Do a lesson on random variables, the most poorly worded quantity in all of statistics since it is neither random, nor is it a variable but rather a function that maps a sample space to the real numbers...it's the source of a lot of confusion for people learning stats
Glad you call me "friend", a big compliment coming from Mr Taleb to a low profile future student of probability and statistics, thank you for yoru videos and contribution, my honor to watch you my friend
“In anything at all, perfection is finally attained not when there is no longer anything to add, but when there is no longer anything to take away.” -- Antoine de Saint-Exupery Thanks a lot for perfecting these videos and keeping them under 10min. Truly refreshing.
Not. He does not understand QM. The LLN only washes out uncertainty in quantum mechanics if the system is decoherent. A coherent state can exhibit macroscopic superposition - extreme uncertainty.
Very fortunate to have access to his teaching. World leaders, captains of industry, and kings of finance pay a lot for his input on high stakes matters that could cost huge fortunes and even sovereignty.
Excellent lecture by the BOSS. Well done Professor! My Professor in Finland introduced me to your work and wow! Thanks for the lectures. It is our dream as students to be taught by you!
Less is more. Your approach is precisely the sweet spot relative to the internet, the majority attention span and the duration of the pure pleasure of learning that you supply in such a clear rational simplicity of congruent ideas. Nice.
Thank you Mr Taleb, i'm a trader and the lightbulb moment for me (and still is) the law of large numbers, great lecture. Chapter 7 of Mark Douglas's book "Trading in the Zone" where he talked about trading like a casino explained it perfectly about sample sizes. Love the channel.
Professor Taleb this comes with such great timing, I just got the Technical Incerto and this works as an excellent complement class. Thank you so much.
thank you, watched 3 times and then I understood it well, I guess. To approach certainty in FAT TAILED distribution is nearly impossible in real time unlike gaussian distribution. comment to rectify my understanding.
that's because you already know most of the underlying math. its easy for you now but back when you were learning it your professor probably did just as good job as this guy did.
Sir, please make more video Lindy content and make a playlist out of it, We all love you please take UA-cam more seriously, you are a great person we want to know everything from you Love all your work, better mike, better board Thank you for Reading
Thanks teacher Nassim talib. I learned in some situations the bell curve dont apply. That is no. Of observations don’t influence a big deviation. Like the stock market. Or me falling off my bike after 1000 rides! Hahaha
Sir, you are doing a wonderful job. However, please adjust your light as it black board is not clearly visible most of the time. Doing up technical aspect of the video production will help a lot. This is the common observation for all your videos. Warm regards and thanks.
Wonderful content, thanks! But I'm having the same playback problem with all the moocs - plays for 4 - 5 seconds, buffers for 10! Very frustrating. Doesn't happen with other You Tube content. Anyone else getting this difficulty? Perhpas a transcript just before the 'comments' section? Thanks!
@@yington Thanks for the question. On this presentation absolutely nothing. I was referring (in my mind) to other videos that I have watched where he severely criticizes the concept of variance. He prefers the average mean deviation. But it is hard to work with theoretically. The key to unraveling and getting info out of the variance is Tchebyshev's Inequality and one can do some nice analytical work with that. This man has revolutionized the understanding of statistics and warned us about "statistical" assumptions like normality in modelling. Statistics is not understood properly by many practitioners (as in the medical profession) and in fact there is a "Reproducibility Crisis" in science itself. Harvard, Oxford scholars have in fact looked into these matters of great concern. J Froelich Ph.D mathematics
@@WmTyndale Thank you very much! Would you say your point on variance vs MAD applies to the financial markets here? I’ve never heard of Tchebyshev's Inequality, I’ll have to look it up
@@yington The way I see the application of statistics to markets is qualitative like the qualitative theory of differential equations. Numbers that would come out of formulas would just be one of several indicators to assess the condition of a market. But there are many other necessary factors for success, like paying attention to what the Fed is doing with rates etc. To answer your question both MAD and variance will give the same type of qualitative understanding of a distribution in my mind. The abstract idea they are reaching for is "spread" away from the mean. I just use that qualitatively. Volatility is another word that portfolio managers are interested in. My knowledge is limited but I am very interested in Taleb's revolutionary approach to these ideas. Many decades ago I read "Fooled by Randomness". That was an eye opener for this mathematician who had a strong bias against statistical reasoning. Your question gave me a very interesting idea. Can the variance not exist while MAD is finite or vice versa. Probably a simple inequality involving integrals. For finite probability spaces of course both are finite and each can be estimated in terms of the other. However the mathematician Jim Simons used statistics to milk the markets. Lots of computer number crunching but quite secretive in how exactly they used stats to see patterns in the prices. But then there was the famous Long Term Capital Management run by a Nobel in Economics that blew up and had to be bailed out by 14 banks to the tune of 3.6 billion. Prevention of dominoe effect I suppose.
