Yes, you'd of course want to do due dilligence on that trade with your account or more on the line! However, if it's an oversight on your part of poor risk management during a price shock as you described, then unfortunately you or I would have to wear that loss, and they'll get their money in the end, even if you've got to flip burgers for 20 years. There's no such thing as going bankrupt to your broker.
risk:reward and Kelly's criterion- Holy schmoley !
- Thread starter Scotty2Cues
- Start date
- Watchers 3
Yes, you'd of course want to do due dilligence on that trade with your account or more on the line! However, if it's an oversight on your part of poor risk management during a price shock as you described, then unfortunately you or I would have to wear that loss, and they'll get their money in the end, even if you've got to flip burgers for 20 years. There's no such thing as going bankrupt to your broker.
Hotch
Experienced member
- Messages
- 1,409
- Likes
- 256
The point is if you calculated your kelly and it's < 1% then even if you can "accept" a 1% risk, then you will blow up your account, be sad and die alone.
Kelly is the max you can risk, personally run some EAs at around half kelly easy because there's no emotion involved and it's rather consistent (for reference, that's 10% risk).
wackypete2
Legendary member
- Messages
- 10,211
- Likes
- 2,058
If you blow up your account with 1% risk then you have bigger problems with your trading and none of the obscure calculations will help you. For all the years I've been trading I have never used most of these, nor do I know how some of them are calculated. I preferred to concentrate on how to make money consistently while being able to sleep well at night.
If it works for you then fine, I have no problems with that at all. My personal opinion is that new traders don't need to be confused by all that math.
Peter
Hotch
Experienced member
- Messages
- 1,409
- Likes
- 256
I would say if you get confused by the maths then you have bigger problems 
The point stands for any number, the point is you can be comfortable with a risk which is bigger than your Kelly, and that's the no-no. I thought that was obvious.
You're right it's not a topic needed for beginners, but once you are consistent, it doesn't hurt.
The point stands for any number, the point is you can be comfortable with a risk which is bigger than your Kelly, and that's the no-no. I thought that was obvious.
You're right it's not a topic needed for beginners, but once you are consistent, it doesn't hurt.
Scotty2Cues
Established member
- Messages
- 737
- Likes
- 33
Not sure of the best way to use it, I've put the formula in terms of % edge where % edge is the difference between probability of success and the payout odds.
So maybe place SL and TP levels on a high probability setup. Lets say risk:reward is 1:1 and we estimate probability of win to be over 70% so our edge is 20%. We could lower this to 5% edge (we shouldnt be in the trade if our edge is 0%) and calculate the Kelly number from that.
Or play around with different risk values (below the Kelly number with the 70% success rate) using the 70% and see how the account grows...
So maybe place SL and TP levels on a high probability setup. Lets say risk:reward is 1:1 and we estimate probability of win to be over 70% so our edge is 20%. We could lower this to 5% edge (we shouldnt be in the trade if our edge is 0%) and calculate the Kelly number from that.
Or play around with different risk values (below the Kelly number with the 70% success rate) using the 70% and see how the account grows...
But i think there are some important assumptions which serve as the basis for Kelly, like each trial should be independent and i think the probability is always the same (i think) which makes it less applicable to discretionary trading, but more adaptable for system trading and obviously casino games. Edward Thorp has written some good stuff on this, i think his original work for Beat the Dealer was inspired by reading something like Kelly.
Yours
Yours
wackypete2
Legendary member
- Messages
- 10,211
- Likes
- 2,058
Scotty...it appears what you are doing is trying to optimize your Kelly number to fit your trading. To my feeble brain this makes no sense. If you input realistic numbers and get the Kelly answer but don't like it then you feed it different numbers or scale it down? If that's what is happening then it makes no sense. If I'm reading it wrong then someone try explaining it to me...sincerely, i'm not wisecracking.
Peter
wackypete2
Legendary member
- Messages
- 10,211
- Likes
- 2,058
It's worth noting that one of the biggest proponents of optimal bet size and kelly related formulae, Ed Seykota, doesn't actually use any of the maths used to derive these calculations for his position sizes. He bases it on what drawdowns he's able to tolerate and the probable drawdowns which will occur trading a certain % size.
meanreversion
Senior member
- Messages
- 3,398
- Likes
- 537
I know Seykota is a bit of a legend, but I'm not sure I'd want to have dinner with him.. too much of a wiseguy.
Yeah, i could do without being psycho-analysed about how much mustard i have with my ribeye steak.
