Putting predictions on this blog allows testing of prediction against results. In this post I’ll look back at what I said before England’s tour of the West Indies in 2019.
I was surprised how few concrete predictions were included in previous posts. Next series I may include player by player predictions, so there are more data points.
1.No reason to model Jennings’ expected Test average as anything other than 33.
❌ Jennings averaged 16. Though it was only four innings, it’s hard to see that prediction as a success! The extra data takes his expected average down to 32.
2. One spinner is the right choice
✅ Rashid’s match figures of 26-1-117-0 with the ball and 12 & 1 with the bat showed England the error of their ways.
3. History says expected average by bowing type Spin 32 Pace 26
✅ Actual averages: spin 35 pace 21, which reflects the quality of bowing on display – both teams have better quicks than spinners.
4. West Indies’ best chance will come if their fast bowlers can keep England under 225 in one innings
✅ Both West Indies victories included innings where England scored under 225. England won the third Test scoring 277 & 361-5. I don’t really like this kind of prediction though: Cricket is won by taking 20 wickets and scoring more runs than your opponent. How you do that is unimportant.
5. England 2019 are at about the level of the 2005 Ashes side, by having no weak links rather than being packed with world-beating batsmen.
❓Most would say that England’s batting was stronger in the past, but the current team has huge potential. My view is that England’s current batting is fragile because it is not that good, while some pundits would have you believe that England are afflicted by “amazing-but-collapse-too-often syndrome”.
6. England have a one in three chance of Whitewashing the West Indies.
⚠️ I stand by this prediction- though hard to appraise the success of this. Just because it didn’t happen doesn’t mean there wasn’t a 33% chance of it. Equally if it did happen that wouldn’t tell us much from one prediction.
There has recently been interest in Keaton Jennings’ average against pace. Two failures in Barbados have stoked this discussion. His average (26) in 16 Tests is below his expected average (33) based on County performances over the last three years. Generally, I would choose the big sample size (County Cricket) over the smaller sample size (Tests), and so rate his expected average at 33, not 26.
But – can we learn anything about technical flaws from Jennings’ Test performances to change that view? Specifically his average against pace:
Keaton Jennings‘ average against pace (16.90) is the lowest of any opener to have played more than 15 Tests, for games in which ball-by-ball data is available.
Wisden (Jan 26th 2019, via Twitter)
I’ve had a look at his performances over the last 3 years on the county circuit. The hypothesis is that there are some very good pace bowlers in County Cricket, and as an opener Jennings will face them (a middle order batsman might be able to make hay without facing much of the best bowlers).
The data supports this hypothesis – 68% of the time he faces at least one opening bowler with Test experience.
Keaton Jennings has played two of the last three seasons in Division 1, scoring 11 hundreds, and making runs in a variety of conditions (including April and September- when the deck is stacked in the bowler’s favour). His three year average isn’t amazing, but the key point is that one can’t look at the above data and conclude that Jennings has a problem against pace bowling.
As an aside, this piece is a reminder that I need to build a way to combine the Test performances to the First Class performances to ensure I’m using every available data point in appraising batsmen.
Conclusion: There is no reason to model Jennings’ expected Test average as anything other than 33. Plenty of people will disagree with that!
One of the benefits of twitter is hearing new ideas. Jonas (@cric_analytics) has suggested the third innings should pause when the lead reaches 300, then the fourth innings takes place.
That way, a team that’s winning doesn’t have to pointlessly bat until the lead is over 500, before crushing an inferior opponent. Here’s how Jonas puts it:
I’ve modelled how this would work in practice, with the aim of answering two questions:
Does this make the strong team more likely to win? (Probably)
Is the game over sooner? (Generally)
Here’s the summary from the single scenario I looked at:
Scenario: West Indies vs England, Bridgetown.
England have batted first and scored 360. West Indies slipped up and were bowled out for 210. We join the action at the lunch on day three. England lead by 150. Two versions of this were modelled: under the existing laws, and temporarily declaring the third innings if they score 150 more.
Let’s see what happens:
In 92% of cases England made it to 150 without being bowled out – and so, with a lead of 300, temporarily declared
West Indies scored under 300 83% of the time – so the third innings did not need to re-commence
When the West Indies scored more than 300, sometimes the game meandered to a bore draw because the West Indies couldn’t confidently declare
Here’s the distribution of match end times depending on which rules apply:
We can see that there’s a big shift towards Day 4 finishes under compulsory declaration at 300 – mainly from the team batting fourth being bowled out for less than 300.
Worth noting the result wasn’t significantly affected by the rules being used. This would be different in other scenarios – such as if there was less time in the game.
Conclusion – This could be very useful in county cricket (where matches are only 4 days long). Suggest more modelling is required (especially scenarios where the odds are shifted from the draw being favourite to a result being favourite). A trial in County Championship Division 2 would be fascinating.
