Page 1 of 10

Craigaudio's ER2 Buyzone Strategy

Posted: Tue Dec 11, 2007 5:19 pm
by craigaudio
Eudamonia asked me if id put add a post about the buyzone setup Ive started using on the ER2 which I pretty much copied off him, so Ed, thanks on the starter for ten :-)

Goes without saying id also id like to thank Avery for all of help, sharing his strategies, indicators and advice. Big Kudos.

With that said lets move on...

Buyzone Parameters:

Trigger Long: Open Price + 11 ticks
Trigger Short: Open Price ? 11 ticks
Target: 10 ticks
Stop Loss: 12 ticks

When I see the open price for the day I subtract 11 ticks in my head and place a limit short order with my broker. I add 11 ticks to the open price and place a limit buy order with my broker. When one order is filled I pull the other order.

I use Transact futures and have the orders automatically setup so that if filled the profit and targets are automatically set when I getting filled.

I make maximum 2 trades per day. If the first trade is a winner then I go enjoy myself.

If the first trigger is a loser, no stress, I will only take a second trade if it hits the opposite side of the buyzone.

For example, lets say I went short in the sellzone and was stopped out for a losing trade, I would take no more shorts that day, the only possible trade for me would be to go long if the buyzone trigger was hit. If that trade is a loser then I quit for the day. As I live in the UK I only to trade the morning futures session.

I do not trade every time price crossed the buy or sellzone.

If the second trade wins then I have drawn for the day, 1 loss vs. 1 win, and take the drawdown on commissions.

This is such an easy strategy to back-test, taking minimal time and effort.

From my own testing by eye and a few horizontal lines I have gathered the following results:-

02/11/2006 to 11/12/2007

# Trading Days: 282
# Trades: 385
# Winners: 241
# Losers: 144
Win Ratio: 63 %
Ticks Per Trade: +1.9
Max Consec Losers: 3
Max Consec Winner: 11

Starting Capital: $10,000
Risk Per Trade: 4% (compounding, max limit of 30 contracts per trade)
Ending Capital: $148,320
Profit: $138,320 (inc. commissions $5 round trip)

With an average win of 1.9 ticks per trade, with say 22 trades per month, thats a profit of 41.8 ticks per month. With 10 contracts per trade thats a smooth $4180.

With 10 contracts you make $1000 for every winning trade (excluding comms). Over the year I have tested so far there are occasions where you have had 11 max consecutive winners before a draw or loss (max 3 consecutive losers). Thats a tidy profit of $11k in just over a fortnight.

If you compounded your capital (like I do) and traded a percentage of your account for each trade, lets say 3%, then as you win your stakes get higher after each success and also reduce if you encounter a losing. Due to the high hit rate and the attendant liquidity in the futures market, with an account of size $100k taking 3% risk per trade, you could be trading 25 contracts making $2500 per trade before commissions. Nice.

I would post the spreadsheet here but to be honest, if youre seriously considering trading this method you should definitely do your own due diligence and research. It will certainly give you confidence that the system works well. Sods law you will encounter 3 consecutive losers when you start trading and be totally put off by the approach.

Note that just because the ER2 has behaved in this volatile manner for the last 12 months it doesnt mean that it will continue to do so, it just gives us more confidence that the pattern will continue. Continuation is more likely than change. Monitoring the entries and exits are necessary in order to tweak the entries and exits if required.

Also you might be thinking ?hey that trade isnt even 1:1 risk reward, the stop is bigger than the potential profit. Well from the research i've done so far the strategy works much better with a bigger stop due to the fact that the open price often acts as support or resistance and can help keep you in a trade.

As far as the target goes, Im currently looking into the effects on the numbers if the target is raised by a few pips more but havent quite got round to that yet. A lot of times it seems you could squeeze out more but I dont have the figures to back that up yet. If you find that you have better results with a different targets and stops then id love to hear more.

As you can see, this strategy is extremely good on emotional capital. You dont spend all day long staring at charts, winning trades, losing trades, clocking up commissions, and eventually ending up breakeven or even worse... at a loss for the day. I?ve had plenty of days where Ive sweated blood and tears only to end up with a negative P&L :-(

With this approach you are trusting the statistics, the emotion is totally taken out of the trading. I spend maximum of one hour per day on this and the figures backup that it will be able to provide me and my family with an excellent standard of living. You cant really ask for more than that

Trade to live, not live to trade.

