Australia were clear favorites going into the ICC World Test Championship Final, but what were the actual chances of the Proteas walking away with the mace?
I’m Johannes, 18, and I’m working toward a career in Quantitative Finance. To stay accountable (and share some cool insights), I’m documenting my learning journey every day.
Today is Day 1, and I decided to jump into the core of Quantitative Finance: probability and statistics.
The most straightforward way to determine the probability is through a concept called Conditional Probability. It sounds intense, but it’s actually a useful and fairly quick way to figure out the likelihood of an event, given that several conditions or factors are in play. Formally, that’s written as:
P(Proteas Win) = P(SA Win | Factor 1, Factor 2, …, Factor n)
To start, I used the overall probability of the Proteas beating Australia in Test matches, without digging too deeply into recent form. This serves as a good baseline. The number I found was 26.47%, the exact percentage of matches South Africa have won against Australia in Test format. I would’ve usually assumed 25%, but when it comes to exact probabilities, assumptions don’t take you very far. I’ll show why in a moment.
Next, I looked at South Africa’s recent form in Test cricket. The most logical percentage to use here is their win rate during the World Test Championship: 69.44%. But that includes draws, which adds more complexity than I’m ready to deal with yet. So I used their record without the draw: 8 wins from 11 games, giving a cleaner 72.7%.
Another important factor is how South Africa have performed against Australia recently, regardless of format. I took the last three SA vs AUS matches from each format (Tests, ODIs, and T20Is): Draw, Australia, Australia, Australia, South Africa, South Africa, Australia, Australia, Australia.
That’s 2 wins in 9 games, giving a 22.2% win rate, which starts to show just how unlikely this win actually was.
Now the venue: Lord’s is neutral ground for both teams, but the historical win rates are telling. Australia’s overall record at Lord’s since 1896 is 15 wins, 3 losses, and 14 draws. South Africa’s record is 6 wins, 8 losses, and 4 draws. Clearly, conditions lean in Australia’s favor.
Another detail to consider is the toss. South Africa won the toss and chose to bowl first, but Lord’s is historically a pitch to bat first on. The win rate when bowling second is just 44.4%.
This gives us a decent amount of data to estimate a reasonably accurate probability. The next step is to assign weights to each factor, since some have more impact than others. I gave both recent form and recent head-to-head 30% weight each. The venue got 20%, and the historical SA record plus overall head-to-head each got 10%.
Here’s the final calculation:
P(Proteas Win) = (w_recent_form × P_recent_form) + (w_recent_h2h × P_recent_h2h) + (w_venue × P_venue_SA) + (w_overall_h2h × P_overall_h2h)
= (0.30 × 0.727) + (0.30 × 0.222) + (0.20 × 0.4286) + (0.10 × 0.2647)
= 0.2181 + 0.0666 + 0.08572 + 0.02647
= 0.39689
So, according to my basic probabilistic assumptions, the chance of South Africa winning the WTC Final was roughly 39.7%.
