Expected goals are a bedrock of an analyst’s toolkit. However, it is simply not possible to look at totals and infer sufficient information about a game to understand what really occurred. Extra layers are required to round out the story. One additional tool that can help analysis is the Race Chart.

These charts enable us to examine the shape of the game not just with regard to the volume and value of chances but also when they took place. We plot every chance in the game in relation to the time in which it took place (x axis) and the cumulative expected goals (xG, y-axis). Take for example this match from May 2017 between Philadelphia Union and New York Red Bulls:

On the surface a straightforward 3-0 win for Union, but immediately expected goals tells a different story, with the Red Bulls creating shots to a value of 2.43 xG and Union 2.14. The Race Chart tells us plenty more than that. Prior to their first goal, Union created only few low value opportunities, and did not create a single shot between the 22nd and 70th minute. During the time the game was 0-0, the Red Bulls logged chances worth around 1.5 xG, with one huge chance in the twentieth minute and further medium sized chances during the second half.

In chasing the deficit, the Red Bulls continued to create shots, until they conceded a second and eventually a third, each of which were high quality chances for Union. The percentages towards the upper left of the chart are powered by simulations of the chances and show how frequently each of the three possible results would be expected to occur. In this instance, we can see that Red Bulls can

consider themselves unfortunate not to have taken anything from the game. Given the balance of chances, they had a probability of around 75% to secure at least a point.

Quick takes:

• Red Bulls were highly competitive for 75 minutes, especially in defence
• The scoreline ultimately flattered Philadelphia
• Red Bulls opened up too much when chasing at 0-1

Here’s another match, again from May, with DC United facing Montreal Impact:

Here we have a situation in which an early goal dictated the entire shape of the game. A combined xG total of around 1.5 is extremely low and it looks likely that after scoring, Montreal attempted to shut down the whole game in two ways. Firstly, in defending, they allowed DC only low value shooting opportunities. Indeed the highest value shot was in the 76th minute, and registered a probability of scoring of around just 13%. Secondly, Montreal had just two shots in the second half and none after the hour mark. It looks likely that they were told at half time to defend their lead and stop attacking entirely. The percentage simulations reflect a close game, but the Race Chart gives us the information to understand how it played out.

Quick takes:

• Montreal exploited an early lead by defending to keep it
• DC struggled to create valuable opportunities against a team with a defensive outlook

Lastly we have a recent match between FC Dallas and Colorado Rapids in which Dallas won 2-0:

FC Dallas scored two goals within the first ten minutes and were the only team to carve out viable chances from there on in. Colorado created very little in the entire game apart from extremely low value shots. The Race Chart makes it clear that the home team was dominant and comfortable throughout while the simulations suggested that given the balance of opportunities, they were 87% likely to win this game.

Quick takes:

• Straightforward victory for Dallas
• xG representative of final score
• Colorado created next to nothing

The Race Chart adds in a straightforward but significant layer to post-game analysis. In combination with a Shot Map, we now have the game information required to tell us where shots were taken from, when they were taken, and how valuable they were. The ability to digest information quickly about multiple games means they are simple but powerful tool for analysts.

#### Article by James Yorke

james@statsbombservices.com