I can really recommend that second volume book of Shreve. That is where I learnt from when I was new to the subject. It explains things cleary, but don't go into allt the more technical and subtle details to prove everything rigourosly.. This is a good start to understand what is happening. After that I recommend a more detailed book. Shreves other book together with Karatzas "Brownian motion and stochastic calculus" is one such option. This is not a light read, but also a good book.
Excellent Lecture. In 21:30, we know Ln(St) is Normal. How about St itself, is it Normal too? Not sure what the distribution of e^X is when X is Normal.
Negative stock prices can be avoided by using exp(brownian motion) instead. Still, I don't think a normal distribution is a good fit, since it severely underestimates the probability of bigger price movements; so called "black swan events", which happen all the time. A distribution with fatter tails is more appropriate. I think the pareto distribution probably has a too fat tail for most stocks though - so a semi-fat tail distribution is likely the best fit for modeling most assets.
[note:I don’t know what I’m talking about when it comes to finance.] What about the Martingale assumption? Would that being violated contradict the Efficient Market Hypothesis or something? Or... Well, I guess momentum trading working (iirc?) sorta maybe implies violating the martingale assumption? Hmm... now I’m wondering how one would measure how well a model is doing. I think maybe you could look at like, the surprisal of the observations under the model, minus the entropy of the model? Or, uh, Hm, well, if things at separate intervals are independent (or across different stocks? Or something that would let you take an empirical average), you could maybe take a KL divergence? Of course, the real test in practice would be “does it make you money”, but pretending that wasn’t the goal and you just wanted an “accurate” model for its own sake, how would you measure its “accuracy” (when it is all stochastic like this)
This is actually true, and priced in the market aswell I highly recommend the video "the volatility smile" by professor patrick boyle here on youtube, he talks about how this is dealt with as the volatility is actually changing all the time
@@drdca8263 Martingale is in most instance a game you can't play long enough to make back your investment to play the game twice before losing all in the exponential increase of risk. Could be applicable to wider game theory/dynamics like e.g. arbitrage systems.
Great video! I do have a question. In the first 30ish seconds you mention that stocks can’t go negative. Why is this so? WTI crude went negative in the year 2020.
Crude oil are futures and not stocks. The residual claim right of stocks on earnings just stops at $0 because the company is limited and does not need to come up for additional capital if it files for bankruptcy because it is legally limited. For the future crude oil however the future contracts expired, which means the oil barrels are delivered and because of the storage problem at the time, you did not pay for the oil, but you where paid to store the oil. As you can see very different issues, and barely comparable.
Unlike the equity market, commodity markets have quite inelastic demand. The extreme example of this is in the power market where instantaneous power cannot (on a large scale) be stored. In commodity markets it is essential to have negative price signals. As in the comment above, companies are entities that have limited liabilities.
An advise when you teach math: To avoid making your equality signs to appear as a slight of hand, try distinguishing more rigorously asymptotic equalities by actual equilities. Otherwise I like the scope and how you motivated your topic.
Thank you for the fantastic video. Can you please upload a video on Partial averaging in conditional expectation using fairly basic language. Thank you once again for your videos.
Nice lecture, but in the intro for Wiener stochastic processes, though I'm not a professional quant, I always thought that the stochastic variable was the log of the asset value rather than the asset value itself. This is what it says in Wikipedia too, and using the log has two effects (1) the variable can't go negative and (2) the statistics look similar whatever the lever of the asset, whether its $1 or $1000.
Thanks John, this lecture starts with the Wiener process and progresses to the Geometric Brownian Process where assets prices have the log of asset price as you’ve mentioned
Just my opinion neither, financial math would be a good choice 👍 CFA & MBA are programs not specific to quantitative finance in terms of derivative pricing theory ect. Great if you want a job as an analyst in the financial industry though.
@@QuantPy The next year I’ll finished my bach in Financial Engineering and now I'm looking for the CFA or Quantitative Finance MSc in Glasgow. Which MSc could be better for the Quant path? Or in your opinion, what could be the next step?
@@Gerard91999 Don't worry about CFA or finance at all. If you are serious about Quantitative Finance then do MSc either in Quantitative Finance or Mathematics (better if focused in PDE or probability theory) and do lot of good quality computer science projects either in Python, C++ or JAVA. With good skills in Computer Science and Math you could easily crack it.
