I wrote this post because a friend of mine sent me a link to an article introducing Kelly’s formula.
I didn’t read the link. I read it years ago.
This formula uses a bunch of seemingly sophisticated derivations and calculations, which is particularly deceptive.
In particular, people who ask, If it doesn t work, how can there be so many articles online? Are more likely to be fooled.
Practice is the sole criterion for testing truth.
I haven’t actually met any traders who use Kelly’s formula, which means it’s probably “useless.”
I was going to write an article about ripping up Kelly’s formula, but then I realized it wasn’t necessary.
A waste of my energy and a waste of the reader’s time.
It is better to talk about my position calculation formula carefully, to provide some ideas and references for your partner.
There may not be anything as detailed online as what I’m talking about, so there’s a risk of “no one else has said that, so I’m probably not right.”
My position calculation formula is borrowed from the Turtle trading law, to start with a comparison: Turtle position calculation formula: Position size unit = 1% of the account /N* Dollars represented by each point My position calculation formula:
Position = principal * Risk Control/stop-loss/point value is essentially the same, but I have a different understanding and handling of each of the four points, and I will explain each one in detail.
1. Principal “Principal” corresponds to “account” in the Turtle formula.
According to Turtle, 1% of the account means 1% of the net worth.
We usually say net worth = balance + profit and loss.
If no position is held, the net value equals the balance.
If there is a position, the net value is changing, may float surplus may float losses.
According to this understanding to calculate open position size, floating surplus state than floating loss state can open more positions.
The amount of risk a single trade can take will also increase.
This is because undetermined profits are pre-booked into the principal that is used to take risks.
The turtles actually use a “nominal account.”
Dennis gave each turtle an initial $1 million account.
If he loses 10 per cent, Dennis shrinks his account by 20 per cent.
If it loses another 10 per cent, it shrinks by another 20 per cent.
Note that the second reduction is based on $800,000, and a loss of $80,000 reduces the account size to $640,000.
So my guess is that Turtle’s net worth is not necessarily the net worth related to the profit and loss of the position, but the size of the account adjusted for Dennis’s distribution.
Some books suggest a 2% risk limit for a single transaction, which should not be taken to mean that you have to lose 50 in a row to break even.
This is because 2% is not always calculated against the initial principal, but is based on the smaller amount of the current account balance and net worth during the transaction.
If you look at it, it’s actually a “death-proof” risk control strategy.
You can lose 50 trades in a row at any given time until you don’t have enough money left over to open the minimum trading unit of the contract.
This is the same logic of risk control that Dennis used to control the turtle’s account funds.
If you continue to lose money on a trade, reducing your position is not the answer. You should stop trading, find the reason for the continued loss, and then start trading once the problem has been resolved.
So, I made some adjustments based on the turtle.
Determine the initial principal at the beginning of each year.
The initial principal determines the lower limit of the position size.
Even if I lose money, I will not continue to reduce my position beyond the initial principal.
Increase the principal amount in the position calculation formula for every 10% of earnings.
I will artificially control the adjustment speed, maybe the actual return reaches 15%, I will increase the principal amount by 10%.
Or it may be a market after the adjustment of the principal amount.
The advantage is that even if there is a continuous loss after the increase, there is no need to adjust the principal amount repeatedly.
Because my single risk control limit is small, the withdrawal of funds is not large.
In addition, my strategy refers to the fundamentals and position data, only to do a single direction with obvious tendency varieties, the probability of continuous losses is relatively small.
So basically there will be no reduction in the principal amount.
If the strategy has a low win rate but a high expected return, it is better to handle the principal adjustment more carefully, as the Turtle did.
2. Risk control Risk control corresponds to the “1%” in the turtle formula.
The Turtle simply states that the risk factor is 1% (why not risk control will be explained later).
My usual risk control for a single trade is 0.5%, with some of my favorite species being given 1.5% initial risk control.
After locking in some profits will continue to add positions.
One reason I use tighter risk controls than a turtle is that I have more strategic trading opportunities.
Another reason is that my trading varieties are mainly foreign exchange gold and silver oil and stock index, the overall correlation is very high.
Turtles limit their positions in highly correlated multiple markets to no more than six.
