chasingthedon
International Regular
'Clutch' is a term used widely in the American sports media to define those situations in a match when a higher than usual level of intestinal fortitude is necessary - or as we English might say, squeaky bum time (SBT).
A discussion on the forum following my feature on batting impact, notably highlighting Tom Graveney and Hanif Mohammad, which chiefly involved, myself, Bolo and Starfighter, regarding whether or not the phenomenon of 'clutch' ability even existed and could be shown objectively, or was it simply 'a spurious [concept] founded on subjective impressions', as Starfighter put it.
I have spent some time in the ensuing months trying to get to the bottom of this, however I'm having some trouble deciding on what exactly a 'clutch' situation is in Test cricket. I do have a database which includes all of the FOWs from the first Test in 1877 to Test number 2310 or so, from which the win probability (WP) can be determined for any given situation. The difficulty is determining which of the many situations is a clutch situation, or SBT. I can base this on WP, but I would like to get a feel for how the forum sees it.
Initially I split the various match situations into easy (70%-100% WP), neutral (30%-70%) and tough (0%-30% WP). However, in situations like Botham at Headingley '81 and Nathan Astle in 2002, they had both stated that they felt there wasn't much pressure, i.e. nothing to lose by going for it.
So I changed the lower cut-off to 10%. By the way, these cut-off values aren't arbitrary, rather they're based on looking at the various situations at those percentages and discerning how I thought a batsman would consider the situation from an SBT point of view. The problem with excluding the lower WP value performances is that, while the give-it-a-go smash-fests are excluded, so are a lot of important knocks, such as Lara's 153* in 1999 and Ranji's 154* in the 1896 Ashes. So I re-assigned the tough category to 0-30%.
Upon further reflection, however, it appeared to me that a batsman would be sufficiently motivated if the match was close or, even, slightly ahead. So I decided then to nominate SBT as 0-60%.
I was also at this point restricting to the team's second innings, or innings three and four. The reasoning behind that was because the system may think that the team batting second was behind in WP if they were following a large total, because early on the system hasn't adjusted to how the pitch is. This resulted in the proportion of SBT runs at something like, for example, 15% for David Warner, who would not be batting with his team behind in WP as often as, say, Mushfiqur Rahim, whose ratio of clutch runs is almost double Warner's at around 29%.
But a close look at the innings which the second innings restriction would exclude meant we would classify an innings like Frank Worrell's 191* carrying his bat in the second innings at Trent Bridge in 1957 as non-clutch.
So where I am now is using performance in all four innings, where the batting team's WP is at less than 60% when the batsman came to the crease, as clutch. This of course means that a significant portion of the batsman's career Test runs are classified as clutch, i.e. as much as 87% for Rahim, but even Warner is at 65%.
So are these runs 'important' runs, rather than 'clutch' runs? Should these be separate categories? Does it seem reasonable that 35% of Warner's runs were less important in terms of contributing to the state of the match, given that the team was already well ahead when he came in?
I await the views of the CW cognoscenti.
A discussion on the forum following my feature on batting impact, notably highlighting Tom Graveney and Hanif Mohammad, which chiefly involved, myself, Bolo and Starfighter, regarding whether or not the phenomenon of 'clutch' ability even existed and could be shown objectively, or was it simply 'a spurious [concept] founded on subjective impressions', as Starfighter put it.
I have spent some time in the ensuing months trying to get to the bottom of this, however I'm having some trouble deciding on what exactly a 'clutch' situation is in Test cricket. I do have a database which includes all of the FOWs from the first Test in 1877 to Test number 2310 or so, from which the win probability (WP) can be determined for any given situation. The difficulty is determining which of the many situations is a clutch situation, or SBT. I can base this on WP, but I would like to get a feel for how the forum sees it.
Initially I split the various match situations into easy (70%-100% WP), neutral (30%-70%) and tough (0%-30% WP). However, in situations like Botham at Headingley '81 and Nathan Astle in 2002, they had both stated that they felt there wasn't much pressure, i.e. nothing to lose by going for it.
So I changed the lower cut-off to 10%. By the way, these cut-off values aren't arbitrary, rather they're based on looking at the various situations at those percentages and discerning how I thought a batsman would consider the situation from an SBT point of view. The problem with excluding the lower WP value performances is that, while the give-it-a-go smash-fests are excluded, so are a lot of important knocks, such as Lara's 153* in 1999 and Ranji's 154* in the 1896 Ashes. So I re-assigned the tough category to 0-30%.
Upon further reflection, however, it appeared to me that a batsman would be sufficiently motivated if the match was close or, even, slightly ahead. So I decided then to nominate SBT as 0-60%.
I was also at this point restricting to the team's second innings, or innings three and four. The reasoning behind that was because the system may think that the team batting second was behind in WP if they were following a large total, because early on the system hasn't adjusted to how the pitch is. This resulted in the proportion of SBT runs at something like, for example, 15% for David Warner, who would not be batting with his team behind in WP as often as, say, Mushfiqur Rahim, whose ratio of clutch runs is almost double Warner's at around 29%.
But a close look at the innings which the second innings restriction would exclude meant we would classify an innings like Frank Worrell's 191* carrying his bat in the second innings at Trent Bridge in 1957 as non-clutch.
So where I am now is using performance in all four innings, where the batting team's WP is at less than 60% when the batsman came to the crease, as clutch. This of course means that a significant portion of the batsman's career Test runs are classified as clutch, i.e. as much as 87% for Rahim, but even Warner is at 65%.
So are these runs 'important' runs, rather than 'clutch' runs? Should these be separate categories? Does it seem reasonable that 35% of Warner's runs were less important in terms of contributing to the state of the match, given that the team was already well ahead when he came in?
I await the views of the CW cognoscenti.
Last edited:
