Aren’t those simple SMA crosses not history yet?
I was really surprised after testing some basic building blocks for strategies (arbitrary crosses for signal purposes) and firsts results confused me totally.
What’s wrong with these, not even over curve fitted!
First run for 1Y and same for 3Y and still same results. (that was the surprise part)
What are those Crude Oil traders doing? Is oil so simple to trade with basic TA?
Here you have the fresh ecuity qurve in my hobby site:
(http://www.facebook.com/pvoodoo if previous does not work)
Is that a long only strategy, or always in? Add a volatility stop, position sizing logic, pyramid entries for a few entries per stock and retest. Find another asset or 2 to build a trading portfolio
By the way, a simple moving average cross can indeed be curve fitted. Careful…
Mostly allways in (exit on close one start new secuence every day), long or short. no real logic, only one simple cross, 1200 trades. It didn’t work “so well” against other assets. Although nice 1 year curves OK and traditional curve fitting syndrome was possible to see but not with CL and CL had the best results originally. So this might be pure random but I have to check this more … ( and start to trade oil instead of index futures).
(BTW: I try to put some info to site http://pvoodoo.com/Images/Images.html as well)
there is no “easy market”.
Although the chart looks great, there are some “issues”. After 8 months or so, you are still flat (lost all of your gains of about 40%). Would you have continued with the strategy ?
Your performance peaked a first time around June 2010 (+ 180%) to fall to about +105 %. Same question. It took you 8 to 9 months to recover from that drawdown.
So, although the performance looks strong with 20/20 hindsight, you will have a tough time selling it to investors I fear.
I know and easy market turns to hard market immediately when you have an open position there.
Anyway, actually this isn’t even any strategy yet, (no stoploss or targets used or anything, I’ll run those most simplest one later to this) only quick setup to test some signals or amount of those, not even the effectiveness yet… only one building block to be filtered with different filters (trend
as you progress and bring other parameters into play (stops, targets, …), you will probably see your performance drop significantly. Trying to control volatility of your performance comes at a cost.
And probably, but still some nice results with “basic stuff: stop loss /target” (added to facebooksite just) from 1 Y although that might be (read: IS) sort of over optimized already.
As you look at stops, check multiples of percent and a couple of different volatility measures (10 or 15 day vol converted back into a daily rate and ATR). Look at current bar’s/range values, not when the order was placed. As you test single day volatility violations, initial and trailing stops, you may be surprised at both the control on volatility and increase in returns. In my strategies, a fairly tight initial stop loss and an extremely large trailing stop increased APR by more than 5% for both optimized and out of sample testing across 6 portfolios over more than 10 years of historical data – significant money.
Once the volatility adjusted positioning and stops are fine tuned, start playing with your slow moving average and long moving average periods – you should find that you can change them SIGNIFICANTLY without much change in your returns. You may, however, find nicer equity curves – nicer being a more pronounced slope rather than parabolic moves through periods of curve fitted returns.
Your stopped equity curve looks pretty good. How far back have you ever tested that? What I would do, is optimize that for several years in history not including the past few years. Then check it for out of sample performance over the past few years and finally look at the results for the furthest back history in/out of sample data available. You can expect to run a strategy for about 1/3 of the time frame used in testing/optimizing/out of sample testing for your finalized startegy.
I wouldn’t worry too much if you find cyclical periods of outperformance and underperformance – these are known/accepted with trend trading systems.
You may find it interesting to run your equity curve vs. a buy and hold or some other index. Just as some like to compare performance to sharpe/risk free, as if they’d really put money in a 3 month product to achieve a RFR, you may like to see your peformance via a benchmark.
I have not been able to find a target profit entry that provided better returns on any of my strategies however. Once you account for commission and slippage, rentry to a trend (if you allow it) combined with missed continuation of the trend eats up the profit in missed opportunity.
very valid points.
My exits or position sizing has never been, not even near, optimal.
Although this ( this current example “SMA” strategy) haven’t been under my development more than slightly it is so interesting that I have to spend some time with it. More history data and
FOSC (not Stochastics) sort of oscillator filter added and not so bad results…
The SMA crosses is not your friend stop dreaming…
You simply need to build a strategy with more than 33% winning trades and with a reward vs risk >2.
More than 50% of my trades are loosing one but winning trades are two times bigger than loosing one…
Did you check that last summary:
Profit Factor 2.4 , 60% profitable, ratio avg win / avg loss = 1,57
Of course, over optimized BUT Walk Forward results from last 5 Y -> PF 1,65 and that’s quite good, indeed.
Anyway, I’m skeptic to that SMA method but surprised about backtest results.
HOW DO YOU START A PROFITABLE TRADING BUSINESS? Read more NOW >>>NOTE
I now post my TRADING ALERTS
into my personal FACEBOOK ACCOUNT
. Don't worry as I don't post stupid cat videos or what I eat!