Because the pure technical index tracking strategy of sea turtles cannot identify the dominant species and the following species.
There is a situation where one breed is strong and the associated breed is continuously cheating the line.
I don’t do position limits on highly correlated varieties.
Because I only open a position when there is an advantage, I can accept the behavior of related variety line cheating, I will interpret it as a bet on the wrong time.
In the last article, I used the example of shorting the United States and Canada when I was talking about the large-entity K-line strategy. At the same time, I also long Australia and the United States and NewYork and the United States, which are all commodity currencies with a high degree of correlation.
But each variety showed enough strength to justify going long.
Whether it’s foreign exchange, futures or equities, there are problems dealing with the risk of trading highly correlated varieties.
If the strategy is based solely on technical analysis, it is better to limit the total position size and the position size in the same direction.
The purpose of the restriction is risk management.
But to be honest, not necessarily.
It is because of the high probability of rising and falling that people are highly correlated.
The risk is that the strategy may signal much more frequently than the synchronized rise and fall of related varieties.
Does the first signal have to be the leader?
The dark horse may not be able to race because of the “limit number”.
There is no technical solution to this problem. The solution is to look for strength in the fundamentals or in the money side, bypassing the concept of correlation and betting directly on strength.
Most, however, pretend that technical indicators signal strength.
This issue has been discussed many times before. The strategy signal provided by technical indicators is probability, and probability is not the same as advantage.
I did not understand the “probability and Advantage” series of articles written last year.
3. The number of N-value stops corresponds to “N” in the turtle formula.
N is the average daily fluctuation of the underlying trade.
I made an Excel table, which can be more convenient to calculate the daily N value of trading varieties, judge whether there is a big entity K line, and other market information derived from the N value.
In the Turtle’s position calculation formula, the N value is simply used to calculate the size of the trading position.
When I said “1%” I was careful to emphasize that 1% is not a turtle’s risk control because the turtle uses an initial stop loss of 2N.
Taking a position at 1N and a stop loss at 2N means that the Turtle’s initial risk control is not 1% but 2%.
Turtle also has up to 3 times to add positions, add positions after the stop loss moves up, so after adding 4 positions, the turtle’s overall risk control is not 8%, but 5%.
I thought that was too risky, so I replaced the N with a stop loss.
Even though I use 1N most of the time as a stop loss, the number of stops equals N.
The advantage of this is that the risk control on the trade can be locked in at 0.5%, regardless of the actual stop loss.
In some cases, I would bet on the price and it would take anything less than a pricking to trigger a rally.
The margin of this price to the current price is very small, maybe only 0.5N, or even 0.3N, do not need to give a 1N stop loss.
I’m more than happy to bet on that.
If you’re wrong, you lose 0.5%.
If the bet is right, just give 1N margin and you can make a profit of 1%-1.5%.
Compress the number of stops to 1/3, which is equivalent to enlarging the position 3 times.
To me, if you don’t make 1N margins, you’re doing something wrong.
Sometimes a 1N stop-loss is too much and I’ll replace it with a reasonable fixed stop-loss.
Us stock continuous circuit breaker that time, the Dow average daily fluctuation of 1500 points, March 12 a day of 2971 points.
It’s totally unnecessary to give a 1N stop loss, and it’s a waste of trading if you don’t.
I would lower the trading cycle, look for short opportunities on the 5-minute chart, and put a fixed stop loss of 300 points.
The turtle is floating surplus plus warehouse strategy.
I don’t go out of my way to add to a bullish position. If there is a new opportunity to step in, I will continue to place a 0.5% risk control bet, usually after a stop has moved to parity protection or after a certain profit has been locked in.
Also does not exclude floating losses to add warehouse, but not after floating losses to think of adding warehouse, but a good plan to batch entry.
I would only add to a position if there is a strong bias in position data.
Linking risk control and N-stops to positions solves the dilemma of whether to pursue profit/loss or win ratio.
In this binary choice, neither the profit/loss ratio nor the win ratio is the right answer.
Either way, the goal should be to add a few percentage points to the account once the deal is completed.
After a series of transactions, the contribution rate of net profit to the growth of account funds is the most core index.
