If you are a regular reader of StrataBlog then you will be very familiar with our concept of chances when discussing data collection in football.
In previous pieces we have explained what our definitions of chances are and how we record them, while also demonstrating the various conversion rates of the different categories.
We currently display three of these categories in StrataBet, but actually collect fifteen different levels of chance in-house. The three categories are broken down as follows:
Great Chances are situations that players would be expected to score from.
Good Chances are situations that players could score from but would not necessarily be expected to.
Attempts are situations that players would not typically be expected to score from.
A recent trend our analysts had picked up on was that the first goal scored results from the first chance of note in the contest.
This means that they had seen the first Great Chance or Good Chance in a game typically leading to the first goal. Attempts have an average conversion rate of just 2% and accordingly we do not consider them as noteworthy.
The observation appeared a worthwhile one to investigate from a trading perspective, though it seemed implausible that the first chance in a match would have a higher conversion rate of any that followed:
Using data collected on this season’s Premier League it is clear that the conversion rate of Great Chances/Good Chances is actually lower for the first and second chances in a game than for those that follow. Indeed, conversion rates increase as you progress through the first five chances in a game. This could be explained by the fact that players settle into their stride as a game unfolds, but there are undoubtedly numerous other factors to consider as well.
Unsurprisingly, reducing the scope of investigation to just Great Chances yields similar results:
Although we are dealing with progressively fewer events here, the relatively poor conversion rate of early chances remains and the fifth Great Chance in a match is converted to a goal 9% more than the first on average (compared to 8% more when considering Good Chances as well).
Given that these were relatively disappointing (and indeed potentially blog-stopping) findings, we thought we should take the investigation a little further to see if we could find any valuable information relating to the first chance in a game.
The next step is to see if having the first Great/Good Chance in a match impacts upon a Premier League teams’ probability of winning:
The importance of having the first chance in a game is very clear: home teams who have the first chance go on to win 48.7% of the time and avoid defeat 81.3% of the time.
Furthermore, away teams who have the first chance go on to win 45.1% of the time and avoid defeat 70.8% of the time.
There is an expected drop in the numbers from home to away here, which is repeated when we look at Great Chances on their own:
If a home team has the first Great Chance in a match they go on to win an enormous 54.5% of matches and avoid defeat 80.4% of the time, while an away team goes on to win 48.1% of matches and avoid defeat 76.9% of the time in the same scenario.
Given that the conversion rate of Great Chances alone jumps to an average of 44.8% compared to 27.4% for Great/Good Chances combined, it is no real surprise to see this increase in win probability, as many of these instances will lead to a team taking a 1-0 lead.
Similarly, when there is no Great Chance in a match it is not surprising that a draw becomes much more likely, with 50% of such games ending in stalemate. We must acknowledge here that the sample size is significantly smaller, but the finding is interesting nonetheless.
At the most basic level this gives you something other than goals to look out for when trading football and could explain why the market can spike off the back of chances alone, regardless of their outcome.
When using first chances instead of first goals as an indicator of match outcome, there remains the potential of bias in the findings due to team strength. A natural assumption would be that the better teams have the first Great/Good Chances more often than not and would inherently have a higher chance of winning anyway.
This is the case to some degree, but by no means as clear-cut as you might expect:
Astoundingly tenth-placed Stoke City are comfortable leaders in this table, having the first chance in a game an enormous 74% of the time and typically creating it within the first fifteen-minute segment.
Fourth-placed Arsenal are less surprising runners-up, though Liverpool ranking alongside Manchester City and Crystal Palace matching Leicester City have to rank as minor shocks.
Looking further down it is also a little jarring to see Tottenham Hotspur in mid-table, despite Mauricio Pochettino’s men having now made a case to be considered as genuine title contenders.
Other teams who look to be out of position in a quick eye test are AFC Bournemouth, Sunderland, West Ham United and Chelsea.
Things become a little more as we might expect when we narrow the focus down to Great Chances alone, with title favourites Manchester City leading the way:
There are still some surprises here too, though: Stoke remain impressively high and Watford jump nine places to match Tottenham and Southampton in creating the first Great Chance in a game 57% of the time.
At the opposite end of the table the quartet of teams who look significantly out of place are Manchester United, West Ham United, Crystal Palace and Everton.
These final two tables can give some good cues for trading in the first team to score, time of first goal and first goalscorer markets. In the context of this piece, though, they primarily indicate that the first chance in a football game remains significant to the outcome, regardless of the teams involved.
In the near future we will dive deeper into the importance of first chances for trading, taking a close look at the players who tend to have them and exploring their individual conversion rates.