West Indies can beat England against the odds, but they’ll need their pace bowlers to perform.
The blueprint – Bridgetown 2015. 1-0 down in the series, with a first innings deficit of 68, the West Indies were about to be batted out of the Test.Hearing a wicket fall, a reveller in the Party Stand asked “Was that Trott or Cook?” and was baffled to learn that it was in fact Root, and England were 28-4. The new ball had done the damage, and by the time 20 overs had been bowled it was 39-5 and the game was back in the balance.
West Indies were eventually set 192. Darren Bravo marshalled the batsmen to the target with five wickets in hand. The hosts had accrued only three scores over 30 in the Test, but somehow pulled off an unlikely victory, and drawn the series 1-1.
With that surprise firmly in mind, let’s make some informed predictions for the upcoming series.
1) One spinner is the right choice. This decade the average is 32 for spinners, 26 for pace bowlers. It may be that pitches are turning more than they used to, and it’s true that spinners get 37% of wickets in the Caribbean, but this turn hasn’t delivered cheaper wickets. That said, if a team can reliably judge a pitch as more spin friendly than the average West Indian pitch, then they should go with two spinners – selectors just need to be sure there will be more in the pitch for spinners than quicks before making that decision.
2) West Indies’ best chance will come if their fast bowlers can keep England under 225 in one innings. Turning pitches or not, the West Indies have no elite spinners. If they are going to win this series it will be through devastating fast bowling.
They are unlikely to amass buckets of runs – so Holder’s bowling unit needs to neutralise England’s batting. Specifically, if England score fewer than 225 in one innings, that sets up a target within the range of the West Indian batting.
Taking all factors into account, modelling suggests the probabilities for the first test are: 24% WI. 7% Draw. 69% Eng.
West Indies will probably lose: their batting and spin bowling is inferior to England’s. But if we’ve learned anything from the 2015 series, it’s that home advantage is real, and the new ball could do some serious damage, leaving mystified England supporters to ask “was that Burns or Jennings?” as Stokes returns to the pavilion.
Clive (@vanillawallah) was looking at Kohli’s scores in ODIs since the last World Cup, suggesting that:
Kohli is consistent
He succeeds more than he failures
To check this, I compared Kohli’s performances against what my model would expect him to do – Kohli’s run ranges are broadly in line with what you would expect given his average. His consistency is a consequence of his ability, rather than a specific trait of his batting.
I modelled 1,000 innings for Kohli batting at 3 for India,
with an assumed average of 95 (his average over the last 54 games / 3 ½ years).
The results show slightly more single figure scores in the
real world vs model, offsetting slightly fewer scores in the teens. This is
likely due to small sample sizes.
Two interesting observations:
In a quarter of innings he would (and did) score a hundred. Phenomenal.
The run distribution is skewed towards the 30-50 range by Kohli running out of time – caused by India successfully chasing down targets and the match ending while he is mid-innings.
Rest of the Top 3
Clive also pulled in data on all other top 3 ODI batsmen since the last World Cup. This is a much larger sample size- and worth checking the distribution as a way of verifying my modelling.
Simulating 1,000 innings with two openers: one of whom averages 35, one of whom averages 45 reasonably reflects the real world distribution of scores that Clive showed.
– The real world having more low scores (probably from the
times when weaker openers have been selected)
– More hundreds modelled than seen.
P.S. Appreciate this is White Ball ODI Cricket rather than Red Ball Data. Don’t tell the Branding Police.
In this post I consider the evolution of England’s batting – how it steadily improved through the 2000s, peaked in 2010-11 (as England became World Number 1), tailed off from 2013, and is only recently recovering.
I took the career averages of the top 7 batsmen for each England Test since 2000, and adjusted them for the age of the batsmen (I’ll cover how I do that in a later post). To eliminate artificially low results, Nightwatchmen are excluded. Where someone only played a few Tests, I made a judgement about what their long term average would have been had they played more Tests.
To bring out the trend, the chart above is smoothed with a moving average of the last five Tests.
Evolution: 2000 to 2019
Weakest Team: 29th June 2000, vs West Indies (Home) – Age adjusted Average 218
Vaughan, Hick, Stewart, Knight, White
If you don’t want to remember how bad England used to be, I suggest you skip to the next paragraph. Don’t worry, we’ll be talking about 2005 soon enough.
Let’s reel off why this team was the weakest this century: England had no batsmen in the top 10. Over five tests the highest total England could manage was 303. Only Trescothick and Atherton averaged over 29. Ramprakash and Hick never settled at Test level. By 2000 Hick was 34 and Stewart was 35. The weakest link was White – averaging 25 in 30 Tests is not enough for a number 7. The need for “the next Botham” was real.
