Kelly Criterion Trading Strategy : Used by Buffett, Munger, Pabrai
Вставка
- Опубліковано 12 вер 2024
- Interactive Brokers: My Main Brokerage Platform
CLICK HERE -
www.interactiv...
Undervalued - My monthly Commentary and up to date watchlist information. Check out the February edition for free
undervalued.gh...
Inaugural Money Managers Game 2022
Join the Discord Group for instructions - / discord
TIKR: Financial Data on All Companies plus Super investor portfolios plus an International screener.
www.tikr.com/an...
HAMISH HODDER’S AMAZING COURSE
70+ Hours of Stock Reports
Private Investor Community with Hamish
Judgement Day BONUS Module
Use discount code ‘BROWN’ at checkout for 30% OFF
More Information - www.hamishhodd...
Andrew Brown’s Intrinsic Value Calculator
mailchi.mp/02b...
My other trading platform that I use. Reliable and regulated trading platform.
SAXO BANK LINK: www.home.saxo/...
Twitter:
@abrowninvesting
FYI Some people refer to the Kelly formula as the Kelly Criterion, but it’s the same thing.
Warren Buffet, Charlie Munger and Mohnish Pabrai use the Kelly Criterion as a safe way to guide their decision into the size of the investment they should make. So if it’s good enough for the best investors, it is good enough for us.
Kelly Percentage = probability of winning - (1 - probability of winning)/ winnings as a percent
So let’s enter the information for this game.
K = 0.5 - (1 - 0.5) / 1.5
= 0.167 or 17% of your available balance.
So in this game, it would be wise for your first bet to wager 17 cents.
Because if you lose, you still have bankroll to try again.
Now something that is VERY VERY VERY important. Is this Kelly Formula only works when you have a profitable opportunity.
But our coin flip game gives you 50/50 odds and the win is bigger than the loss. So it’s a profitable game.
So how do we use it for the stock market? Or any investment?
So I’m going to go through the example of Global Cord Blood Corp, which is a Chinese company I have been interested in for some time now.
It’s a good example as the risks are higher than a lot of other stocks.
Right now there are 156 Chinese based companies listed in the US.
So if 1 in 8 companies have fraudulent activities, that would be 12.5%.
But let’s presume it’s higher than that and the fraudulent numbers are at 2010 levels of 35%.
Let’s also assume, that these companies are so fraudulent, that their stock price should be 0.
Even though a lot of fraudulent companies have different degrees of fraud, some minor some major. But let’s assume if the company is fraudulent, we get wiped out.
So far we have our win probability at 65% and our loss probability at 35%.
I think it’s better to be conservative here and say there is a 20% chance of something like this happening.
So now our win percentage is down to 45%, and our total loss probability is 55%.
Now what do we win if the 45% probability event takes place?
But again, let’s be conservative, so I think at a minimum, the win would be that the company reaches the level of the Take over offer on the table. Which is $7.50 per share.
The current share price is $3, so to get to $7.50, it would be a 150% increase.
So let’s put all of this into the Kelly Formula and see what happens.
Kelly Percentage = probability of winning - (1 - probability of winning)/ winnings as a percent
K = 0.45 - (1 - 0.45)/1.5
K = 0.083 or 8%
Let me summarise by saying this.
No investing system is perfect. The Kelly formula will help you allocate funds to opportunities effectively, but it will not pick winning stocks for you.
But when a great opportunity comes along, The Kelly formula will help you take advantage of it. If the odds are in your favor, and the rewards are really high, well you should be aggressive. The Kelly formula will definitely help you be more decisive.
Sometimes you need to go big, but just not so big that you get wiped out and can’t recover.
If you enjoyed the Kelly formula, don’t forget to subscribe and I’ll see you in the next video.
Fantastic! Thank you for pointing me to the Kelly Criterion, and explaining it!
Thanks :)
Okay, last comment now I promise. For anyone interested. Unless I am mistaken (in which case I would greatly appreciate being corrected) you could figure out the odds of independent events occurring using a Venn diagram or similar (that's just what I used to make it easier) and give each event a probability and a return/loss. You could then group them together into 2 groups of winning scenarios and losing scenarios. Using the formula in the link of my other comment (Where there is not an absolute loss assumed) you could calculate a far more accurate and flexible allocation percentage.
This would make sense to do as a company would be made of components working often independent of each other. Just make sure not to compound the gains and losses unless you purposely want to do that as it would just make no sense. If one was dedicated enough they could even figure out how these independent events may increase or decrease profits beyond their individual sums (eg a product that competes slightly with another new product of the company)
Great comment, appreciate the effort a lot
One of the best videos on this topic! Thanks 😊
Thanks, appreciate the kind words
good explanations! thank you. i wish i could see another kelly criterion applied during the covid-19 in US equity market
glad it helped :)
I dont mean to be offtopic but does anyone know of a way to log back into an instagram account..?
I stupidly lost my login password. I would love any assistance you can offer me!
Maths overload! I feel so smart now :)
Sorry, you probably get a lot of maths in your daily life....
The best tuition of investment! Thanks, красавчик 😁
Большое спасибо 🤘, первый комментарий 🥳
Thanks for the video. Very clear. I have one question: how are you calculating the odds of a company being profitable and compute that into a number that you can plug into the formula?
You gave an example of a weighted coin. How does the coin being weighted change the odds to specifically 0.51 and not, say, 0.53?