@@yington PS Tchebyshev allows you to estimate the probability that your measurement will be within sigma, 2sigma, 3sigma etc... of the mean in terms of the variance. Like confidence intervals. It makes no assumption on the probability distribution. But you have to have a finite variance and that is exactly the problem. I've got to think some whether MAD can be finite when variance is infinite. Thanks for the question.
Good Taleb example: You can poll a million Mexicans about their individual net worth, but what's the chance you included Carlos Slim in the observations?
Lovely talks on probability. But dude, you do not understand quantum mechanics. It is not the LLN that washes out quantum uncertainty, it is decoherence. Coherent macroscopic systems can exhibit decent superposition. Quantum computers for one would be impossible without superposition in a many-body system. It is the incredibly short time scale over which even a well isolated system decoheres that brings the LLN washing out into play in QM. But the LLN effect acts classically, only "after decoherence" if you like. Before decoherence there really is zero uncertainty, or extreme uncertainty, depending on your point of view: extreme uncertainty if you want to know the final eigenstate after observation, "zero" uncertainty (modulo prior information) if you only care about the wave function.
If people are pricing securities using statistical analysis as if the law of large numbers works and their standard deviations are valid... Then this means they're giving away free money in the market, right?
"10 minutes here is 365 days in university"
- Interstellar (2014)
Never fails to disappoint with sweater covered shoulders
Under 10 minutes is like saving a semester in college. These lessons are great thank you.
Do a lesson on random variables, the most poorly worded quantity in all of statistics since it is neither random, nor is it a variable but rather a function that maps a sample space to the real numbers...it's the source of a lot of confusion for people learning stats
yes please!
Yes!!!!
What rubbish are you saying?
Glad you call me "friend", a big compliment coming from Mr Taleb to a low profile future student of probability and statistics, thank you for yoru videos and contribution, my honor to watch you my friend
“In anything at all, perfection is finally attained not when there is no longer anything to add, but when there is no longer anything to take away.”
-- Antoine de Saint-Exupery
Thanks a lot for perfecting these videos and keeping them under 10min. Truly refreshing.
"We do not have uncertainty, look at the particles here" - Always with the lowkey genius quips.
Not. He does not understand QM. The LLN only washes out uncertainty in quantum mechanics if the system is decoherent. A coherent state can exhibit macroscopic superposition - extreme uncertainty.
Really enjoying these videos. Informative, concise and no BS.
Very fortunate to have access to his teaching. World leaders, captains of industry, and kings of finance pay a lot for his input on high stakes matters that could cost huge fortunes and even sovereignty.
Please if possible, give a real world examples (society wise would be great) on these concepts toward the end of your videos. Great series!
Absolutely loving these basic, intuitive lessons! Please continue churning them out! Much love.
Thank you very much Taleb sir!Just thinking how can someone be so humble and make these videos
Excellent lecture by the BOSS. Well done Professor! My Professor in Finland introduced me to your work and wow! Thanks for the lectures. It is our dream as students to be taught by you!
Love these series! Thank you Nassim
Less is more. Your approach is precisely the sweet spot relative to the internet, the majority attention span and the duration of the pure pleasure of learning that you supply in such a clear rational simplicity of congruent ideas. Nice.
Thank you nassim, finally i can Listen to you and understand what you mean. Little by little 😊👍
Please record and upload this for reference - brilliant!
NNT, truly appreciate you putting out this content. Time permitting, keep 'em coming!
Thank you very much for the lesson professor Taleb.
Thank you Mr Taleb, i'm a trader and the lightbulb moment for me (and still is) the law of large numbers, great lecture. Chapter 7 of Mark Douglas's book "Trading in the Zone" where he talked about trading like a casino explained it perfectly about sample sizes. Love the channel.
Brilliant! Thank you, Taleb!
2nd or 3rd time watching - finally starting to make sense. thanks
😍😍😍😍 Keep it coming, you are a reference in so many levels
Thank you Nassim! Keep them coming please :)
Xristos Anesti Nassim, these introductory videos are excellent, please keep them coming
"You never get information as you're adding random variables". Thanks, very important. #Cauchy !
