How many more runs to come?

[Victor Ian]

If you were to tune into a test match and found a batsman on a particular score, how many more can you bank on them making, on average. All things being equal, then you can expect them to score their average. However, all things are not equal. Different batsmen perform differently once they get going. Take a look at the following. We can see that root performs quite poorly in comparison between 50 and 100 runs. However he then comes back with a vengeance, while Kohli performs badly thereafter in comparison to his peers.

	0	25	50	75	100	125	150
williamson	50	69	66	85	97	75	162
kohli	53	73	86	84	76	68	55
smith	60	77	90	96	106	150	220
root	53	64	58	67	127	159	236

I did a similar scenario for Tendulkar vs Lara, because we all know that Lara is king of staying in once he got going. As it turns out Tendulkar was better than him across the graph. It seems that not outs really hurt this analysis as they are more frequent in larger innings while larger innings become less frequent.

So basically, is there any use to such an analysis? Are there other better ways of looking at this kind of thing?

 

[harsh.ag]

Wait how is Kohli worse after 100s? Has 6 double hundreds for god's sake.

 

[NUFAN]

I swear the calculation method isnt the same for all 4 bats. How is root going to make a further 236 runs if you tune in and he is on 150.

 

[harsh.ag]

If you were to tune into a test match and found a batsman on a particular score, how many more can you bank on them making, on average. All things being equal, then you can expect them to score their average.

	0	25	50	75	100	125	150
williamson	50	69	66	85	97	75	162
kohli	53	73	86	84	76	68	55
smith	60	77	90	96	106	150	220
root	53	64	58	67	127	159	236

Something wrong with this for sure. Is it because of a large not out score followed by a hundred which has pushed Root and Smith and Williamson's numbers that high? Can be the only thing I can think of since Kohli is the best of the lot at going big.

 

[Starfighter]

I'm quite confused by this tstl.

 

[harsh.ag]

Root's 150+ scores - 180, 200*, 154*, 182*
Smith's 150+ scores - 162*, 192, 199, 215, 165*, 178*
Williamson's 150+ scores - 161*, 192, 242*, 166,
Kohli's 150+ scores - 169, 200, 211, 167, 235, 204, 213, 243

Yeah, it's because of not outs.

 

[ajdude]

i love tuning in to see root on 75 and he goes on to make 67

 

[Gnske]

i love tuning in to see root on 75 and he goes on to make 69

Good one

 

[Victor Ian]

Something wrong with this for sure. Is it because of a large not out score followed by a hundred which has pushed Root and Smith and Williamson's numbers that high? Can be the only thing I can think of since Kohli is the best of the lot at going big.

Yeah, it's the not outs that ruin things. For Tendulkar and Lara, they both average over another 500 runs*, once they get to 150. Absurd!

So I am curious how you solve this problem. The idea in itself seems sound, so I was hoping some of the mathematically, or practically, inclined can solve this. Do you throw away not out innings (or count them as out) - seems like unfairly distorting the data. Should I use other forms of average? Should I consider balls to face instead of runs made? Should the columns beyond 100 be ignored because they use too few data points and are most likely heavily affected by declarations or hitting out.

Can anyone shed light on Kohli, in the sense of match scenario? Is he often hitting out at the end, to the point he gets out but gets his team to stupid totals?

Does this highlight that Root has a mental thing with the number 100. Once he goes past that, he is rather remarkable, yet getting there, once he gets close enough to think about it (past 50) seems to make weird things happen.

edit: *(this is wrong - corrected below)

 

[harsh.ag]

Just treat the not outs as outs imo. It's about how much you would expect them to score after 150, which should take into account things like declarations and running out of partners as well.

 

[Victor Ian]

with not outs counting as outs....So just how many more runs expected, not average expected, which is the point of this thread

	0	25	50	75	100	125	150
williamson	46	55	46	50	39	35	37
kohli	49	62	66	64	50	68	55
smith	51	56	53	56	44	32	35
root	47	53	46	45	57	45	43

and just for a bit of color...


Interesting. Does this show, that quite contrary to popular opinion, that Kohli is the best batsman going around...for the team? All that matters to the team is how many runs you get for them. Who gives a **** if you are not out. Your team can not use the runs that you didn't score on the ground.

 

[Spark]

Whenever I do an exercise like thus, I jut take an average of all innings that are above the cutoff.

 

[Spark]

with not outs counting as outs....So just how many more runs expected, not average expected, which is the point of this thread

	0	25	50	75	100	125	150
williamson	46	55	46	50	39	35	37
kohli	49	62	66	64	50	68	55
smith	51	56	53	56	44	32	35
root	47	53	46	45	57	45	43

and just for a bit of color...
View attachment 23832

Interesting. Does this show, that quite contrary to popular opinion, that Kohli is the best batsman going around...for the team? All that matters to the team is how many runs you get for them. Who gives a **** if you are not out. Your team can not use the runs that you didn't score on the ground.

That's not necessarily the case though. Most countries, you'd declare before ou hit the sorts of figures India hits at home.

 

[Daemon]

Whenever I do an exercise like thus, I jut take an average of all innings that are above the cutoff.

View attachment 23833

Chanders sure loved playing for himself didn't he

 

[srbhkshk]

Nice analysis (although Kohli is not better than Smith, whatever the graph says, Smith has also had situation where he basically ran out of partners and ended up not-out at less score for no fault of his.)

Could you post the results for Lara / Sachin (and if you are not doing this manually, might be good to see the data at 5 runs interval rather than 25.)

 

[OverratedSanity]

Pretty sure Sangakkara and Lara's 100+ stats will be insane. They both converted a third of their 30+ hundreds into doubles.

 

[Victor Ian]

Whenever I do an exercise like thus, I jut take an average of all innings that are above the cutoff.

View attachment 23833

So you do what I did at the very top. (I did that in response to your answer in the maths thread). However it quickly became apparent that it becomes silly beyond a limit, most likely due to small samples and increasing not outs. For instance, Chanders is NOT going to score 200 once he gets past 50. Once past 100 runs, it is most likely irrelevant anyways, as that graph showing likelihood of not losing for various scores posted somewhere else shows. I imagine a bookie would not use this simple average if they were taking bets on the expected increase in runs.

Excepting that it shows Kohli as the superior being, the method Harsh suggested seems closer to the truth.

 

[Victor Ian]

Pretty sure Sangakkara and Lara's 100+ stats will be insane. They both converted a third of their 30+ hundreds into doubles.

I didn't check Sangakara but Lara averaged more than 500* extra runs once he reached 150, and stupid extra runs at 100 and Tendulkar was better than Lara for all the brackets I counted. I might do Sangakara, Lara, Tendulkar and Ponting tomorrow, just for the lols.

edit: *(this is wrong - corrected below)

 

[srbhkshk]

Also, about Kohli, his main failing is generally linked to tough batting conditions, pressure / good bowlers etc. don't really bother him if the batting conditions are at least fair. So it makes sense that he either gets out early or just keeps on going.

 

[Victor Ian]

Sorry, I'm doing this manually. I thought of bracketing it out in smaller bits but .....i'm lazy.