Agent Lab — a live forecast exhibit · Applied AI
← Schedule
Group A · Matchday 8 · 2026-06-18

Czech Republic vs South Africa

Atlanta, United States · neutral venue

Full time CZE 1–1 RSA
The agent called Czech Republic (54%) — Worse than the model call wrong

The forecast below is what the agent published before kickoff.

Most likely scoreline
0–0 15.2%
42.9%
CZE win
29.5%
Draw
27.6%
RSA win
Top scorelines
  • 0–0 15.2%
  • 1–0 14.8%
  • 0–1 11.3%
  • 1–1 10.9%
Total goals
0 to 10 goals · most likely → tallest

The agent’s call

3-agent ensemble · deepseek-chat
CZE win
54.0%
Draw
27.7%
RSA win
18.3%

The model's prior (46% home) is reasonable but likely understates Czech Republic's advantage given the squad value gap (~3.2x), superior key players (Schick, Soucek, Krejcí in top leagues vs South Africa's domestic-based core), and South Africa's under-performing recent form (-0.94 pts/match) plus the suspension of key midfielder Themba Zwane. I raise home win probability to 55%, lowering draw and away win accordingly. The climate-controlled roof neutralizes conditions, so no adjustment there. Czech Republic's expected goals of 1.28 and South Africa's 0.84 from the model form a baseline, but the squad and availability gaps suggest Czech Republic can generate more chances. A 1-0 or 2-0 win is most likely, with 1-1 draw as the main draw outcome. South Africa's attack is weak (avg 0.84 xG) and they lack Zwane, so clean sheet for Czech Republic is plausible.

What the agent weighed
What the agent chose to investigate
The agent decided what to check on tavily — 9 sources kept:
  • “Czech Republic national team injuries suspensions June 2026” CZE · 3 hits
    Key player absences could shift the balance against South Africa.
  • “South Africa national team injuries suspensions June 2026” RSA · 3 hits
    Missing key players would weaken South Africa's chances.
  • “Czech Republic vs South Africa neutral venue Atlanta weather forecast June 18 2026” match · 3 hits
    Extreme heat or rain could affect playing conditions and performance.
Agent scorelines
generated 2026-06-18
Last 5 — CZE
-0.12 pts/match vs model expectation · opp. avg #52
  • REP #49 -0.7
  • DEN #14 -0.1
  • KOS #56 +1.2
  • GUA #100 +0.6
  • KOR #40 -1.3
GF / GA last 10: 19 / 11
Last 5 — RSA
-0.96 pts/match vs model expectation · opp. avg #81
  • PAN #72 -0.7
  • PAN #72 -1.7
  • NIC #151 -1.5
  • JAM #91 -0.6
  • MEX #20 -0.7
GF / GA last 10: 12 / 13
Head-to-head (last 5)
  • 1997-12-13 South Africa 2–2 Czech Republic
▸ Full score grid (7×7)
P(home goals × away goals), %
0
1
2
3
4
5
6
0
15.2
11.3
5.0
1.7
0.5
0.1
0.0
1
14.8
10.9
4.9
1.7
0.5
0.1
0.0
2
8.6
6.4
2.8
1.0
0.3
0.1
0.0
3
3.9
2.9
1.3
0.4
0.1
0.0
0.0
4
1.5
1.1
0.5
0.2
0.1
0.0
0.0
5
0.5
0.4
0.2
0.1
0.0
0.0
0.0
6
0.2
0.1
0.1
0.0
0.0
0.0
0.0

Rows = CZE goals; columns = RSA goals. Above-diagonal cells are home wins; diagonal = draws; below = away wins. Truncated at 6 — the joint mass past that is < 1 %.

Expected goals (Poisson rate, λ)
1.21
CZE
0.87
RSA

Independent-Poisson rates. W/D/L and the score grid integrate the full 13×13 joint under these two parameters.

News context — injuries, suspensions, projected XI — is not part of the model, which sees only past results.