Bellissimo sentire parlare il Signor Taleb in Italiano

So, here is how I would answer this, The first part of the equation is concerned with the cash flow of the Bitcoin network. While we can argue on this, let's agree that it is in fact, 0. I should also add at this point, that I am talking about the Bitcoin network, denoted by Bitcoin with a capitalized B, while the underlying asset, which is a unit of account on the Bitcoin network, is also called bitcoin and I will write it with a lower case b. Now let's move to the second part of the equation, which is concerned with what amount a rational actor at a time (t+1) is willing to pay. In this case, we are talking about the asset bitcoin which trades on the Bitcoin network. So, a rational holder of bitcoin will have to calculate the second part of the equation to know the price of bitcoin at a given date, since the first part of the equation is 0. Before we do this calculation, I would like to post a bit of wisdom from NNT's books, of which I am a huge fan. If a given model predicts the price of bitcoin to be 0, meanwhile bitcoin continues to trade for a non 0 price, the more rational thing to do is to explain why, for 16 years bitcoin has continued to trade for a non 0 price, rather than to continue being loyal to the model and simply saying something like 'markets are just irrational'. Now getting back to the topic at hand, the second part of the model which deals with the expected price at a future date, is the price which rational (supposedly) actors give to bitcoin. This is a function of many things, and some of them are based on the information these actors have at the given time t+1. This information also may contain noise, and perhaps even false signals. So in order for a rational actor to make a rational choice, we must assume here that he/she has the ability to ignore the false signals and ignore the noise as much as possible and focus only on the signal. Now I can post many reasons here as to why bitcoin is valuable and it's valuation by a rational actor should be non zero, but that is irrelevant to this discussion, so I will focus primarily on the available data. The available data suggests that supposedly rational actors have valued bitcoin at a non 0 valuation for around 16 years now, so I must therefore explain how that may be possible for something with a 0 cash flow, without resorting to rhetoric such as 'markets are too irrational'. This, I believe can only be done by admitting that the rational actors have information which I do not have access to (especially if my model predicts a bitcoin price of 0) I must also add here that every now and then the price of bitcoin gets into bubble territory, at which point, markets are indeed irrational. However, taking a 4 year moving average of bitcoin can eliminate this noise and the signal is clear. The price of bitcoin is trending up with time. The only explanation (without speculating as to what it is) is that more and more actors are participating in the Bitcoin network and have access to this aforementioned information, based on which they are making the rational decision about what the price of bitcoin should be.

Gotta bring it back to the tulip bulb mania. I remember asking my dad why he didn’t invest in crypto when I was 7-11yo and at 12yo I started telling him to invest in bitcoin, not because I thought it had intrinsic value, which I told him (this was 8 years ago), I told him I thought it was a volatile market that was most likely going to go up because it’s perceived value is given by the hopes and dreams of people, I didn’t know all the terminology at 12 but kept developing. Anyways my dad never invested even though I kept insisting, some may say he is responsible for not listening to a 12-18yo but he wouldn’t even have invested more than 1-3k, he was broke so it’s whatever. Bitcoin has absolutely zero value, what Mr taleb says is true according to the rational market but this is highly irrational because it is a mania. Worst of all, people comparing a crypto coin that has zero value over a work of art. It’s a commodity, a fool selling to another fool. I still would like to see bitcoin continue to succeed though, I hope it reaches a high valuation because that will help promote the rest of the crypto currencies, specifically and hopefully the ones that have value. I view satoshi nakamoto as an accelerationist and I appreciate them for their work in that aspect (unintentional or not).

Bitcoin hit $100K tonight. This presentation looks like a midwit who doesn't comprehend that it is a new asset class who is trying to contain it in a box of his own preconceived notions about money. But what do I know, I'm just a polisci major with some worthless Bitcoin.

How can we purchase a license ?

Unironically I came back to this video because there may no longer be the suppressing of real rates if FED buying of treasuries is limited and this could kill bitcoin inadvertently 😂

I am modeling this for my product right now. I want access so bad. haha. Thx for sharing!!

Can someone clarify why there is no mean or variance in some condition of power law distribution? Is it not the case that you can always count the mean or variance of the sample (just by applying the formula to the data)? So, then what Mr. Taleb mean is that the REAL-WORLD data of something that come from certain power law distribution have a (relatively high) probability of containing a veryyyyyy large number (infinite/many times away from standard deviation) that the mean/variance of sample is very unstable and never converge as in Gaussian. Does this interpretation correct?

lol. This is sooooooo good. so much better because I am high too.

Signore Porfesor Docktore Taleb, what a wonderful resource to have you oersonally teach us, the youtube crowd, this concepts. Now we can all be direct students of the technical incerto. Thank you so much.

Yes, we all know the Gaussian distribution doesn't describe outcomes we observe in equity markets. That said, what practical use does this video/software have, if any, for the self-directed retail investor - is this just more marketing material from Universa?

haha

Bonjour cousin ça va bon courage

even better: "The criticism stated by Dr. Taleb involving "fat tails", "via negativa not via positiva" and being a "concave" measure is a series of non-issues and expresses profound ignorance on his behalf. A whole other lengthy thread might be initiated on this alone. The fact that many of the so-called "real-world" performance indexes do not usually show a Gaussian distribution (indeed, most often one finds a Pareto or similar distribution) does NOT imply that the association between a normally distributed IQ (or any Gaussian variable) either "doesn't exist" or "is uninformational". At worst, it just means that such associations are better assessed through nonparametric techniques. The same reasoning goes for nonlinear associations, which can be analyzed through nonlinear methods. It is ludicrous to suggest that a nonlinear association between IQ and the SAT is in any way indicative of the uselessness, inadequacy or fallacy of the first. One must also observe that non-Gaussian distributions can frequently be "Gaussianized" through simple mathematical transformations such as taking a natural logarithm or, with a bit more complexity, a Box-Cox transformation, among other methods. Nonlinear association can also be linearized through usually simple transformations. For example, the graph cited from Frey and Detterman (2004) can be easily turned into a strong linear association if one uses Ln IQ instead of "raw" IQ scores. And since when do nonlinear associations imply in pseudoscience?"

