Quick question: what proportion of teams that have led at the end of the 1st Quarter of a Grand Final have gone on to take the Flag? Supplementary question: how big does the Quarter-time lead need to be before the probability reaches 90%?

Every time I revisit the topic of leads in the AFL I'm reminded how important is even the smallest of leads at any of the breaks, and how rapidly a team's prospects tend to improve as the size of its lead increases. 

At least, that's been my experience when I've analysed the data for whole seasons at a time. What I've not done explicitly before is to perform a simple lead analysis using data solely from Grand Finals (although, this piece from 2013 pursues a similar theme). What we find when we do this analysis is that leads in Grand Finals are of about the same importance as leads in all other games.

In the table below, the block of results on the left provides the summary for all 116 Grand Finals, excluding those GFs that were drawn, and excluding, for the analysis of particular Quarters, games where the scores were tied at the end of that Quarter. We can see in the summary that, for example, teams with any lead at all at Quarter-Time have gone on to win 71% of Grand Finals, that teams that have led by a goal or more at Quarter-Time have gone on to win 74% of the time, and so on. The answer to the supplementary question I posed at the start of this blog is that a lead of 4 goals at Quarter-Time is enough to bump a team's victory chances up to 90%.

A slightly smaller lead at Half-Time - around 3 goals - is enough to produce a 90% victory probability, and any lead at all at Three-Quarter Time has been sufficient to get 90% of teams across the line.

Also included in the table is the same data but for whole seasons at a time, for each of the past five years. 

These results provide some context for the Grand Final results, but also allow interesting inter-year comparisons of their own.

In the current season, for example, compared to 2014:

• leads at Quarter-Time have been less safe than leads of equivalent magnitude
• Half-Time leads have been about equally safe
• small leads at Three-Quarter Time have been safer, larger leads about equally as safe

This season, in fact, teams leading at Three-Quarter Time have fared better than teams in this position during any of seasons 2010 to 2014.

Returning to the Grand Final data, we can also present it in the form of a pair of charts, one tracking the proportion of teams that have won after leading by X points or fewer at one of the quarter breaks, and the other tracking the proportion of teams that have won after leading by X points or more at one of the quarter breaks. (Note that I've shown these proportions only where they're based on at least 20 Grand Finals.)

One interesting characteristic of the lines in the left-hand chart is their slope, which reveals the rate at which a team's victory chances increase as it extends its lead. In this respect it would seem that scoring to extend a narrow lead to a moderate one by Half-Time has the largest impact on a team's chances of victory.

The chart on the right provides a more granular version of the data from the earlier table.