Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Know Yourself, Know Your Setup, by Boris Schlossberg 9 replies

I dare you show your statement 13 replies

Twoblink tweaks and mods 61 replies

Changing Your Self Talk By Talking to Yourself 2 replies

Attachments: First-Strike & One-Night-Stand users, show yourself and your mods, if you dare!!!
Exit Attachments

First-Strike & One-Night-Stand users, show yourself and your mods, if you dare!!!

  • Post #421
  • Quote
  • Edited Jul 24, 2010 7:14am Jul 23, 2010 7:40pm | Edited Jul 24, 2010 7:14am
Hi All,

I wonder how many people out there are still trading FSP and ONS?

I am hanging on there and continually looking for ways to try and improve them.

I trade ONS with a relatively tight volatility-based stop which, of course, has a lower percentage of winners but does better overall (according to my backtesting). Annoyingly I transferred some balance to an Oanda sub-account causing one of the orders to not be filled due to insufficient margin. Even though the overall risk is lower, the smaller stop causes much bigger positions to be taken... Unfortunately a losing trade was taken but not the winning USDJPY trade on the 16th July. A learning experience...

Anyway, onto the subject of my post. I have been playing around with reducing the position size for FSP based on the distance from the open to the buy/sell price compared to the 6 week average. This seems to improve the results considerably. I attach the results of my backtests reducing the percentage to values between 0% and 2% with 2% being a fullsize position.

The rule is if the 6 week average divided by the current distance is less that 1.25 then use the reduced position size. In other words, if the current week has a distance to entry that is a bit less than average then use full size.

Using a "reduced" position size of 0% gives the best risk-adjusted results but could miss a really good move.

The logic seems reasonable though - if you have a large move required to enter a position then, i) more of the move will be exhausted moving to the point of entry and ii) the position size will be smaller.

The change seems to eliminate a worthwhile number of whipsaw moves.

I've tested on all my hourly data from difference sources and it consistently improves the results, which is encouraging. The results attached are using hourly data from FXCM from 2005 onwards (GBPJPY is more recent).

Has anyone had similar experiences or have any thoughts?

Regards,
Timek

Attached Image (click to enlarge)
Click to Enlarge Name: FSP_stepping.png Size: 11 KB


To clarify the meaning of the columns: CAGR% = annual percentage return, MAR = CAGR% divided by max equity drawdown and Max TE = Max equity drawdown
I've been trading FSP and ONS in the same account for the past 57 weeks taking all the trades using 1% risk per trade and following the rules as close as I can and my account is down 52%. The last three months looked like I was coming out of the draw down until the last couple of weeks. Tough system to keep taking the trades but I'm planning on doing just that. It appears to me that there's not a lot of traders left using these methods.

Thanks for sharing Timek
Quoting Crickets
I've been trading FSP and ONS in the same account for the past 57 weeks taking all the trades using 1% risk per trade and following the rules as close as I can and my account is down 52%. The last three months looked like I was coming out of the draw down until the last couple of weeks. Tough system to keep taking the trades but I'm planning on doing just that. It appears to me that there's not a lot of traders left using these methods.
I admire your steadfastness and with a mechanical system such as this (when you have the back-test data to prove its long term profitability) it's the consistency that will eventually see you into profit.

I think it's safe to say the market conditions have been different in the last couple of years which is why you're seeing a prolonged drawdown.

I'm not speaking for Joel, these are only my assumptions, but I cant imagine Joel working a system that ended up being down on the year. This means one of two things - either you're doing it wrong (probably not) or that Joel would have adjusted his approach to suit the prevailing market conditions - perhaps something to think about?


Edit: I originally mentioned Peter rather than Joel as I got confused however it still stands - perhaps there is some adjust you can make that will provide more wins and might be worth doing some research on.
one year does not a system make.


A year is 52 trades per pair.......hardly a large enough sample to determine a mechanical system's profitability.

I have been playing for the last 40 some odd weeks and am down as well.

The only way I am considering tampering with the system is by dropping the GBP/JPN cross from it. The reason being is that if you trade the majors. You are trading vs the USD. Once you introduce the crosses, you start throwing additional variables in to the mix and I suspect you are just increasing your risk exposure without a comenserate increase in profitability. When you play the GBP/JPN, you are really trading the G/U and U/J. So essentially you are doubling down on those pairs when you throw the GPB/JPN into the mix.





Quoting MattW
I admire your steadfastness and with a mechanical system such as this (when you have the back-test data to prove its long term profitability) it's the consistency that will eventually see you into profit.

I think it's safe to say the market conditions have been different in the last couple of years which is why you're seeing a prolonged drawdown.

I'm not speaking for Joel, these are only my assumptions, but I cant imagine Joel working a system that ended up being down on the year. This means one of two things - either you're doing it wrong (probably...
Quoting Crickets
I've been trading FSP and ONS in the same account for the past 57 weeks taking all the trades using 1% risk per trade and following the rules as close as I can and my account is down 52%. The last three months looked like I was coming out of the draw down until the last couple of weeks. Tough system to keep taking the trades but I'm planning on doing just that. It appears to me that there's not a lot of traders left using these methods.

Thanks for sharing Timek
Great steadfastness in following the system through the drawdown! I stopped almost a year ago. I trade another system full-time now (check my post history if you want to know more).

Are your results in a form such that you could see what it would be like had you only risked 0.2%/trade for 1% total risk/week?
Hindsight is 20/20
Quoting razorboy
...
A year is 52 trades per pair...[does not a system make]
You're absolutely right, I wasn't really factoring in the thought that it's really only one trade a week. Provided your backtest data covers several years with a profitable outcome then one bad year is no reason to tamper with the method.