Using my position calculation formula, everything becomes very clear.
At the risk end, a trade would yield a pullback of at most 0.5%.
On the yield side, each unit of stop-loss profit increases the account by 0.5%.
It solves the following three problems: a lot of points, but a low position, no significant growth of funds.
The stops were small, but the positions were too large and the funds pulled back significantly.
With a fixed number of hands, trading objects with different contract values, risk tolerance and return are not balanced.
After determining the maximum loss on a single trade, I only have to think about one thing, how many N’s there are potential profits on this trade, how many N’s I can eat.
Assuming a 1N stop loss is used on every trade, a 1N profit is a 0.5% increase in capital, even if the values of N are completely different.
1N stop loss for 1N profit, the profit/loss ratio is 1, the profit is 0.5%.
A 1N stop-loss for a 3N profit gives you a profit ratio of 3 and a 1.5% profit.
When the profit and loss ratio is three times, the profit will also be three times. Only then can the profit and loss ratio really affect the growth of capital.
To grow your money by 10%, you only need a net profit of 20N.
Since the principal amount will be adjusted for every 10% increase of capital, the capital will increase by 10% in the form of compound interest for every 20N net income achieved.
Based on this condition, the win ratio and profit/loss ratio of the recalibration strategy are more meaningful.
4. Point values Point values correspond to “dollars for each point” in the Turtle formula.
For the small partner to do the inner disk, the point value is the yuan represented by each point.
So there are limits to this description of turtles.
The value of the contract may vary from subject to subject and must be understood in detail before trading.
The same is the foreign exchange direct trading, the United States, pound and other currency pairs, 1 contract a point is worth 10 dollars.
On the other hand, the point value of the US-Japan, US-Canada and other currency pairs will change with the rise and fall of the exchange rate.
To calculate the position, first to query or calculate the current exchange rate point value is what.
Contract specifications vary from subject to subject.
The primary gold spot contract is for 100 ounces, and a $1 move in the gold price is a $100 change in value.
The first silver spot contract is for 5,000 ounces, and a $1 change in the silver price is a $5,000 change in value.
If the second decimal point of silver is a point, the value of the point is $50.
Trading the same primary contract, with the same $1 fluctuation, the silver contract would make 50 times as much money as the gold contract.
Traders who fail to understand the contract specification for silver and assume that gold and silver have the same point value are at great risk.
Contract specifications vary from trader to trader.
For example, the Dow futures contract is divided into large and small Dow.
I have two platforms. A market maker provides CFD contracts with 10 contracts.
Another market maker offers a CFD contract of five contracts per lot.
The same 1 hand contract, the former point value is $10, the latter point value is $5.
Other traders do offer Dow contracts, but the minimum unit of trading is a standard hand, not a mini hand.
If the size of the calculated position is less than one standard hand, you have to abandon the trade, otherwise you will take excessive risk.
Some people struggle with which point of the mark is a standard point.
It’s not our job to harmonize measurements.
When calculating the position, the point at which the stop-loss points are calculated is the point value corresponding to that point.
In single-digit gold stop-loss terms, a move of $1 is one point and the corresponding point is $100.
In the second decimal place, a $1 swing is 100 points, and the corresponding point value is $1.
1 times 100 is 100 times 1. Remember that, and you’ll be right.
How to calculate the size of a trading position, this article is basically done.
You can adjust the principal, risk control, stop loss points and other points in the formula according to your own trading strategy.
If you only do a single variety, you can moderately increase the risk control.
If you do a variety of different cycles, you can use different stop points.
If multiple policies are running at the same time, different principal quotas can be allocated to different policies. There is no need for efficient and inefficient policies to use the same principal quotas.
The beauty of this formula is that once you combine principal, risk control, stop-loss points, and point values into these relationships, no matter what type of trade you make, no matter what strategy you use, no matter what period you follow, stop-loss and gain can finally change your money curve according to your actual profit/loss ratio.
As a heavy user of OneNote, I usually calculate my position directly in OneNote. If I write the principal, risk control, stop loss and point value into an equal sign after the formula, I can directly get the result, which is super convenient.
Disclaimer: The copyright of this article belongs to the third party author