Good point about Law of Large numbers failing in a fat tail situation. That was an angle I didn't previously consider
Professor Taleb this comes with such great timing, I just got the Technical Incerto and this works as an excellent complement class. Thank you so much.
thank you, watched 3 times and then I understood it well, I guess.
To approach certainty in FAT TAILED distribution is nearly impossible in real time unlike gaussian distribution.
comment to rectify my understanding.
Literally 2 lectures-worth of content explained better than my professor, in 10 minutes.
that's because you already know most of the underlying math. its easy for you now but back when you were learning it your professor probably did just as good job as this guy did.
noise washes out - remember this gem!
This was very cool!
This is great Thanks !!! Mr Taleb, could you do a video on skewness and kurtosis? God Bless
Priceless
Very nice and concise. Thanks
Can we do entropy next?
Thank you for sharing your knowledge Nassim.
Beautifully done
Nice lesson on the convergence to zero of the mean of the distribution of the sampling mean. Make a lesson on the LLN 🙂
Sir, please make more video
Lindy content and make a playlist out of it,
We all love you please take UA-cam more seriously, you are a great person we want to know everything from you
Love all your work, better mike, better board
Thank you for Reading
Wow nice thanks 😌😊
I’m reading statistical consequence of fat tails, very interesting
Thanks Friend !
Thanks teacher Nassim talib. I learned in some situations the bell curve dont apply. That is no. Of observations don’t influence a big deviation. Like the stock market. Or me falling off my bike after 1000 rides! Hahaha
Great lecture! Taleb. The implications of the law of large numbers in scientific observations is grossly neglected! Keep them coming in small bites!😁👍
Concise and informative. As great as ever.
Thank you kindly ✍️
Make a comparison against CLT and the Russian School of Inequalities. That would be another round of bashing the asymptotic n-> infinity....
Of course!
Thanks a lot for those wonderful videos!
By the way, I just borrowed your book Skin in the Game from the public library in St Petersburg, FL.
great video, as always
Epic numbers at the end
Now I want to know what are the other tricks to deal with the 80/20 distribution.
Gracias Nassim!
Keep making more videos.
Loved it
Thank you very much.
How does diversification helps against the low probable out come of a fat tail ? Does it even help ?
Why is there a statistics field in universities!? You are true genius not some of them that won Nobel prize!👌🙏🌷
these are great
Sir, you are doing a wonderful job. However, please adjust your light as it black board is not clearly visible most of the time. Doing up technical aspect of the video production will help a lot. This is the common observation for all your videos. Warm regards and thanks.
Wonderful content, thanks! But I'm having the same playback problem with all the moocs - plays for 4 - 5 seconds, buffers for 10! Very frustrating. Doesn't happen with other You Tube content. Anyone else getting this difficulty?
Perhpas a transcript just before the 'comments' section? Thanks!
Must be your connection.
Love your books, love your thoughts, and happy to see you on YT. Spend the money; get good Audio equipment and a stand for your camera.
Hi Taleb, could you please elaborate the use of Cauchy distribution in finance? Many thanks for your educational videos!
brilliant
Bravo sir. UA-cam > Twitter perhaps?
Can someone explain the 10^13 of the 80/20 distribution?
Yes, I'd like to see the math behind it.
Thank you Nassim Taleb
despite my occasional disagreements he is "the Maestro" of this subject.
What would you disagree on
@@yington Thanks for the question. On this presentation absolutely nothing. I was referring (in my mind) to other videos that I have watched where he severely criticizes the concept of variance. He prefers the average mean deviation. But it is hard to work with theoretically. The key to unraveling and getting info out of the variance is Tchebyshev's Inequality and one can do some nice analytical work with that.
This man has revolutionized the understanding of statistics and warned us about "statistical" assumptions like normality in modelling. Statistics is not understood properly by many practitioners (as in the medical profession) and in fact there is a "Reproducibility Crisis" in science itself. Harvard, Oxford scholars have in fact looked into these matters of great concern.
J Froelich Ph.D mathematics
@@WmTyndale
Thank you very much!
