But, recently at least, Headingley’s fourth-innings record shows it is fertile ground for similar acts of last-gasp heroism.

In every one of the last five Tests England have played in Leeds, they have fielded first and gone on to win the match. Out of those, two wins came on the final day as they chased down a total of more than 290. The last match they lost at the ground was Shai Hope’s contribution to fourth innings masterpieces, where West Indies chased down 321 thanks to his unbeaten century. Go back even further, to 2001, and you’ll find Mark Butcher’s name in the book, his best-ever innings giving England victory in a seemingly impossible chase of 315 against a great Australia side.

Clearly, outside of Stokes’s most famous offering, there are plenty of footnotes that give substance to Headingley’s reputation fruitful in producing freakish backs-to-the-wall rescue efforts.

Delving into the stats, it’s not just individual performances that make it more likely for a side to be able to chase down a high target inside the Yorkshire border. Since 2000, Headingley has the joint-highest average for runs scored per wicket in the fourth innings of a Test match of any ground in the UK at 33.09 (equal to The Kia Oval). For the teams batting in the fourth innings, it also has the joint-second highest win/loss ratio at 1.4 – showing teams win batting fourth there more often than not.

If the criteria is changed to post-2010, the stats lean even further to the side batting last’s favour. Headingley has the highest average runs per wicket scored in that time frame, increasing slightly to 34.72. Edgbaston is the next highest at 31.16. The win percentage stays stable at 1.333, making it one of four out of eight grounds with a positive ratio (the side batting last wins more times than it loses).

In fact, the number of runs scored per wicket on the Headingley pitch gets higher as the game progresses. Since 2010, the side batting in the first innings averages 25.39 per wicket, in the second innings it’s 29.57 and in the third, it’s 33.78. Based on those averages, Headingley has been the worst ground to bat in the first innings in England since 2010, and the only one where the average runs per wicket in the first innings is less than 30.

Out of all the grounds that have hosted a Test match in England since 2010, batting first at Headingley has proved the most disastrous to a side’s chance of winning. The team batting first has a win/loss ratio of 0.666 (joint-lowest of all with Edgbaston). The side fielding first has a win/loss ratio of 1.5.

[breakout id=”0″][/breakout]

Overseas sides have been slower to recognise the benefits of putting sides into bowl first at Headingley than England have. New Zealand, India and Pakistan, the three most recent touring sides to win the toss at the ground, have all elected to bat first. All of them have gone on to lose those Test matches. Whereas, aside from the ongoing contest, England most recently won the toss at Headingley in 2019 and elected to field. Famously, that ended quite well for them.

It hasn’t always paid off. In 2015 England elected to field first against New Zealand, conceded 350 in the first innings and went on to lose the match by 199 runs. The following year, Sri Lanka opted to field and ended up losing by over an innings. Stats can get you so far but nothing makes up for poor execution.

But there are also signs that Headingley is getting even better for sides to chase totals on. In Yorkshire’s first County Championship game of the year, Leicestershire chased down 388 in the fourth innings to win by three wickets. The next match at the ground saw Yorkshire post 412-9 batting last, still falling well short of a win.

All this means Australia will need plenty to be sure the scars of 2019 aren’t reopened. With significant time being lost to rain and more scheduled over the next two days, a declaration could potentially come into play. Joe Root and Adam Gilchrist may still shudder at the thought of their own declarations at Headingley. It would be an unenviable task if Pat Cummins has to decide how many runs are enough.