Super Overs are bad and boundary countbacks are worse, so here are some better ideas

I’m 100% NOT HERE to relitigate That Final, I promise.

England won the World Cup and that’s that. They won the first final to go down to a Super Over (an idea which has always been a crappy way to decide matches) let alone a boundary countback (an idea which had, until that exact moment, been too obscure to matter). If the World Cup ever gets so close again, it would be a mistake to re-employ these methods, neither of which reward what ODI cricket ought to reward. That’s what got me thinking about better ways to split tied matches.

Cricket World Cup: England beats New Zealand in stunning final
Whatever the rules were, we LIKE ’em!

When you think about close cricket matches and the question of “who’s winning”, you quickly realise that Duckworth, Lewis, and Stern have already done heaps of really high-quality thinking about exactly this topic. Heaps of it. That’s why the second half of this post stands entirely on their shoulders. But first we need to get through Part One.

Part One: Super Overs and boundary countbacks suck

A lot of people won’t need to read this part. But if you don’t believe that Super Overs suck, let’s spend a few paragraphs correcting things.

Super Overs are a really bad version of what other sports call extra time

In soccer, 'stoppage time' can lead to confusion and ...
“90 minutes wasn’t long enough to find out who’s the better team, so we’re giving you an extra 1:48s. It was the ICC’s idea! Go! Go! GO!”

A lot of sports have extra time when they need to separate two finalists, so isn’t the Super Over just another example of this common practice?

No.

I can’t think of anything comparable to adding a single over to a 50-over match. It’s another 2% of playing time. Imagine FIFA putting drawn finals into overtime for 54 seconds each way, or the French Open replacing tie breaks with a Super Point.

Instead FIFA adds on two 15 minute periods, or an extra third. Rugby union is less generous but still builds in an extra quarter (10 minutes each way). In an ODI, the equivalent would be 75 or 100 deliveries each.

St Kilda comes from behind to draw with Collingwood in AFL ...
This is what everyone looks like when they just got told the whole match was worth nothing and we’re all coming back next weekend.

(Don’t look to the AFL for a useful model here, though. When Collingwood and St Kilda were tied at the end of the 2010 Grand Final, they just blew the whistle and invited everybody back for a rematch next week. Seriously.)

We haven’t even mentioned that cricket isn’t like soccer, tennis or rugby in another important way: Extra time (or a tie-break) in these sports keeps the current action going. A Super Over doesn’t do that at all. The team that was chasing becomes the total-setter. Dismissals are all wiped – any batsman might return to the crease. Most of the momentum of the match is lost. Even the ball is changed. This makes it unfortunately similar to the AFL’s ridiculous rematch.

Good ODI cricket is very different to good Super Over cricket

When extra time is a significant proportion of the match, the team that eventually wins is likely to have been the one that played the sport best. Over an extra 75 or 100 deliveries (i.e. quarter or third), the two teams would have time to do what ODI cricketers are supposed to do – pace an innings, decide when (and who) to attack, use bowling changes as a weapon, and react to shifts in momentum and confidence. Players would have time to recover from a bad ball, or even a maiden.

Image result for bowling change
I wanted to illustrate one of the things that you don’t get in a Super Over. This is what happens when you Google “bowling change”.

Six balls from one bowler encapsulates none of this. It’s one day cricket on homeopathy. Rather than adding anything meaningful to a match, it captures zero of the qualities that the best ODI teams exhibit.

Good ODIs are tactical battles through hours of pressure. Batting teams balance acceleration phases and wicket-preservation. Bowling teams balance wicket-taking and run-saving. Bowlers operate in pairs. Captains react to events and wrestle to control the tempo. Get these things right and you have a strong chance of winning. Turn up to a Super Over with any of this in mind and you’re playing the wrong game. Super Overs reward short-lived pinch hitting and luck.

Boundary countbacks are even worse

Whoever came up with the idea of comparing the total number of boundaries hit probably didn’t think that there’d ever be a World Cup on the line. So let’s not be too harsh on their dumb idea. Let’s just agree that it’s dumb. What makes the over . . 4 . 1 4 more valuable than 2 . . 6 . 1? How come sixes and fours are suddenly equal?

Some matches can’t even use this method. Any time the original target was altered by DLS (due to rain), the whole countback is scratched from the rulebook.

Probably the worst thing about having a boundary countback as a second method of separating tied teams is that you’re still not guaranteed a winner. If you’re going to throw some random stat into the mix to break a tied match with a tied tie-breaker, at least make it one that can’t be tied.

Tim Southee in 2019 ICC World Cup gear
Tim Southee: Should not bat higher in tests.