That's a great question and practically impossible to get exactly right in stocks. It is more of a estimate. For example, the odds of bank of america going to 0 would be less likely than some start up. Hope that makes sense. Ball park figure is the aim
I love the unemotional structure to your methods!
Hey mate, thanks, appreciate the feedback. I'll get back into some company analysis soon :)
complicated the formula simply 2p-1=x can be used where p is the probability of winning and x is percentage of investment of your net worth/portfolio
Thank you very much. Subscribed & will watch more of your videos
Welcome aboard!
@Andrew brown thank you for that! But how do you calculate the P and B in the first place for an investment? Also- Does this require additional sensitivity analysis?
Great point, it is very difficult and more guess work than accurate maths. Honestly, this is only helpful if you have many opportunities to invest. But most people won't have 100's of things to invest in, so this isn't the best strategy. I have learnt this since making the video.
You can’t use what’s in this video in real life. To estimate these parameters, it’s hard. There are companies just to estimate these, even they aren’t that accurate even you built some complicated models . So in modern portfolio theory, you have an expected return and a covariance instead of the probability of the return (P & B) and the distribution (in %), hence you can get one of the parameter (expected return in stock) and say you magically get the second one. Even just some difference in your parameters make your bet size changes, which will result in changes in your return! But we can take a step backward and forget the math, if you somehow trust your company’s estimate on the return, why don’t you yeet yourself and do leverage onto the position?
in conclusion, 1) you can’t estimate these. 2) even if you think you have it, the strategy isn’t robust, small difference in the estimate will result in big difference in your return. 3) this thinking doesn’t take covariance into account, the buy/sell time isn’t well defined, and most importantly it isn’t the same bet over time as the market change overtime.
This video assumes the parameters are trusted and true, but stock market isn’t like that. Same way university level investment finance class teaches you about portfolio theory, it’s more like a mental framework that helps you think about these stuff and you can’t really implement it unless you dive really deep into it
I like this one. Very technical to me, but explained so well I can follow along.
Thank you for your comment. Glad you liked the video, I was trying to explain it the best I could because it is a bit technical.
Thank you Andrew!! Great stuff!!
Thanks, glad you liked it
Great explanation! Thanks!
Thanks, glad you enjoyed it
thank you Mr. Andrew it is very useful
Thanks, appreciate the kind words
What’s 92% divided by 2? It is not 41%.
so what if returns are not from binomial distribution ? such that tail risk is heavy?
I think you may like SSL if you havent looked into it yet. Its the second largest company in South Africa. They had a huge mishap on an giant investment amplified by negative oil prices. They look to be recovering. In any case, a company SA sees as too big to fail at 8 dollars a share is enticing.
Hi, appreciate the comment and suggestion. I haven't looked into SSL, or any South Africa company for now but I'll add it to my list.
this is great! well explained;)!
Thanks appreciate the feedback. :) Happy investing
Didn't pabrai disown method of kell criterion for investing, isn't it just for betting when you are putting money several times.
Nice video dude!
Thanks a lot 😊
with R:R 2 to 1..and winning rate 50%..risking 25% still make you a lot..more on trade expectancy
Why did you use 1.5 as denominator when referring to a 150% final possible result in the coin example but used 0.15 when referring to a final balance of 115% for the Berkshire example, shouldn´t it have been 1.15?
Thanks!
great report
Thanks a lot
Excellent
binging these
Thanks mate
It still works when the odds are against you. You just have to flip your perspective. If you take it from the "house's" perspective, they will be at greatest risk of loss, despite their odds advantage, when you put all your money on a single bet. This is because it is their most likely way to go bankrupt which is your most likely way to win.
So basically, if you know the odds are against you, just go all in!
not sure it works like that... haha
Issue with Kelly in stocks is you don't know the payout or probability. Makes it almost useless.
great example is Birkshire returning 15% for the next 50 years @ 90%. Holy fk. maybe 25% and thats high.
Surely this would break down when looking at for example treasury bonds. Even at crazy low interest rates, the formula would probably tell you to go near all in due to the extremely low risk. Anyone know a modified formula that could account for this?
So you are right, but only if you can make the same bet over and over again therefore it will compound quickly. I actually think the kelly formula isn't a good way to think about investment size. Because it assumes unlimited bets. And we all know that's not possible in most investing. It works at the casino.
If you have a trend following system that is profitable 54% of the time with a mean profit of 19% per trade and median profit of 9% per trade, with a mean loss of 4.5% and median loss of 3.9%.
How will this be calculated?
I'm getting strange numbers:
.54- (0.46/.09)=-4.57
Should I be looking at cumulative return over 1-2 years instead?
Or is this saying it's a bad system 😃
I think it's a bad system
I'll just ask my friend Kelly
Friends?
Hella assumptions, which I thought mathematical equations are supposed to reduce. Not sure I can recommend this strategy; however, after I heard x5 assumptions, I stopped trying to understand
Totally understand, it is suppose to get you thinking about how to allocate as best as possible. Definitely lots of assumptions but that is how predicting the future works I guess. If you find a better way, let me know. I'd be open to learning any other methods.
@@AndrewBrownInvesting Ray is right here, Kelly assumes constants and not changes in market conditions. Consistent probability of payout is not true in a market, and it doesn't take into consideration the correlation between securities. Also, if you would assume the potential profit equal to the potential loss (efficient market hypothesis), the probability of profit is 0%. There are other ways to use this, but it's just not that simple.
dude na, they never use this formula
Funny that I own this also.
Berkshire?
Are you broke yet?
Not yet 😊
my nigga sid the sloth is wise