Science Explains Rohit Sharma Losing 12 tosses in a Row? - Future IQ

5,432 views Wait, is this logic right? • May 01, 2025

Slog Reference: Fooled by Randomness #2: The Null Hypothesis

Description

Rohit Sharma loses 12 coin tosses in a row. A Philippines lottery draws only multiples of 9. An ayurvedic doctor claims he can guarantee a male child — just take his "ancient" medicine through your left nostril during the 9th month of pregnancy. Sounds suspicious, right? But here’s the twist: what feels like a conspiracy might just be random.

In this episode of Fooled by Randomness, we dive into the null hypothesis — a powerful concept from scientific thinking that helps you separate real evidence from coincidence. Most people either believe wild claims or just ask for proof. But the real power comes from asking: what’s the simplest explanation? Could this be chance?

This episode will change how you look at luck, pseudoscience, and even your own beliefs. You'll see how randomness can fool you, how your brain jumps to conclusions, and why the null hypothesis is one of the best tools for critical thinking in everyday life — whether you're dealing with lottery results, medical claims, or just drama in the family WhatsApp group.
Watch now to start thinking smarter.

More Videos:
Why We Only Hear About The Winners? Survivorship Bias Explained: https://youtu.be/QjDXyuBJ0UY
Believing is seeing - How Bayesian Priors Trick Your Senses: https://youtu.be/bxx-My8J_kM

Hope you enjoyed FutureIQ by Navin Kabra and Shrikant Joshi. Do hit us up on Twitter:
@ngkabra http://twitter.com/ngkabra
@shrikant https://twitter.com/shrikant

Listen it on the podcast provider of your choice: https://tapthe.link/FutureIQRSS

Chapters:
00:00 Intro
01:50 Let's get into it!
05:10 How do you figure out it’s a coincidence?
06:56 But.. even the Baba who promised male children tho?!
10:22 BUT COMPARED TO WHAT?
15:03 More examples in case you are still not convinced
19:22 So don't be fool and ask the right questions!

Sources:
https://www.iafaforallergy.com/ayurpediatrics-a-to-z/pumsavanam/: website with information about “ayurvedic” medicine to be taken by left nostril in 1st month of pregnancy for male child!!!
https://en.wikipedia.org/wiki/Null_hypothesis: Wikipedia page
https://towardsdatascience.com/hypothesis-testing-explained-as-simply-as-possible-6e0a256293cf: Simple explanation of Null Hypothesis and Hypothesis Testing
https://www.forbes.com/sites/toddessig/2018/12/30/serious-problems-with-rcts-and-ebts-exposed-by-the-satirical-parachute-study/: Parachute use to prevent death and major trauma related to gravitational challenge: systematic review of randomised controlled trials
https://terrytao.wordpress.com/2022/10/03/what-are-the-odds/: Mathematician Terry Tao on a Suspicious Lottery Result, and using the Null Hypothesis to analyze it

#futureiq #nullhypothesis #believing

Related Slog Matches

Fooled by Randomness #2: The Null Hypothesis

Manual

100.00

Transcript

- Rohit Sharma lost 15 coin tosses in a row in one day internationals. Do you think this was fixed? No, but 15 in a row is weird. No. 3 years ago, the Philippines lottery where six numbers were picked out. The six numbers were 9, 18, 27, 36, 45, 54. All multiples of 9. Yes. Does that sound rigged? Might be rigged. I mean, how can you get six multiples of 9? Another example. Okay, there's a guy in Karnatak who charges 5,000 rupees for an ayurvedic medicine described in the Vedas which is to be taken in the left nostril in the first month of pregnancy and he guarantees a male child. Can this be real? Uh, no. I

0m 48s mean I I ask what you want to know the evidence? Yeah. Yeah. It comes with a money back guarantee. He has hundreds of photos of happy customers including some celebrities and you personally know at least five happy mothers of spoiled male brats. Now does it sound like there is enough evidence? I mean that is solid evidence. So maybe see the problem with all the situations is that we tend to ask the wrong questions right? Okay. What I mean basically the believers they don't ask any questions at all. They're like, "Oh, Vedas, so must be true." Right? On the other hand, smart people is like, "Oh, let me get the evidence." And then when they see evidence, oh,

1m 30s 5,000 happy customers, and they're like, "Oh, that's a lot of evidence." Right? Both of these are mistaken, right? What you really need to do is the topic of this episode, which is the null hypothesis. Is there an alternative explanation? Right? Okay. Yeah. So let's start with a simple example. Okay. Okay. Uh you are in a movie and Amitab Bachan comes and he tosses a coin and it's a heads right. Four times in the movie he tosses the coin and it's a heads. Do you think that coin is rigged? Is that movie cholelay? Well are you Dhandra? No.

