Hello,
i have just opened an account and would like to hear thoughts and feedback on the best way to put together a portfolio of systems for AutoTrade.
Would love to hear advise of the best way to select systems, pitfalls to avoid etc.
I am selecting systems by the following filtering process:
1. FX only.
2. the system must be atleast 90 days old,
3. the system must have a fairly high turnover (looking for many small trades not one big winner every few weeks). min of one trade a day.
3. the pay off profile of the system must be nice and smooth - 45 degrees.
Is this a good way of going about things? The point i worry about is persistence…how likely is it that a system that has performed well in the last 90 days is likely to perform well in the next 30.
Hi,
90 days tells you absolutely nothing. Ask the provider for a backtest result including commission and slippage. Of course that is not verified and you have to decide if you trust the vendor.
Remember that more trades means more commission, more slippage and more autotrading fees (if you use Gen3). With these counted in, many profitable looking systems are actually losing money for subscribers.
Whatever you do, the most important thing to keep in mind is leverage. It is very very easy to overleverage here and and blow your account. A good rule to use is that if the system loses 50% of its account value, you lose 10% of yours.
When you look at some equite curves here, you might think making money is easy. However the truth is that making money by trading is never easy. Concentrate on preserving your capital and if you select the right systems, you might get decent gains (10-20% pa. longer term).
- Panu
Michael -
Some random and not so random thoughts…sorry if you know this stuff, but this is also for readers who might not know as much as you:
To start, the fact that you actually have criteria beyond “I want to make a lot of money” is great. Most people don’t even get this far. I think you are on the right path.
1. FX only.
OK. I’ll assume you understand the risks/pitfalls with most Forex brokers, where the brokerage takes the other side of your trade, rather than routing it through the interbank market.
Also, I’ll assume you realize that a great deal of diversification in Forex is not practical, since there are only a handful of major pairs, and most of those are correlated.
Lastly, I’m sure you’ll stay away from exotic pairs, since the spread you have to pay is huge, and will be very tough to overcome.
So, once you have your candidates, you can combine them by using the “Build a Portfolio” feature on the right side on the page that has one of your candidates on it. It helps you create equity curves from multiple systems.
2. the system must be atleast 90 days old,
90 days is certainly better than 30 days, but not nearly enough, unless it is from a developer you already know and trust. I’d say 6 months minimum, really 1 year. The more the better, since you want to see many different market conditions.
Don’t bother asking for backtests, since 90% of vendors don’t have a proper one (purposely or not, it is easy to “cheat” with backtests. I used to have the same issue myself, and I always wondered why realtime never equalled backtest results). If you do rely on a backtest, cut profit by 50%, and increase max drawdown by 100%. If you see something âtoo good to be true,â it is â ignore it.
3. the system must have a fairly high turnover (looking for many small trades not one big winner every few weeks). min of one trade a day.
Ask yourself why you want this criteria. If it is profitable, and there are a statistically significant number of trades, why would this matter? I suspect you are looking for some excitement with trading. If so, this is a warning sign. Great trading is boring.
4. the pay off profile of the system must be nice and smooth - 45 degrees.
This is a great criteria, if you can find it. I looked at the Grid, for forex systems > 180 days, <40% drawdown and Calmar > 1. 14 systems popped up. MAYBE 1-3 of them have even close to 45 degree equity curves. After your other due diligence that I hope you do, you’ll be left with maybe 1 or 2. So much for diversification.
So, that might mean you have to back off your criteria…
If I was forced to pick a system (that I did not develop) to trade with real money, here’s what I would look at:
Age > 1 year
DD < 30%, unless it recovered very quickly. Many did during financial crisis.
Calmar>1 at a minimum
Vendor’s other systems - no spectacular failures with other systems
Vendor reputation - has to have some sort of good history outside of C2
Adding to Losing Trades - never, unless vendor has record of doing it on small scale, responsibly
Avg Trade - at least 2x the slippage and commissions
I hope this helps.
Kevin
Hi all
Thanks for your guidance and thoughts - i will refine some of my criteria based off the feedback and try out 4 systems on very low leverage and see how it goes.
Here’s my 2 cents:
For each system in your basket: analyze intratrade risk (the risk indicator in the trade list with the colors) it indicates the vendor might hate taking losses. You’ll most likely see this happen with systems that have +50% correct trades. It’s a real mental fight sometimes to keep the win% high. A winning trade that went deep in the red before closing out with a relatively small profit is a recipy for disaster. I look for rather low win%.
You need high win%? pick only true mechanical systems.
Good luck!
Martijn Deketelaere
Michael -
It would be interesting if you post your new criteria when you create it, and also keep a running log on how you are doing. You wouldn’t have to mention system names. It would help educate others, and also help you stick to your plan.
Kevin
great write up Kevin !
Personally, I think you would be better off choosing the systems with the higher equity curve in the Portfolio Maker. JMO.
