Research continues

The past few days I’ve been doing more reading and research for interesting techniques and strategy ideas.

1)    One important thing I’ve learned is that risk dictates either the position size or stop loss. So if the acceptable loss/risk amount chosen relative to the account balance of $7000 is 1%, that’s $70.

–       If I open a position of 10 lots, the stop loss should be calculated such that I can only lose 1%. Obviously if my position size is 100 lots then the stop loss will be an order of magnitude closer… this might be too close! If so, then…
–       A stop loss is chosen and a position size is calculated such that I cannot lose more than $70

2)    Another brief thing to learn is that we should never ignore the USDOLLAR, even when focusing on cross currencies.

3)    Today I learned there is a difference between volatility trading and spread trading. Essentially both are trading a correlation ratio between two streams, however, in volatility trading we are correlating the object against itself (say a moving average of its own price). That way if the spread moves too far from the moving average, effectively we’re trading the volatility. Spread trading would be trading the spread between two currencies.

4)    Tips to improve your strategy:

  1. One way to prevent over-optimizing is reduce the number of tunable parameters in your strategy so such a perfect fit cannot be made. Deliberately choose reasonable values instead of optimizing them too well.
  2. Use “out of sample” data. So optimize your strategy against a set of data and then backtest against another set of data not seen by your strategy.
  3. Your system should not “break” with minor changes in your system parameter values. If it does, then it is more likely to break against new market data. You should test this.
  4. Take into account price slippage, transaction costs, and rollover interest charges – they can make the difference between a profitable trade or not – especially for low expectancy (range trading for example).
  5. Selling is not the same as buying, and it is very common that some strategies perform well on one side compared to the other. This is not a problem.

5)    I also read an article about trading systems used in SAXO bank. The article talks about the frequency that trades are made. Some strategies only make one or two trades a year, while other strategies make several a day. So therefore, I should keep this in mind that it is okay for some strategies to not trade often, the important thing is that they make money!

6)    If you take the log of the price, a change from 9-10 is equivalent to a change from 90-100… therefore, you can more accurately compare price shifts. I guess the same could be achieved if you take a percentage of the difference as well but that would mean you need a base to calculate the differential from. The log simply allows you to plot the price as is and look at the change visually. If you simply wanted to calculate the relative change, then you can calculate the percentage.

7)    We can calculate an index by using the formula
INDEX = 100 * (100 / [1 + x]) where x is the ratio of Delta(A) / Delta(B)
This is equivalent to
INDEX = 100 * (Delta(A) / [Delta(A) + Delta(B)])
So this basically gives you a percentage of Delta(A) on top of the sum of both Deltas…. You can then use the extreme values (under 20, over 80) as possible signals?
Example: The money flow index           http://en.wikipedia.org/wiki/Money_flow_index

8)    To calculate a ratio of a fixed number, we can do something like this…
RATIO = [Count(A) – Count(B)] / [Count(A) + Count(B)]

So basically we calculate the total and then the ratio of the difference. If we are mostly A, then we will be largely positive, and vice-versa. The good thing about this is it stays between 1 and -1.

Posted in Backtesting, Money Management, Optimization, Research and tagged , , , , , .

One Comment

  1. Pingback: What’s the aim of optimization anyway? | MooMoo Forex Trading

Leave a Reply

Your email address will not be published. Required fields are marked *