2m 10s Right. So then it's not show. Okay then uh might be rigged. Well, see, let's ask for an alternative explanation. Yeah, your primary explanation is that this coin is heads on both the sides. Okay. The alternative explanation, which is like the most common thing possible, is that's a normal coin. Mhm. And you got four heads in a row just by chance. But what are the odds of that? I mean, statistically, you can very simple, right? Each head is 50% chance. four of them in a row is one upon 16 which is around 6%. Okay, 6% is not a very low number, right? It can happen randomly by chance without it being very surprising.

2m 56s Yeah, 6% is fairly significant and believable also. Yeah, I mean because you know something which is 6% possible you it can happen every once in a while, right? True. Uh so there is no need to directly jump to a conspiracy theory for that. Agreed. But then 15 heads in a row is a lot less possible, right? I mean it's like 1 in 30,000 or something like that, right? The point is though, how many ODI matches have happened, right? India has played around 1,600. Yeah. Okay. And at each one of those matches there was a possibility of 15 heads in a row. So we have had 1,600 chances and there are other countries playing. Each of those countries would

3m 40s have had a chance and anytime you get 15 heads in a row, no not heads, 15 losses in a row, you would have been jumping up and down saying conspiracy, right? So just if you add up all those chances, it comes to around 3 or 4%. which is again not that uncommon. And you if you factor in all of the other matches that were played between these 15 coin toss losses, then there might even be a greater chance of 15 losses in a row happening with other captains also involved. Yeah. So yeah, now I see that it's not really as rare as we think it is. One in 30,000 is rare, but then a lot of matches have also been played.

4m 20s Yeah. See, that's the problem with human intuition, right? I mean like as Goldfinger in the James Bond movie said, once is happen stance, twice is coincidence, three times is enemy action, right? So when something surprising happens three times in a row, we just jump to the conclusion that this is a conspiracy against you, right? But you look at some people, right? It's like their entire life there is enemy action against them, right? They don't stop to think that this is probably coincidence. We are not pointing fingers at anyone. We're just saying it. I mean this is what is the usual all the people who think we are pointing fingers at you. Yes, we are pointing fingers at

5m 00s you. So this episode is for you. Yeah. All right. But then uh let's look at the other kind of people who think it might be a coincidence. How do you figure out if it's a coincidence? Well, what you do is that you don't just generally ask what is the probability of this, right? What you do is you say okay what is your explanation right one explanation is that there is somebody rigging the tosses okay then you have to ask what is the null hypothesis which means what is the most simple what is the most normal explanation for this right like Okam's razor like Okam's razor right so but here we get into a little more detail so

5m 42s you ask what is the simplest most non-conspiracy explanation Okay. And then for each of those explanations, you can use probabilities and attach a probability to it. Right? Right now the simplest non-conspiracy explanation if it has a reasonable probability. Right? Then we say that you know what this thing can happen. So then you just go with it. Then you say well there is not enough evidence to say that this is a conspiracy. Right? like 6% is a reasonable probability for four times heads in Amitab Bachan's coin case or 15 losses in a row is still a reasonable probability for Rohit Sharma losing a coin toss in one day international correct so in science what you are supposed to do is that you

6m 33s are supposed to have a hypothesis which is what you your research is trying to prove and then you also mention what is the null hypothesis and then you have to prove that your null hypothesis has less than 5% chance of being the re explanation right that is what is known as the p value is less than 0.05 05 that 5% is very interesting but I want to quickly step back for a minute and uh take this example of hypothesis and null hypothesis that you just gave or explanation and apply it to the Baba who promised male children. How would you go about applying this hypothesis and null hypothesis in that scenario? So hypothesis is that Baba has a medicine

7m 17s which actually gives you male child. Okay. Null hypothesis is that Baba is a sham that medicine does nothing. Okay. Now let us look at the evidence that is why is this a null hypothesis? Because that's the most common thing that happens in the world. Right? The assumption being that birth is a genetic lottery statistical lottery and whatever the Baba is doing is not affecting the birth and the gender of the child sex of the child. And uh a more sophisticated uh explanation for the same thing is that of all the things we have learned in science and a lot of experiments and in our textbooks and past experience, right? All of them together tell us that

8m 06s something like this is unlikely to be true. Right? If you were paying attention, this is a prior in Beijian inferencing. Check out our previous episode. Right? Yeah, you should. So yeah. So now the null hypothesis is that Baba is a sham that medicine does nothing. True. Now the question is the evidence that is being shown to us can that be explained by the null hypothesis with a reasonably high probability. Right? Okay. Let's take this situation. Okay. 10,000 couples go to the Baba. Baba gives them all medicine. Right.

