# Bakwaas Trading

#### augubhai

##### Well-Known Member
Fractional Kelly:

Here's the link to a very lucid article about the applying the Kelly Criterion.

The Kelly Criterion
How to apply the famous betting formula in investment, trading and professional gambling.

by Nicholas Yoder

When leverage is increased, the Edge of a bet grows linearly with the amount of leverage but the Negative Geometric Drag (NGD) grows as the square of the leverage.

Here's is a chart that I should be interested in. I still have to get to learn about calculating this, but for now it can be a handy reference.

#### augubhai

##### Well-Known Member
Fractional Kelly:

Below is another figure that shows the same thing in a different way. This one is by William Ziemba, who has done a lot of research in this area.
Source: ResearchGate

Probability of doubling and quadrupling before halving and relative growth rates versus fraction of wealth wagered for Blackjack (2% advantage, p=0.51 and q=0.49)

1. Kelly Criterion gives the optimal bet size for maximizing return.
2. Geometric Drag associated with higher bet size may make application of the full Kelly Wager unfeasible.
3. In no case, should the bet size be greater than the full Kelly Wager, because there is a decrease in return, as well as an increase in drag.
4. The incremental benefit of increased bet size decreases as we approach the full Kelly Wager size.

#### Happy_Singh

##### Happy
Hello augubhai

Reading the content from your link . . .
Kelly isn’t the goal, but rather the boundary.
Seems, limiting the bet size to 1/5 - 1/3 of Kelly Criterion is what is being suggested
By always betting 30% of the Kelly-optimal bet size, a gambler reduces the chance that he will drop to 20% of his peak bankroll (80% draw down) from 1-in-5 to 1-in-213, but he still keeps 51% of the growth of a Kelly-optimal bet.
Wouldn't it be nice to get a calculator where we plugin the past results of our system
and then get an exact lot size to be traded on next trade . . .

.

#### augubhai

##### Well-Known Member
Hello augubhai

Reading the content from your link . . .

Seems, limiting the bet size to 1/5 - 1/3 of Kelly Criterion is what is being suggested

Wouldn't it be nice to get a calculator where we plugin the past results of our system
and then get an exact lot size to be traded on next trade . . .

.
I actually do this using a spreadsheet. I calculate the value for every month to see the variance.

For calculating the Kelly Criterion, you only need the win rate and payoff ratio. Can be easily done on historical data.

#### Happy_Singh

##### Happy
I actually do this using a spreadsheet. I calculate the value for every month to see the variance.

For calculating the Kelly Criterion, you only need the win rate and payoff ratio. Can be easily done on historical data.
OK

The win rate is 50%, so conservative adjusted rate for calculation is 40%
the payoff ratio is 2.5, again conservative adjusted value 2.0

These values give a bet size (risk) as 25% (For an online calculator click here)

The document you shared, recommends around 30% of Kelly's bet size
so in the above case it will be 8 %

Is this right?

.

#### augubhai

##### Well-Known Member
OK

The win rate is 50%, so conservative adjusted rate for calculation is 40%
the payoff ratio is 2.5, again conservative adjusted value 2.0

These values give a bet size (risk) as 25% (For an online calculator click here)

The document you shared, recommends around 30% of Kelly's bet size
so in the above case it will be 8 %

Is this right?

.
Correct, but check it for different months/years. You will find drastic variation in the Kelly Bet Size - including -ve values for losing months. In trading, we will never have an accurate value for the Kelly Bet Size.

The historical value can at best give us a level of confidence to bet a higher size.

#### augubhai

##### Well-Known Member
I mainly use the Kelly Criterion to compare backtest results of different systems. The system with a consistently higher Kelly Value can be traded with more confidence.

I also use it to fine tune systems. For example, whether to keep a target or not, whether to trail or not, performance on different scrips, etc.

Always consider the brokerage and slippages while inputting the values for the calculation.

#### Happy_Singh

##### Happy
Thanks for this discussions, helpful .

I am in the process of tinkering with the MM for my carry trades,
want to increase size after a -ve streak

Don't have any predictive model for winning/losing streaks,
only if I had one then won't need all this kelly-belly

.

## Intraday Higher Leverage

Save up to 90% in brokerage and get higher leverage for intraday trades.

 Name: Phone: Email: City: State: Select StateAndaman and Nicobar IslandsAndhra PradeshArunachal PradeshAssamBiharChhattisgarhDadra and Nagar HaveliDaman and DiuDelhiGoaGujaratHaryanaHimachal PradeshJammu and KashmirJharkhandKarnatakaKeralaLakshadweepMadhya PradeshMaharashtraManipurMeghalayaMizoramNagalandOdishaPuducherryPunjabRajasthanSikkimTamil NaduTelanganaTripuraUttarakhandUttar PradeshWest Bengal Are you a day trader?