Sorry that should be higher lows when going up and lower highs when going down. I think I may have messed that up. As far as the equity curve Id be wary of listening to others and see for myself if it works or doesnt. The thing about those sorts of tests is the assumptions that you use to calculate them. I can tell you for a fact that this system requires an extremely good execution system in order to make money. At a place like fxcm it will lose money hand over fist no doubt. But really Im just trying to elaborate on the way that you need to think about money in order to be successfull. In this case I would liken it to splitting the atom. You split the thing in half and then you can divide and conquer it. Maybe... If I was looking to test a system like this by hand Id probably do use a demo account from Interbankfx.com who also suck in order to get a couple of months historical data for free for the 1 minute. Then Id say something like ... "Well its averaging 600 pips per day gross if you just take it point point to point from high to low and vice versa. Then theres the overhead on 200+ round trips per day which maybe with a top notch execution we could get down to 2.5 pips round trip. In other words its largely a matter of getting to grips with overhead and whether the trading software is some shitty bucket or more like a banks liquidity with tight spreads and very deep liquidity. In my experience there is a law of diminshing returns with the euro where daily might net you 1000 pips and 4 hour might net 3500 pips per year. And so on getting more lucrative the lower the time frame but also becoming more dependent on liquidity and exceution spreads the more turnover you are having to wade through.
Lemming pinhead trader or piker is 2 words commonly used to describe someone who trades like this. All products move this way the problem is that futres are really shitty and really only the US stock market and the currency markets are worth looking at. Ive built systems like this for 500 dollars or less I just cant post the pictures as they have my name written on the interface and thats not happening. All I would do is only construct systems using Fix (financial information Protocol). Id use www.quickfixengine.org program the api in C++ the same as they use to code most 3d games these days and then Id pay IB the 500 dollars it takes to use there fix connection thus bypassing TWS completely. A good C++ programmer is 20 bucks an hour these days...
Why would someony call themselves chipmonk? Gee i wonder... Maybe hes a smurf or maybe he like microchips with embedded logic. Actually the greatest secrets kept from the public pertian to money and thats why there always a disturbing lack of clarity over key issues. Probably the biggest national secret the governmant keeps is which systems in the massive global network are amking the most pips per year . Period. Because if you know that and can pyramid then youll soon be wealthy beyond belief.
BGH, Apologies, slapped that together all too quickly from a pre-existing system of mine and to put it simply, I screwed up... I tend to dismiss insanely obvious approaches immediately and should have given this the more critical look it deserved. Anyway, here's a clean write of your system and the accompanying equity curve. It really deserves closer examination, kudos. // Big Game Hunter's EURUSD System // TradeStation 8.3.1634 // Commission: $2.50/trade // Slippage: 0 // Interest rate: 0 // Account size: $100K Inputs: iTradeSize(1), iStartTime(600), iEndTime(1800), fImprovement(0.00001); Variables: iMP(0), iTime(0), fLongStop(0), fShortStop(0); iTime = Time; fLongStop = High + fImprovement; fShortStop = Low - fImprovement; if CurrentBar > 2 then begin iMP = MarketPosition; if iMP = 0 and iTime >= iStartTime and iTime < iEndTime then begin // flat, bracket buy/short orders within active trade times Buy iTradeSize Shares Next Bar at fLongStop Stop; SellShort iTradeSize Shares Next Bar at fShortStop Stop; end else if iMP = 1 then begin // long, update stop Sell iTradeSize Shares Next Bar at fShortStop Stop; // SAR within active trade times if iTime >= iStartTime and iTime < iEndTime then SellShort iTradeSize Shares Next Bar at fShortStop Stop; end else if iMP = -1 then begin // short, update stop BuyToCover iTradeSize Shares Next Bar at fLongStop Stop; // SAR within active trade times if iTime >= iStartTime and iTime < iEndTime then Buy iTradeSize Shares Next Bar at fLongStop Stop; end; end;
Thanks for the attachment... As i say Im pretty distrustfull of that kind of testing and would probaly favour paper trading and forward testing in order to be as accurate as possible with the assumptions that a backtest requires. Put another way I think if you can get the tests to work forward with just a good ecn like ideal fx via paper trading then I know that with a full level 2 montage of 5+ providers my result for bigger sizes will be as good as or better than the testing would indicate since there should be tighter spreads and better liquidity. Originally I was favouring a more laid back approach like 4 hour euro but recognised immediately that if you could really drive down costs such as commssion and slippage then it may be possible to use leverage in a much more aggressive way if the strategy is very edgy.
Big Game Hunter, What would your system do in a situation like this that I've circled, specifically the area highlighted by brackets? Looks like there would be a lot of entry, loss, reentry in other direction, loss, reentry in initial direction, loss, over and over. a) what happens here? b) is your system truly ALWAYS in the market, or is there a filter against being in sometimes? c) also in the choppy area on the middle/right, there are HHs during the down trends and LLs during the up trends. How would this do here? Thanks.
Or here. I tried to draw the lines at the exact points of entry. BGH, I'm not trying to give you a hard time; I'm curious how this system does in periods of chop like this.