8m 42s They go back. But according to our null hypothesis that medicine is a sham. So how many boys will be born? Half of them. 50/50 chance of a boy and a girl. 5,000. Correct. So now Baba has 5,000 testimonials that the medicine worked. Yeah. The other 5,000 where it didn't work. Right. Why aren't they complaining? Why aren't they complaining? Because Baba has a money back guarantee. Baba returned their money. Yeah. Right. Keep in mind though that he's kept the 5,000 rupees of each of the boy childs, right?

9m 15s So he's made a lot of money even though he gave a full money back guarantee. In fact, he could have given uh you know 7,000 rupees back and still made a lot of money, right? Right. Third, let's say you have 10 friends who went to the Baba. Okay. Five of them got girls. So they're staying quiet, not telling anyone about it because they also got their money back. Yeah. The other five got boys and they are going around telling each and every couple they know that you have to go to that baba. You have to go to that baba. Right? So that explains all the evidence without having to break a sweat. True. Right. This is where the

9m 52s null hypothesis comes in. There is an alternative simple explanation. Yeah. And if we were to actually rigorously prove this, I suspect we would have to do an RCT with this medicine from the Baba and then test the null hypothesis in that RCT. Yeah, but if you were paying attention in the Beijian inferencing episode, you know that an RCT is not necessary. The prior on Baba being a sham is so high and evidence doesn't change that. True. That is indeed true. But what I'm sensing here is every time you come up with a hypothesis, you have to have to check it against the null hypothesis.

10m 29s Not the null hypothesis. What do you mean not? There can be multiple null hypothesis and you have to choose carefully. Let me give an example. Okay. So let's say somebody comes along and claims that Wikipedia is not a reliable source of truth because anybody can edit it. Yeah, we get a lot of those comments in our uh episodes also. So this is the hypothesis. Okay. Now let's ask what is the null hypothesis right? What is the null hypothesis? So I mean basically you can think of the null hypothesis as the opposition party. Right? So the party is Wikipedia is not reliable. Is the opposition Wikipedia is 100% reliable, doesn't contain a single mistake or is

11m 12s it that Wikipedia is 99% reliable and 99% is good enough? Is that no? Or it is that Wikipedia is less than 5% uh unreliable and that's P equal to 0.05. Which of these makes the most sense? So probably the second statement because 99% is see I just gave you three choices and you picked the middle choice okay you fall for the standard trick the real problem is that all three of these questions are the wrong questions right okay the correct null hypothesis in this case is compared to what the null hypothesis is Wikipedia is as reliable as encyclopedia Britannica correct and now The opposition statement, the hypothesis is that Wikipedia is less reliable than Encyclopedia Britannica R.

12m 08s See, suddenly now we are doing an applesto apples comparison, right? And I mean from the previous episode you know that it turns out that Wikipedia is actually that reliable as reliable as Encyclopedia Britannica. So the null hypothesis is true and the other hypothesis then has to be rejected as being a conspiracy theory. Right? So when constructing a null hypothesis always try and construct a hypothesis that gives you apples to apples comparison. Absolutely right. So this thing compared to what is a very very powerful statement in real life which most of the time if we go by instinct we forget about it right examples okay let's say people have started using chat GPT for medical advice right lots of

12m 54s people complain that no no no no chpt makes mistakes compared to what right because the alternative isn't that there is like a 100% accurate god giving you medical advice, right? You either ask Chad GPT or you ask a doctor and then you have to ask yourself how many of those doctors can give you wrong advice, right? There's that's one possible null hypothesis. Yeah. A different compared to what null hypothesis is that person won't go to a doctor at all because that person can't afford a doctor or that person lives like 70 kilometers from the nearest doctor. Then you have to ask the question is Chad GPT's advice worse than not going to a doctor at all right or

13m 42s worse than following whatever priors he has yeah just to be clear I am not saying that Chad GPT is a great doctor and you should all be using Chad GPT for advice right all I am saying is that you should know that there is a null hypothesis there and that is the comparison you should do before you decide which is the right thing Right? Essentially what he's saying is don't judge a book by its cover. Ask always what is the alternative? What is the null hypothesis that is causing this hypothesis to gain precedence? Yeah.

14m 17s Couple more examples, right? Yes, please. Self-driving cars are the big thing these days, right? Every time a self-driving car makes some mistake and there is an accident, it makes headlines all over the world. And then they are like, "Oh my god, self-driving cars, the technology is not yet there and it is killing humans." And sure, killing humans is a bad thing. But compared to what? Because if the car is not a self-driving car, then a human driver is going to drive it. And there is evidence that 12 times more humans are killed by humans compared to self-driving cars.