"The observation that IQ explains “only" 13% to 50% of the variance in some tasks merely reflects the fact that performance requires more than just intelligence. Yes, such things include Conscientiousness (which is related to impulse control or the ability to defer satisfaction, i.e., "patience"), but also values, personality, and the way in which all these things relate to sociocultural settings and even physical environment. Actually, one should suspect a "quack" when someone in human or social sciences claims that a single variable alone explains most of the variance of anything. Indeed, the most advanced multivariate statistics in use today were created by psychologists and social scientists in order to deal effectively with such complex problems in their field."

Sebastian Trail

Thank you, Professor Taleb. You are teaching the world how to think with courage. For those struggling with the Mathematica code, there are two functions in there. The first one is solK, which finds the quantile of the Pareto with scale value L and shape parameter α (alpha) such that we have a probability p of finding a value greater than it. For instance, later on in the example mk/. alpha → 1.15 /. p → .2 //N the second function (mk, see below) is called, and from within that second function, this solK function is used in /. solK[[1]] to generate the K value that leaves 20 percent (p→ .2) to the right (higher). That K value is not shown at any points, but it can be calculated by solK /. \[Alpha]-→1.15 /.p →.2//N (it happens to be K → 406.159 L). Notice that the L parameter is left in the solution as a variable. This K remains within the function (local variable), and is used to calculate the output of mk, which is a ratio: The second function, mk, is basically a ratio: the numerator is the expected value of Pareto for values greater than K (from K to infinity), calculated as the integral of x * PDF(x) using the K output by solK. The PDF of the Pareto is alpha * L^alpha / x^(alpha +1). The denominator is the expected value of the PDF, which has different limits of integration, but it has essentially the same integral of x * PDF(x). This allows the L factor to be taken out of the integral on both numerator and denominator, and cancelled. In the end, the ratio measures the tail beyond K against the mean. It essentially represents the conditional expectation of the Pareto distribution beyond the threshold K, normalized by the overall mean of the distribution. ON THE CANCELLATION OF K: Remember that solK's output is a function of L. Now the numerator of the ratio is K * L^alpha * alpha / k^alpha * (alpha - 1). The denominator is the mean of the Pareto (see Wikipedia), which is L * alpha / (alpha - 1). After doing a bunch of high school cancellations, you are left in the ratio of these two with the close expression (L / K)^(alpha - 1). And remembering that K is the output of solK, and is a linear function of L, we finally get L cancelled inside the parenthesis.

It is a lonely journey on the path of truth. But history will prove you right

Thank you both.

Just to mention in case NNtaleb reads the comments - using inv prob / erfc is what is done in Probit models. Logistic transform is done in Logit models. Usually Logit is used because calculation tends to be easier. Just FYI - econometricians have a framework for studying them (but typically without the dynamics which you speak about here).

Lee Estates

Great show guys. You guys are a true inspiration. Can we bring back the Quant finance lectures please

China and India still have more people today than any country around the world Mother f-cker!

542 Treutel Summit

90247 Luis Streets

16220 Larson Path

48357 Kilback Overpass

Thank you both. This looks interesting. We should meet at my local, Mulligans in Manchester, UK. Minimal downside, upside is infinite - depending on the guitarist.

51898 Gerhard Underpass

To catch fast-moving golden fish, even if it costs money, I buy a very wide net and spread it out. Is my understanding roughly correct?

..But how do normal people invest in universa?

833 Clair Streets

9916 Medhurst Extension

2:58 the turkey's statistical department LOLL

I am a very very amateur mathematician, so take my comment as a mark of curiosity more than anything. In the example you gave of {0, 0, 0, ... 10^6}, isn't the standard deviation a better "estimate" of the deviation? Even though neither STD or MAD really makes sense as a descriptor of the set, isn't 1000 a somewhat more accurate estimate of the deviation than 2, precisely because it weights large deviations more?

I'm gonna use what delta my broker provides and Black scholes. Too much work to switch systems at my age... If i can was a younger/wealthier man i would research this stuff.

aistockadvisor AI fixes this (AI Stock Advisor). "Empirical distribution hides real tails."

Thanks for the share, looking forward to seeing Universa institute in the future!

Man the hardest thing about this is actually managing a tail risk hedging. I've listened to just about everything Mark and Nassim have put out on the subject, over and over the refrain is "everyone is doing it wrong we are doing it right, but don't try and do it" 😢

1942 Andrew Heights

52050 Elody Junctions

Brandon is good-looking!

Taleb seems very chill here.

At 1:03 the look that Taleb gave....lol

Haha also korean pronunce is "베르-누-이[yi]"

I think the Lucretius comment was from Antifragile, no?

I missed you, Nassim. Thanks for the sanity

I would like to work for Universa, but no answer when sent my CV. I hope you don't think I support fragilistas hahaha

タレブ！！シェアしてくれてありがとう！！愛してる！！

Amazing!🙂👌