitz precision. Both sides neutralise each other effectively across the board. Only the Komodo matchups push decisive results above 50%, confirming that every other engine pair is closely enough matched to frequently enter draw territory.
5
Mid-field is extraordinarily tight. PlentyChess, Obsidian and Alexandria are separated by just 23 points — less than one percentage point of total possible score. Small stylistic advantages, not raw strength, determine standings in this bracket. The 50.0–50.0 PlentyChess vs Obsidian deadlock is the clearest evidence of this.