Pressure intelligence

Where is the pressure?

Pressure = Leverage (how much the next ball can swing win probability) × Difficulty (behind par on wickets and asking rate) × Stage. Compare three win-probability models and both innings.

1.43

Avg Pressure Index · close finishes

0.90

Avg Pressure Index · one-sided games

25.1%

Balls flagged as pressure

▦Heatmap · overs × wickets down

Share of balls flagged as pressure

Overs ↓ / wickets →

Overs 1-4

53%

55%

43%

50%

17%

Overs 5-8

19%

24%

27%

40%

27%

10%

Overs 9-12

10%

14%

21%

38%

37%

21%

10%

Overs 13-16

21%

36%

48%

40%

39%

12%

Overs 17-20

17%

25%

44%

48%

54%

45%

19%

Pressure rises sharply when wickets fall in the first overs of a chase (3 down in overs 1–4: 64%) and when 6–8 wickets are down late (7 down in overs 17–20: 80%), but not when a side is 0 down late. That is the Leverage × Difficulty logic working.

◧Powerplay / middle / death · 3 models

% of balls flagged

Logistic Regression floods the chase powerplay with "pressure" and ignores first-innings middle overs, so XGBoost, which learns over × wicket × runs interactions, defines the main index.

✓Does pressure really make batting harder?

Dismissal % per ball

The Pressure Index also tracks genuinely close matches: 1.43 in finishes decided by ≤10 runs or ≤3 wickets vs 0.90 in one-sided games.