(The only redeeming feature of the boundary countback is that it’s a better idea than the previous one, which counted only sixes. This has an even smaller relationship to the qualities of a good ODI team performance, and although New Zealand would have won the World Cup (by 3 to 2), it’s impossible to wish that this rule still existed. The person who thought it up probably believes that Tim Southee ought to bat higher in tests.)

Part Two: Three better ideas than the Super Over

Duckworth-Lewis (D/L) turns into Duckworth-Lewis-Stern (D ...
D (L) and L (R). Not pictured: S

The question we want to answer is, “who was the best team in this match which was so weird that we couldn’t decide it according to the normal rules?”

We already have a way to find an answer in other situations: DLS. So here are three DLS-inspired ways to separate teams in drawn ODIs.

Unlike half-a-dozen balls of smash n’ hope, these all reward good ODI cricket.

1. Fewest wickets lost

The fewer people it took to hit the tied total, the better you are at batting (and the worse the other team is at bowling). So you deserve to win. This is easy to understand and uses numbers that are already on the scoreboard.

In DLS terminology, the team that lost the fewest wickets used up fewer “resources” on their way to the same score.

Who would have won the 2019 final? New Zealand (241/8) def. England (241 all out).

Will this definitely separate the teams? No, you could have, for example, 241/8 vs 241/8. That’s when method #2, the DLS Countback, comes in.

What if I wanted to bore people with nuance? Does 241/8 really tie with 241/8 if Team A had batsmen 9 & 10 at the crease, and Team B had numbers 4 & 10? Ooh, or what if the Team A scored 241 all out and the Team B gets to 241/9 with one ball left? Do we stop and award the match based on fewest wickets lost, or does the bowling team get the chance to take a wicket and force a tie? Let’s grab a fresh pint and discuss…

2. DLS Countback

Fun DLS fact: If England had been 9 wickets down after 1 over, they would have needed a score of 231 to be ahead of par. (Huge thanks to Andy Zaltzman for emailing me this table, even if it does look a bit like robot puke.)

The DLS method tells us which team would win if a game was suddenly rained off. Once the game is “live” (i.e. the chase has lasted 20 overs) the DLS par score would become the total to beat at the time. Par depends on the original target, the number of overs remaining, and the number of wickets lost. (If the chasing team is level with par, the match is tied.)

A DLS Countback would look at the chasing team’s position relative to the DLS par score at the end of every completed over, back to over number 20 (rain-affected chases shorter than 20 overs aren’t decided using DLS, so countbacks shouldn’t go there either). Every time a team was in the lead at the end of an over, they get a DLS Countback point.

The most DLS Countback points wins. In a full 50 over chase, there would be 29 overs of countback (20-49), making 15 or more points unbeatable. This system is designed to reward the more consistent team in the second innings.

Who would have won the 2019 final? New Zealand. Counting back from over 49, the end of the 33rd is where New Zealand earn their 15th DLS Countback point.

I know this only because the great Andy Zaltzman responded to a random tweet by personally emailing me the relevant DLS table which, of course, he just happened to have saved on his laptop. Many thanks, Andy!

Will this definitely separate the teams?
Usually, yes. There is the chance of a tie, especially in shorter games. Matches already shortened to 20 overs (either by rain or by a chasing team living fast and dying young) can’t accommodate a DLS Countback at all. That’s when you move down to method #3: Last DLS Leader.

What if I want to bore people with nuance? Try to calculate the possibility of a DLS Countback being a tie in a full 50-over game. Bring wine.

3. Last DLS leader

If a DLS Countback doesn’t find a winner, the Last DLS Leader method uses the same over-by-over par scores to see when the match last had a leading team. Whichever team was ahead most recently would be declared the winner.

Simpler times: Remember the old T20 “bowlout”?

The World Cup final went to the 50th over, so the first par score checked would be at the end of the 49th. If that’s a tie, we go back to the 48th over, and so on. Even if we get to the 20th over, we keep going back to 19th, 18th, etc.

When we reach an over that ended with one team in the lead, we have a winner.

(A ball-by-ball countback would be possible, but an over-by-over countback accounts for the natural practice of timing a chase in overs.)

Will this definitely separate the teams? Not if the chasing team somehow manages to be tied with the DLS par at the end of every over, but if that ever happened we would have to worry a lot about the fateful gods above. Cricket wouldn’t matter anymore.

Who would have won the 2019 final? New Zealand. After 49 overs, England were 227/8. Their DLS par was 233/8.

What if I want to bore people with nuance? You’ve read this whole article, so you’re not the sort of person who needs help in this area…