Episode 957: Martin Shkreli Explains How To Size Your Trades Correctly (Mathematically Proven Method)

**Martin Shkreli** (0:00)
Is it trading waste of time? Well, I'm so glad you asked that, because I came up with this thing called Keli. I don't want to show it to you. Actually gonna adjust it later today, but here's what it looks like. All right. So, this is a coin. It has a 60-40 probability of landing on heads. How much do you want to bet? Your bankroll is $25. How much do you bet each time? Do you go all in? I bet you want to go all in. Let's see what happens if we go all in. Go all in each time. Well, you know what's gonna happen. One of the times you're gonna flip and you're wrong. So, you might get a couple flips right in a row, four. But you're eventually gonna go broke. So, all right. Maybe 80% of your money? Nope. Pretty bad as well. I made it five flips, three flips. Oh, 28 flips. Went broke anyway. All right. How about 60% of the time? Should I flip my money?
Okay. Whoa, that was good. Survived for a little bit there. Still in the game. All right. I went from 25 bucks to 14 Gs. Now, you probably would have taken your money out at the top, right? Let's spin again. Daryl isn't so lucky. He only got up to 40 or 25
Same thing. Rough. Seems like it crashed out pretty fast here. All right. Let me go to 50 50-50. Half the bankroll each time. That should work, right? Infinite. Infinite gains. Oh, shit. Well, he got to 8K. Keep the dream alive? How about 45% of my bankroll?
Oh, shit. Let's go. 50K. 80K. 600 See, I would have taken all my money out at 600 I know that's the number. Oh, never mind. I would have taken it out at 8 million. What a run. 70 mil. Let's go.
Yeah, it looks like it's over. Round tripped from 75 mil to 5K. Dang.
It's over, bros. So the reason this is important is because I want you to think about bet sizing. Generally speaking, everyone over bets. The Paul Tudor Jones lesson. So anyway, if we go to, say, 15% of the bankroll, how do we do?
Do you wave getting crushed? Yeah. Crushed. But yeah, it's done.
So you can see this is a pretty efficient process. You know, if you have an ability to get 60-40, hey, Christina, if you got a 60-40, which a lot of us, I think, you know, if you think about it, you know, can you be right in the market 60% of the time? 60-40? That doesn't sound crazy. And you see, if you control your bet size, look at how the simulator is up to 600 million, now billion. All it had to do is find 900, basically 1,000 bets where it was 60-40. Can you find 1,000 bets over your life that are 60-40? I bet you can. Look at Capper. Look at.
You know, Casava. That's it. You got $2 trillion, 1,300 flips.
So the Kelly criteria tells you how efficient it is. The optimal bet is actually 20%, which is 60-40. But if we make the edge just 55, the optimal bet is 10%. And it's mathematically proven that's the right bet. Not 9, not 11, but 10 And so let's do 9 and see how we do. This is a much smaller edge. It's just 55%. But look how it's able to preserve capital. And now we're back green. Had some really bad luck there for a minute.
The point isn't that Kelly gives you some perfect secret recipe to making money. There's no secret recipe.
Yeah, 5% edge is 55, so 55 minus 45 is the Kelly Criterion, which is a 10% bankroll bet each time. And you'll see how this compounds. You never know with precision your edge. There's many, like I said, this is a idealized concept, but it's supposed to be a concept, nothing more.
The concept is that if you type Kelly or you do, actually, Kelly.godel, this should work. Let's see if this works. kelly.godelterminal.com. Yeah, try this out. There's no timing. That's the idea here.
How's it going? So you see, here, it takes a lot of flips. So the edge is just 55, which is a lot less than 60 And you see just how many flips it takes. It takes forever. Big drawdowns, too.
Microns are getting smoked here. I want to see 700 something. There, feels like it.
Intel 2
And there it is, after 3000 flips, which, you know, 55, 45, I bet you can do one of those a day. So this would take you 10 years, but you would turn 25 into 60 million. So it's that consistency, just constant 55, 45, constant 55, 45 And guess what? Being right 55% of the time, doesn't feel like you're a genius. It kind of feels normie. But look what the results get you. Billion dollars. And the Koreans, what are they doing? They're betting their bankroll way too high. That's why they crash out right away, basically.