That England won that series 3-1 was down to Cork, Gough, Caddick and White dominating with the ball rather than England imposing themselves with the bat.
England vs West Indies in 2000 marked a watershed for West Indies cricket: this was their first series defeat in England since 1969. Their record in Tests in England since is W1 D2 L13. England were on the up.
That’s more like it. Five of the fifteen England batsmen in the modern era to average over 40. This was a good batting side (rather than a great one), with room for improvement at 6 and 7. Flintoff was a proper all-rounder – a luxury England had not had for a long time. He was an all-rounder with a career batting average of 32 however. Geriant Jones was carried that summer, averaging 25 (in line with his career average of 24).
Strongest Team: 26th December 2010, vs Australia (Away) – Total Average 305
It’s too soon for many to realise just how good this team was: World no.1 from August 2011 to August 2012, with a strong enough four man bowling attack to confidently play six specialist batsmen.
In the 2010-11 Ashes series six of the top 7 averaged over 40; they accrued nine hundreds in only five tests.
But the side was aging: Collingwood 34, Strauss 33, Pietersen 30. The eldest (Strauss & the retiring Collingwood) needed to be replaced in 2011. As it was, Strauss stayed on for 18 tests, but would pass 100 only twice more, averaging 31 after the 2010/11 Ashes.
As players retired they were, predictably, hard to replace. England were also unlucky in that Pietersen and Trott didn’t go on to play full careers with England.
By March 2014 England’s ICC Test Rating had slumped to 100: they went from best in the world to average in 38 months. In May 2014 they lost a home series against Sri Lanka. Some stars (Cook and Root), some young players being played too early to succeed (Ali and Buttler) and Bell had gone on a bit too long.
Current Team: 23rd Jan 2019, vs West Indies (Away) – Expected Average 268
Not bad, probably the best selections that could have made, and should be too strong for the West Indies.
It’s important to see this side for what it is: lacking in stars, yet well balanced with three all-rounders. With Ali bolstering the batting at 8, this team are likely to continue the trend of winning at home but losing away against the top 6.
Verifying the Data
To check this model (age adjusted batting average) against reality, I compared this to the ICC rankings. The correlation is clear. Worth noting that since the Age adjusted Batting Average is smoothed using a 5 point moving average, there is a time lag in the orange curve. This correlation is surprising as the ability of the top 7 batsmen makes up less than half of the strength of a team (the remainder being bowling ability and tail batting strength).
2019 are at about the level of the 2005 Ashes side, by having no weak links
rather than being packed with world-beating batsmen.
Managers tend to pick a strategy that is the least likely to fail, rather then to pick a strategy that is most efficient. The pain of looking bad is worse than the gain of making the best move.
In the last 35 years England have had just 15 batsmen who averaged more than 40 over their career. Expectations should shift: aspire to players averaging 40; accept batsmen averaging 35.
The chart below may surprise you – it surprised me. How could barely any recent English batsman reach the benchmark set for them? Averaging 40 (at least in my head) was a minimum, not an elite average.
The data speaks for itself- 45 isn’t the new 40. 35 is the benchmark, and has been for a long time.
We, the red ball loving hordes (and our journalist generals) need to help the selectors by having realistic expectations.
The selectors should return the favour: stick with players that are good enough, even if they aren’t stars, and even if pundits are piling on the pressure.
Next time someone is 10 tests into their career, averaging 34 and with the data saying they would average 35 long term, let’s not call for a change because they aren’t scoring enough. Only remove them if a better prospect comes along – not someone with similar numbers who we might want to gamble on.
There’s a great case study: Andrew Strauss retired in 2012, and received wisdom is that he is yet to be replaced as an opener. We wanted the next Strauss. We should have been looking for the next Rob Key (15 tests averaging 31 between 2003-2005 while we waited for the next star batsman to come along).
Remember who Carberry got his runs against? An away Ashes series in 2013: Harris, Johnson, Siddle, Lyon, Watson. Those 281 runs were well earned.
With hindsight, pretty much every pick between Robson and Jennings was an error. England had viable alternatives for Strauss 3 times: Compton, Carberry and Robson. Having rejected them, playing people out of position (Trott / Ali) and gambling on youth (Duckett / Hameed) as the next cabs off the rank as England moved ever further down the list of possibles.
England chose weaker options because they weren’t willing to settle for a batsman averaging in the low-30s. That cost England runs- and since the selectors’ are employed to pick the best team possible, this is a failure. One they don’t get criticised enough for. Fear not, dear reader, we know England’s best batting options– and will collectively tut if the selectors deviate from them!
Conclusion: England should hold their nerve, even if Burns and Jennings are only averaging 33 coming into the Ashes.