Half-Baked Potato

[TWI]

Incidentally I always start with a range trade and only kick in the breakout trade if this subsequently occurs. Sometimes you sell close to top of range and it breaks out of bottom those are the good days. The reversal value is quite small so there is a high number of losses to profits (Eur, GBP and CHF < 40%, profitable ) but it takes nearly every large break.

 

[JonnyT]

I don't think you can autotrade FX with TS yet. The provide FX brokerage via a thrid party but haven't integrated it yet.

Someone please let me know if they do!

JonnyT

 

[TWI]

You can only autotrade with the futures. It is the only way unfortunately until the FX is automated.

 

[turtle trader]

FXCM have started a 'System Execution' service if that's any use. You have to send them your EasyLanguage code (which you can encrypt) and they will sign a confidentiality agreement. After that you can leave them to it.

I've not used it so can't comment any further - here's the link:

http://www.fxcm.co.uk/system-trading.html

Simon

 

[TWI]

Minimum account size of $100,000
$10 commission per lot round turn for accounts under 100K

I can get enough leverage for my own needs from far less capital.
I wonder whether this is autoexecution or if they have some morons sitting there executing on signals.
Not sure about the 500k desk cost.

 

[tradermaji]

thank you everyone for your contribution. If you trade with a forex broker like oanda, transaction costs will be minimal, but will be there. However, they are part of doing business. My intention is to try to profit from the daily ranges, and that is why I suggested that modification.

Any other modification will complicate the real elegant system you have, but here is another suggestion. Use a narrow range filter for your trades. Thus you will trade only on days where the previous day's range is a percent of the average daily range. For example, if yesterday's daily range was less than 0.75 times the average daily ranges for a 10 day period, place your daily breakout trades. Usually, there is a thrust move after these periods of consolidation. Kind of Toby Crabel's idea.

Also, smoothing out the equity curve is a good thing, and perhaps better than net profits. If you are using some kind of money management scheme, a smoother equity curve will result is better performance as it prevents large drawdowns, and perhaps lower bets at those levels. If you can try a test where your maximum risk is 3% of the total equity per trade. If you use Oanda, you can use the variable size lots, and perhaps the advantage of a smoother equity curve and compounding will be better realized. Again, these are just suggestions.

Thanks again for the great ideas and backtests.

 

[TWI]

Thanks for yours.
Will certainly investigate your range suggestion. I have seen setups that are similar in the past but not experimented much with them. I tend to keep things pretty simple. I tried many things with ATR and although found it improves certain systems on certain markets I have found a problem finding a method that provided a consistent improvement across a number of markets.

 

[FetteredChinos]

knackers. have reviewed the coding on my revised spreadsheet, and there was something seriously kooky with it. correct results would have been a lot, lot less. so please disregard it.


ah well, back to the drawing board

FC

 

[robber]

FC, Hi, look on the bright side; presumably you spreadsheet in your first post is still valid? The results from that are still well worth the effort. It would be good to see a report on the fri to mon trades but using the original strategy. Any chance of running a test?

regards

Neil

 

[TWI]

Your curve from the first spreadsheet is looking the same as mine. A shame about the second one, but I suppose the curve was looking a bit holy grailish.
This thread got me looking at options on days. It creates interesting results if you just start knocking out certain days of the week. I have been convinced for a while now that Friday figure days really do need to be accounted for. It is a real pain to try and code every single major figure day but doing something like narrowing stops on Fridays certainly seems to improve results.
It is interesting just looking at Fridays alone. For example for a while now if you took a position in Eurodollars on a friday based on the direction it moved on the Thursday you would have been doing very well. Also tried this on the Dax with similar results but needed to make the entry criteria based on not only Thursdays but also the Monday to Thursday net move to see consistent results. Certainly the Eurodollar was simpler and more rewarding. Problems came when going back over a couple of years, then it would have been better doing the opposite. When you see things like that it does make following the thing a little harder but thought about the possibility of relating it to interest rate cycles which would flip whether it buys or sells, this is possibly some sort of horrible curve fitting idea that ends in disaster, I do not know. Maybe better to use some sort of moderation device like an Average to determine direction or whether a signal is taken at all.
It is just so easy to spend so many hours at this and wonder where the week went.