I wouldn't, however, trade DIBS for a year and happily see a loss
Yes, it's certainly a challenge to keep on taking the trades. I've been trading FSP and ONS for over two years and am currently in a 18% drawdown (largely because I've kept on adding capital during the drawdowns).

I've programmed FSP into my backtesting software and here are the results using a risk per trade from 0.2% to 2.0%. firstly up until the last equity peak last summer:

Attached Image (click to enlarge)
Click to Enlarge Name: UpToPeak.png Size: 16 KB


Secondly up to today:

Attached Image (click to enlarge)
Click to Enlarge Name: UpToNow.png Size: 16 KB


Note these results do not include slippage or rollover interest costs/income. Start date was mid 2005 for all pairs apart from GBPJPY. Hourly data as before is from FXCM (note that my data from Barclays is considerable more optimistic). Oh for the consistency of exchange-published prices!

I would also like to trade DIBS but currently do not have the time to monitor the market looking for those hourly inside bars. My backtesting has shown very favourable results for it, though.

Happy Trading!

Regards,
Timek
Interesting how the results drop off markedly with the recent data, it's sure been choppy markets for medium to longer term trend followers recently. Timek, did you use TradingBlox to test? I also have TradingBlox but my coding is not good. Are you happy to share the code, or can I buy it from you?
Thanks
Hi Range Trader,

Yes, this system has been in a big drawdown for a loonnng time!

I can post the latest version of the system (though it has not been tested on the latest version of Trading Blox) in the "Blox Marketplace" section of the Trading Blox forum. I placed an earlier version there but I've made some improvements since.

It is amazing how very small things can drastically affect the results. Therefore I'm not sure how accurate the simulations are for determining your actual results but they are good for:

i) comparing modifications
ii) determining whether you have an edge.

I am now trying to incorporate the cost of carry into my simulations and have downloaded all the interest rates from Oanda for this purpose. I was shocked that my interest was over 700 GBP last year....

Best,
Timek
Hi Timek, that would be great if you can post it in the Blox Marketplace, thanks! It'll be interesting to run some tests on this and see what the optimisation space looks like, my guess if it's anything like any of the other medium to LTTF systems then most parameters will have produced fairly miserable results over the last year. All the LTTF systems I've tested (Chan BO, MA cross, BB BO etc), including those using so called adaptive technology have been crushed the past 12 months. Interesting you mention cost of carry, I think you are right it is absolutely necessary to include this in the tests to see realistic results.
Update: still placing trades for this system (see figure)...
Attached Image (click to enlarge)
Click to Enlarge Name: LB01EURUSD.gif Size: 12 KB
Quoting asasa
Update: still placing trades for this system (see figure)...
So that's the data for trading this system with the euro for almost the past 3 years? (1 trade/week? ~145 trades in the chart?)
Hindsight is 20/20
Quoting cosgrove
So that's the data for trading this system with the euro for almost the past 3 years? (1 trade/week? ~145 trades in the chart?)
EURUSD for the period 12/2008..06/2010
I'm also using this system with other 4 pairs
PS: Remember I'm using my modified version of the system, I have already written about this
Quoting asasa
EURUSD for the period 12/2008..06/2010
I'm also using this system with other 4 pairs
PS: Remember I'm using my modified version of the system, I have already written about this
Ah yes, thank you for your continued updates.
Hindsight is 20/20
Quoting cosgrove
Ah yes, thank you for your continued updates.
Another wonderful week this one for FSP...
Hi Asasa,

Yes, it's been not bad. We'll see how it concludes tomorrow morning. I'm currently up around 17% for the week (across all three of my systems).

Would you care to comment how your MM approach compares to mine suggested at the top of page 29? I have not used this approach as (hopefully) we are getting towards the end of a large drawdrown and this is not the time to change the risk/reward parameters greatly. In future I may choose to switch...

Regards,
Timek
Quoting Timek
Hi Asasa,

Yes, it's been not bad. We'll see how it concludes tomorrow morning. I'm currently up around 17% for the week (across all three of my systems).

Would you care to comment how your MM approach compares to mine suggested at the top of page 29? I have not used this approach as (hopefully) we are getting towards the end of a large drawdrown and this is not the time to change the risk/reward parameters greatly. In future I may choose to switch...

Regards,
Timek
Even if it is calculated in a different way, your MM looks very similar to mine, you reduce the position size when the volatility increases, i.e. you increase the position size when the volatility shrinks.
Has anyone considered trading the FSP on a daily basis but using weekly or previous 5 days ATRs?

For example on monday enter the trades as per standard FSP instructions. On tuesday enter trades using 30% of previous 5 days ATR (not previous weeks data), on wed, thursday and friday repeat the trades as per tuesdays instructions. (i,e using previous 5 days ATRs)

During the following week close mondays trades and re-enter new trades. On tuesday close previous tuesday trades and re enter new ones again based on previous 5 days ATR. Repeat the close of previous trades and re entering of new trades on wed thurs and friday

The reason I thought of this was to capture trends that form later in the week where your monday trade may have already been stopped out.

I havent tested this in any way and am still at the "mulling it over" stage. I jiust wondered if those that adopt FSP for day trades use this method and whether it is robust.

Thanks
this guy was doing what you are asking about.

cheers

geoff

Quoting DaBuschi
To answer your question:

I follow the exact same rules on the daily system as I do on the weekly. So I determine the previuos weeks range and all that stuff. The only difference is:

I get up every morning to get the 06.00 GMT price and place my trades and I use two accounts to do so as my broker/marketmaker doesnīt allow hedging within the same account.

The only difficulty I experienced very quickly was, that itīs tough to keep track of all the trades and different stops. So I created an excel spreadsheet, wrote some macros and now itīs more...
Thanks for responding Geoff. I did try to PM him but he is not accepting PM's.
0 traders viewing now