14m 54s Yeah. But then that goes into a very different territory of you can hold a human responsible how do you hold a self-driving car responsible etc etc and that's a whole different area of discussion but I understand the concept of null hypothesis here I understand what Naven is trying to say I'm just saying this discussion is nuanced but the concept of null hypothesis is what you need to take away from this example other examples please cricket I mean we should give cricket examples right excellent that oh DRS Hawkeye makes mistakes right the technology has this much of a margin of error right or V if you are more into football virtual assistant referee which is the same

15m 31s thing as DRS somewhat and so obviously the purists complain that you know you can't use this technology which is fillable and which has this much percent of error and so on but compared to what you're comparing against humans I mean the LBWS these days versus 10 years ago you should know that DRS is clearly better than the humans yeah somebody should actually go back and look at all of the LBWS that happened before DRS and figure out what is the margin of error on those and how many were erroneous and all of that. But one thing I want to point out which is that we started this discussion by saying that the null hypothesis there can be multiple of

16m 10s those. Right? So let me give a really nice example of how there can be multiple null hypothesis, right? So um the Philippines lottery. Okay, this is a lottery in which six numbers from 1 through 55 are randomly picked and anybody who has exactly those six numbers on their ticket wins the big jackpot jackpot. And the numbers that got picked was the table of nine. 9 18 27 up till 54, right? And that clearly looks like it was rigged, right? Yeah. Because getting multiples of 9 that precisely. Wow. So Terrence Tao, one of the greatest living mathematicians, right? Yeah. Uh he has written a detailed post analyzing it. I recommend all of you go read that. I'm not going

16m 55s to talk about the entire analysis. I'm just going to mention a few of the hypothesis you mentioned. Right. Okay. The null hypothesis is that it was not rigged. That's the most common thing that happens in the world. Lotteryies are not rigged. The alternative hypothesis. What is your alternative hypothesis? Okay. The possibility is the lottery is rigged by some corrupt officials who on that particular date decided the random winning numbers in advance so that they could win the money and they were stupid enough to choose 9 18 27 as the numbers. Okay. Okay. That sounds really stupid. Okay. So obviously we need some other hypothesis, right? So maybe the lottery machine has developed a fault and it only outputs numbers that

17m 44s are multiples of nine. Could be except that that fault happened only on that one day and not the day before that and not the day after that and never again and you know then what I mean is that a conspiracy of any type right what are you complaining about? Uh yeah, that's not a conspiracy, but that's a lucky break. And it's as lucky a break as getting multiples of nine in the or the lottery is rigged by a cult who worships the multiples of nine. What are these Ashoka's nine or something, right? Or like you know a divine being decided to send a sign to humanity by placing this unusual pattern in a lottery. All right, let's not even go

18m 29s there. And that God chose only this one lottery in this one country on this one day. Right? Let's not even go. So the point being that people just say, "Oh, this looks weird. So it's a conspiracy of some sort." But when you think through what are you actually implying, right? The there there has to be an alternative explanation that you are coming up with and often that turns out to be ridiculous. Yeah. None of these satisfy the AAMS razor test. And also when you actually think about it in terms of statistics, the possibility of 9 18 27 36 45 54 being the results is as probable as 1 2 3 4 5 10 20 30 40 50 or

19m 16s any of those other combinations. Right? So I will not get into that in detail. Terrence Tao has already done it. Read that article. But what I want to say is to talk about this whole big picture here, right? That whenever there is a claim which seems a little outlandish or surprising, right? Your first instinct should be to say compared to what what right what is the normal simple non-conspiracy explanation and what is the probability that that explanation produced this non-standard output? Right? H often it turns out that it wasn't such a low probability right the second thing is whenever somebody says X is bad you have to say compared to what because right now what

20m 06s is happening is Y so you can't just say X is bad you have to prove that X is worse than Y otherwise that statement has no meaning right like Wikipedia versus encyclopedia Britannica right sometimes the null hypothesis can be as stupid as just dumb luck, right? So, you have to ask yourself, could dumb luck have created this? Yeah. Like the Philippines lottery was definitely dumb luck when you think about it, right? And never ever forget that if something is happening a large number of times, most of which get ignored. Then it is possible that something very rare just happened because there were so many trials, right? Like the 15 coin losses lost in a row by Rohit Sharma. Correct. H one

20m 51s thing to be careful about though is to not take this too far right as in as in this can lead to what about right I say today's leader is bad right and then you come back by saying compared to what and then you point out that you know the leader 10 years ago was equally bad and technically it is true right the null hypothesis is that all leaders are bad but that doesn't mean that we give up right We should still be holding our leaders to high standards. That's how we make progress, right? Absolutely agreed.

21m 26s So when you say compared to what be very careful and cognizant of the slippery slope of what about that you might get into also the other problem with this is that sometimes you might not be able to prove the null. Right? So for example uh if I say that oh parachutes don't work. Okay. Yeah. We've done this before. I'm not getting into this again. If you want to know why parachutes don't work, we'll line that up for you. Check that out. This is Shri Khan. This is Naven. This is Future IQ.