We’ve been working on building a model to simulate and project matches in the PSL this season. Put simply, running simulations allows for a broader view of potential match outcomes. It’s nothing new - betting companies use something comparable in setting up odds, and so does the Win Predictor you often see on TV. We’ve previously done the basics of this with with xG and xPts, but we can be more confident in our hypothesis by replaying some of these underlying numbers over and over. The model is not perfect (but we have to start somewhere), but it does help give a good idea of result and goal probabilities from a match, and throughout the season based on various data points. With time, hopefully we’ll add more variables than xG and xPts.
Some definitions
“Expected Goals calculates the likelihood that a shot will be scored by using information on similar shots in the past. The Opta xG model uses several variables from before, and up to, the exact moment the shot was taken. It evaluates how over 20 variables affect the likelihood of a goal being scored. Some of the most important factors are listed below:
Distance to the goal.
Angle to the goal.
Goalkeeper position, giving us information on the likelihood that they’re able to make a save.
The clarity the shooter has of the goal mouth, based on the positions of other players.
The amount of pressure they are under from the opposition defenders.
Shot type, such as which foot the shooter used or whether it was a volley/header/one-on-one.
Pattern of play (e.g., open play, fast break, direct free-kick, corner kick, throw-in etc.).
Information on the previous action, such as the type of assist (e.g., through ball, cross etc.)”
And more
We’ve taken this xG and related xPts data and replayed the same PSL matches/efforts 10,000 times to realistically see (with slightly more accuracy) how often a team would win a game, and obtain probabilities for the number of goals they would score.
Sundowns v SuperSport
Pre-game, a Sundowns win was the most likely outcome, with the bookmakers giving Sundowns a 70 - 72% chance of victory. Defending champions, playing at home, head-to-head stats, and other information we’re not privy to (possibly xG and more) - a Downs defeat would have been a shock.

In-game, Sundowns managed 10 shots, six of which were on target. 80% of the time they win this game, most likely scoring three goals in the process. 95% of the time Downs score 2 or more goals, which is what happened on the day.
Orlando Pirates v Chippa United
The Pirates game had somewhat similar odds before kickoff. Pirates (75 - 77%) were expected to win by the bookmakers, but Chippa (12% for the win before the game) came very close to grabbing a point (both at 1-1 and with a late chance at 2-1).

Based on how the game played out, Pirates should have won this in a simpler way - they win this 86% of the time (would be lower if we removed the penalty, but probably still comfortable), with at least three goals scored. 13 shots (compared to Chippa’s seven), 39 touches in the box (only 6 for Chippa) and xG of 2.3 compared to 0.6 for Chippa. It’s important to note that there are well documented dangers of looking at single-game xG, but with simulation, we can get better conclusion.
What’s the point?
A week ago we shared the Expected Points table. Here is a summarised version to jog the memory. One standout point was that despite finishing 23 points behind in second, Pirates were 3-4 points behind Sundowns on the Expected Points table.

Last night’s game seems to be a typical case of what happened with Orlando Pirates last season (and other seasons before). A performance with a 70%+ probability of victory only resulted in a draw, and often defeat. This would result in a loss of momentum, and further ground lost to a Sundowns winning machine. In the first 10 games last season, Pirates had dropped points 6 times, sometimes undeservedly.
“Disappointed with the result, obviously I think our performance deserved more than a nil-nil game, like I’m used to saying it’s a scoring game and we didn’t, so that’s why we’re running with only one point,” Jose Riveiro after Pirates 0-0 AmaZulu last season.
92nd minute winners feel good but they won’t (and can’t) be the order of every week, but if Pirates are to challenge, they need to convert their dominance into wins on a more consistent basis.
Fans have suspected that; but the data may have the evidence.
Thanks for reading Laduma Analytics’ Substack! This post is public so feel free to share it.