Would you say your point on variance vs MAD applies to the financial markets here? I’ve never heard of Tchebyshev's Inequality, I’ll have to look it up
@@yington The way I see the application of statistics to markets is qualitative like the qualitative theory of differential equations. Numbers that would come out of formulas would just be one of several indicators to assess the condition of a market. But there are many other necessary factors for success, like paying attention to what the Fed is doing with rates etc. To answer your question both MAD and variance will give the same type of qualitative understanding of a distribution in my mind. The abstract idea they are reaching for is "spread" away from the mean. I just use that qualitatively. Volatility is another word that portfolio managers are interested in. My knowledge is limited but I am very interested in Taleb's revolutionary approach to these ideas. Many decades ago I read "Fooled by Randomness". That was an eye opener for this mathematician who had a strong bias against statistical reasoning.
Your question gave me a very interesting idea. Can the variance not exist while MAD is finite or vice versa. Probably a simple inequality involving integrals. For finite probability spaces of course both are finite and each can be estimated in terms of the other.
However the mathematician Jim Simons used statistics to milk the markets. Lots of computer number crunching but quite secretive in how exactly they used stats to see patterns in the prices. But then there was the famous Long Term Capital Management run by a Nobel in Economics that blew up and had to be bailed out by 14 banks to the tune of 3.6 billion. Prevention of dominoe effect I suppose.
@@yington PS Tchebyshev allows you to estimate the probability that your measurement will be within sigma, 2sigma, 3sigma etc... of the mean in terms of the variance. Like confidence intervals. It makes no assumption on the probability distribution. But you have to have a finite variance and that is exactly the problem. I've got to think some whether MAD can be finite when variance is infinite. Thanks for the question.
Is the last calculation featured in his Fat Tails book?
10^13! Holy hell
merci
Thanks. I’m a novice, but this is very interesting.
exactly 0 people on the street ever discussed a pareto distribution with me
Can you habilite the subtitles? non native english speakers will appreciate it
very good video. Nonetheless could you record with your phone rather your laptop because it’s much more fluid and neat with your phone. Thanks
How can I self-learn statistics?
What is that thing he's wearing? Not being facetious, I'm curious as to what that is
It's a bluetooth microphone/earphone.
❤️❤️❤️
Can someone confirm that around minute 7:27 he says "Cauchy" distribution? If not, which type of distribution did he mention?
Thank you!
That's what I thought
Yes that's correct. He goes on to explain the properties of the Cauchy shortly after.
@@EpicGamer659 thank you.
He should consider doing math professionally
спасибо)
Some of the syllables are elusive in the audio.
Can i Use Kappa with a given data series to determine if it is cauchy or gaussian distributed?
Prof., did you use mathematica to extract 10^13 for the Mean Deviation's drop ? Can you share your code or your approach to get that value ? Thanks
I was going to mention the variation in webcam quality in this series....but thought better of it.
Good Taleb example: You can poll a million Mexicans about their individual net worth, but what's the chance you included Carlos Slim in the observations?
I am skeptical of bootstrapping techniques because of this. I suppose I do not understand them deeply
Somebody get that man a SmartBoard! I am interested to see why the Pareto distribution requires so many more observations.
It's in SCFT the free book
What headphones were you using for this Nassim?
Hey
Does anyone here has a good reference book/site/course on Cauchy Distribution or Fat Tails to get a better understanding on the subject?
Thanks
@Dan Campbell Thank you
Did you come up for this function to find kappa in order to price or attempt to price equities priced at market incorrectly via gaussian models
That's very cool
Not at all
Can someone explain how standard deviation was sqt(p(1-p)) @3:25
Ok understood bc it's a binomial distribution
Lovely talks on probability. But dude, you do not understand quantum mechanics. It is not the LLN that washes out quantum uncertainty, it is decoherence. Coherent macroscopic systems can exhibit decent superposition. Quantum computers for one would be impossible without superposition in a many-body system. It is the incredibly short time scale over which even a well isolated system decoheres that brings the LLN washing out into play in QM. But the LLN effect acts classically, only "after decoherence" if you like. Before decoherence there really is zero uncertainty, or extreme uncertainty, depending on your point of view: extreme uncertainty if you want to know the final eigenstate after observation, "zero" uncertainty (modulo prior information) if you only care about the wave function.
OK
Tell us what you think. Don't show your portfolio.
Holy crap.
If people are pricing securities using statistical analysis as if the law of large numbers works and their standard deviations are valid... Then this means they're giving away free money in the market, right?
10^13.....
You’re a billionaire: invest some money and get a proper microphone and camera.
Your audience will be grateful. ;)
Obviously these lectures are not for you. You should look elsewhere.
Yes, under 12 minutes. University lectures are said to lose the listener after 12 minutes.
Merci