Neither. But the CFA is cheap compared to the MBA, so if you can, take that. For a quant career, the best is a PHD, where it doesn't matter what you choose to study. Whether phyisics, mathematics or computer science. It depends on your focus and your research interests.
Thank you for the fantastic video. Can anyone please help me understand one basic concept. Why do we have break the [0,t] into smaller intervals. What do we miss by not doing so? Does it eliminate any unnecessary /overlapping factors? Thank you
Try out Stochastic Calculus by Calin. It should be simple for you based on what you say your knowledge is, but it should prepare you for further study.
![BLACK BAG - Official Trailer [HD] - Only in Theaters March 14](http://i.ytimg.com/vi/Du0Xp8WX_7I/mqdefault.jpg)
![Blox Fruits Dragon Rework Update [Full Stream]](http://i.ytimg.com/vi/EqDAp8udhm0/mqdefault.jpg)
Very clear explanation. This helped me in my Advanced Statistical Mechanics course I'm doing in my Masters in Physics!
Genuinely in love with this channel, keep up the good work!
15:20 -> This second order term actually adds to the drift term, OVER TIME. Exactly, this is the essence of the case. Well depicted!
have been in love with quant for many years, but man GBM is the cure for insomnia. God help me :) ... Awesome video btw, thank you!
what does GBM stand for ?
Geometric brownian motion
This is really a great channel. Wish I knew about it earlier.
better than my professor, good job man
Best video ive seen so far...
Really love this sorted of content
Nice lecture. The explanation is quite clear and informative.
Great channel !!! Thanks for the work
I can really recommend that second volume book of Shreve.
That is where I learnt from when I was new to the subject. It explains things cleary, but don't go into allt the more technical and subtle details to prove everything rigourosly..
This is a good start to understand what is happening. After that I recommend a more detailed book.
Shreves other book together with Karatzas "Brownian motion and stochastic calculus" is one such option. This is not a light read, but also a good book.
Fully agreed, Martin Baxter's Financial Calculus book is also good for developing an intuitive grasp on these concepts.
Brilliant Lesson Thank you so much
Excellent Lecture. In 21:30, we know Ln(St) is Normal. How about St itself, is it Normal too? Not sure what the distribution of e^X is when X is Normal.
Negative stock prices can be avoided by using exp(brownian motion) instead. Still, I don't think a normal distribution is a good fit, since it severely underestimates the probability of bigger price movements; so called "black swan events", which happen all the time. A distribution with fatter tails is more appropriate. I think the pareto distribution probably has a too fat tail for most stocks though - so a semi-fat tail distribution is likely the best fit for modeling most assets.
Also the interpretation changes when we use an exp(Brownian motion) doesn't it? Which would complicate things.
[note:I don’t know what I’m talking about when it comes to finance.]
What about the Martingale assumption? Would that being violated contradict the Efficient Market Hypothesis or something? Or...
Well, I guess momentum trading working (iirc?) sorta maybe implies violating the martingale assumption?
Hmm... now I’m wondering how one would measure how well a model is doing. I think maybe you could look at like, the surprisal of the observations under the model, minus the entropy of the model? Or, uh,
Hm, well, if things at separate intervals are independent (or across different stocks? Or something that would let you take an empirical average), you could maybe take a KL divergence?
Of course, the real test in practice would be “does it make you money”, but pretending that wasn’t the goal and you just wanted an “accurate” model for its own sake, how would you measure its “accuracy” (when it is all stochastic like this)
This is actually true, and priced in the market aswell I highly recommend the video "the volatility smile" by professor patrick boyle here on youtube, he talks about how this is dealt with as the volatility is actually changing all the time
@@drdca8263 Martingale is in most instance a game you can't play long enough to make back your investment to play the game twice before losing all in the exponential increase of risk. Could be applicable to wider game theory/dynamics like e.g. arbitrage systems.
idea. can you consider doing a video on cases where the underlying is interest rate product, such as an interest rate swap?
Great video with so simple language.
Pretty good strategy for every beginner trader
Great video! I do have a question. In the first 30ish seconds you mention that stocks can’t go negative. Why is this so? WTI crude went negative in the year 2020.
Crude oil are futures and not stocks. The residual claim right of stocks on earnings just stops at $0 because the company is limited and does not need to come up for additional capital if it files for bankruptcy because it is legally limited. For the future crude oil however the future contracts expired, which means the oil barrels are delivered and because of the storage problem at the time, you did not pay for the oil, but you where paid to store the oil. As you can see very different issues, and barely comparable.
Unlike the equity market, commodity markets have quite inelastic demand. The extreme example of this is in the power market where instantaneous power cannot (on a large scale) be stored. In commodity markets it is essential to have negative price signals. As in the comment above, companies are entities that have limited liabilities.
Great video! What is the context behind the mean square limit....why do we need that particular concept for the stochastic integral? Thanks!
I have no idea what you’re talking about since I’m not this advanced, but it’s still cool.
What would you say I should know before learning ito calc?
Don't waste your time if you dont understand what he is talking about, there is an abundance of quants for which this is quite basic
An advise when you teach math: To avoid making your equality signs to appear as a slight of hand, try distinguishing more rigorously asymptotic equalities by actual equilities. Otherwise I like the scope and how you motivated your topic.
Thank you for the fantastic video. Can you please upload a video on Partial averaging in conditional expectation using fairly basic language.
Thank you once again for your videos.
Nice lecture, but in the intro for Wiener stochastic processes, though I'm not a professional quant, I always thought that the stochastic variable was the log of the asset value rather than the asset value itself. This is what it says in Wikipedia too, and using the log has two effects (1) the variable can't go negative and (2) the statistics look similar whatever the lever of the asset, whether its $1 or $1000.
Thanks John, this lecture starts with the Wiener process and progresses to the Geometric Brownian Process where assets prices have the log of asset price as you’ve mentioned
What a brilliant explanation. Well done.
keep up this type of content
For quantitative finance which could be better? CFA or MBa?
Just my opinion neither, financial math would be a good choice 👍
CFA & MBA are programs not specific to quantitative finance in terms of derivative pricing theory ect. Great if you want a job as an analyst in the financial industry though.
@@QuantPy The next year I’ll finished my bach in Financial Engineering and now I'm looking for the CFA or Quantitative Finance MSc in Glasgow.
Which MSc could be better for the Quant path? Or in your opinion, what could be the next step?
@@Gerard91999 Don't worry about CFA or finance at all. If you are serious about Quantitative Finance then do MSc either in Quantitative Finance or Mathematics (better if focused in PDE or probability theory) and do lot of good quality computer science projects either in Python, C++ or JAVA. With good skills in Computer Science and Math you could easily crack it.
@@siddharthchaudhary1266 thank you m8!!!
Neither. But the CFA is cheap compared to the MBA, so if you can, take that. For a quant career, the best is a PHD, where it doesn't matter what you choose to study. Whether phyisics, mathematics or computer science. It depends on your focus and your research interests.
Thank you for the fantastic video. Can anyone please help me understand one basic concept. Why do we have break the [0,t] into smaller intervals.
What do we miss by not doing so?
Does it eliminate any unnecessary /overlapping factors?
Thank you
That is why it is called infinitesimal...
I fucking love this channel
Can I read volume 2 without volume 1? I have a high energy theoretical physics background.
Is there a software to do this? Or this is just theory stuff for intuition?
So why is it an adaptive process into the filtration? Why can’t it be static?
u can be perfectly a teacher xd
Guday to you😉
13:13 should be k>2 not k>3
true, cause Dw*Dw=Dt
I am an undergrad and taking intro to stochastic modelling class, how can I go deeper to understand these topics, what should I begin with?
I suggest buying Stochastic Calculus for Finance II by Steven Shreve
Try out Stochastic Calculus by Calin. It should be simple for you based on what you say your knowledge is, but it should prepare you for further study.
watching this for High School.... everyone else here is in college and some are doing masters...... IBDP is HELL
What was your undergraduate degree in?
Chemical Engineering
@@QuantPy did you get a masters? How was the transition from chemE to something more math heavy?
@@QuantPy Hi, I´m a chemical engineer too. Bur I´ve got interested in finance. Good to know that I can follow your path
What are you studying?
@@QuantPyim currently doing my undergrad in chemical engineering. How did you get into the quant space from this, I want to follow suit 🥺
سلام عليكم
شرحك رائع
هل يمكنني السؤال؟
References a textbook, but doesn’t show it. Nice.
Wienner and Riemann are pronouced with short "i" like prison!
A bit too abstracted
Pessima traduzione dei sottotitoli
that's a cute girl on the video cover
Your ito calculus is overrated…
Dude,,, calculus is boaring part of mathematics 😏 😏
Too bad it's the largest part of Mathematics in Quantitative Finance :p