Posted: Fri Dec 14, 2007 2:12 am
by Ken_S
Thanks for sharing Craig.
I coded this in TS last week and tested it on ER2 using Ed's parameters, but the equity curve didn't look too good. I only used one entry, so maybe your reverse entry is the difference.
I'll check my code and post the results.
Ken

Posted: Fri Dec 14, 2007 6:56 am
by fcg0lfer
Thanks alot for sharing.
After playing around with some backtesting, i too was not able to duplicate your results. What time does your buyzone start? I am assuming 930EST when NY market opens, is that correct?
Frank

Posted: Fri Dec 14, 2007 6:22 pm
by eudamonia
Craig,

Cheers for posting your results. It's great to see someone be able to back up what Avery and I have been saying for awhile - the BuyZone works!

Ken and fcg0olfer,

If you are using TradeStation you CANNOT correctly test limit orders. These trades have been tested with LIMIT entries, and LIMIT STOP orders as well as PROFIT TARGET LIMITS. You might think that 1 or 2 ticks a trade doesn't make a difference - well it does!

Look at the results. The average trade makes 1.9 ticks. If you SURRENDER 1 or 2 ticks every time you trade what will happen to your results?

Have you read my thread "Why most aspiring traders fail"? Rule number 1. B.S. backtesting! Don't be fooled by programs that can't do tick precise entries (like TradeStation)!!!!! I have personally wasted thousands of hours with TradeStation's B.S. backtesting - I know what I'm talking about.

There are three ways to test a method with real TICK precision: 1) Walk forward real-time testing, 2) REALISTIC tick by tick simulation (aka NeoTicker or similar software), 3) Sub-minute candle-by-candle backtesting with volume (to ensure fills). Thank you!

Edward

Posted: Fri Dec 14, 2007 8:14 pm
by Ken_S
Craig,
"From my own testing by eye and a few horizontal lines I have gathered the following results:-

02/11/2006 to 11/12/2007

# Trading Days: 282
# Trades: 385
# Winners: 241
# Losers: 144
Win Ratio: 63 %
Ticks Per Trade: +1.9
Max Consec Losers: 3
Max Consec Winner: 11

Starting Capital: $10,000
Risk Per Trade: 4% (compounding, max limit of 30 contracts per trade)
Ending Capital: $148,320
Profit: $138,320 (inc. commissions $5 round trip) "

This looks like computer generated results, not testing by eye over 282 days. Correct?
Edward,
I welcome a discussion on realtime and backtested results of any method shared on this forum. I'm sure I and others have much to learn about the pitfalls of backtesting especially with TS. I look forward to discussing the subject more.
Respectfully,
Ken
Today would have been stopped out in the morning correct? Would you have entered short when price hit the short zone at about 14:04 or 14:50 est?

Posted: Fri Dec 14, 2007 9:50 pm
by eudamonia
Ken_S wrote:Craig,
"From my own testing by eye and a few horizontal lines I have gathered the following results:-

02/11/2006 to 11/12/2007

# Trading Days: 282
# Trades: 385
# Winners: 241
# Losers: 144
Win Ratio: 63 %
Ticks Per Trade: +1.9
Max Consec Losers: 3
Max Consec Winner: 11

Starting Capital: $10,000
Risk Per Trade: 4% (compounding, max limit of 30 contracts per trade)
Ending Capital: $148,320
Profit: $138,320 (inc. commissions $5 round trip) "

This looks like computer generated results, not testing by eye over 282 days. Correct?
Edward,
I welcome a discussion on realtime and backtested results of any method shared on this forum. I'm sure I and others have much to learn about the pitfalls of backtesting especially with TS. I look forward to discussing the subject more.
Respectfully,
Ken
Today would have been stopped out in the morning correct? Would you have entered short when price hit the short zone at about 14:04 or 14:50 est?


Ken,

Understood. Sorry to come off as being really harsh, it's just that I've seen too many problems with TradeStation backtesting to have much tolerance for the many guys who say "The XYZ method you trade every day and make money with doesn't backtest well in TradeStation". :roll:

The same folks have been telling Avery that the BuyZone doesn't work on the stocks and currencies. Sorry Avery stop making money - TradeStation's backtesting says it doesn't work...

To answer your question, yes Craig did do an "eyeball" test of the BuyZones and asked me if this corresponds to my own realtime results. I have been trading this setup on the BuyZone for 9 months with real money and for another 3 months before that with a demo account. His results are remarkably close to what I've actually seen. Does it seem strange to you that TradeStation's results are so off?

Ken this rant is not specifically or only directed at you but at the many people who get fooled by B.S. Backtesting. I just want traders to spend some time thinking about HOW their backtesting platform actually works. To do that you need to see where TradeStation originally came from (Omega Research) and understand that it was originally designed to test EOD (end of day) data. Once you see that, you will see that TradeStation has been modified to do more precise data increments (modified like dilapidated low-rider). And yet TradeStation is still a slave to the 1 minute bar. Nothing within 1 minute is KNOWABLE in TradeStation for backtesting. Why? Because there is NO timestamped tick data. NONE! So TradeStation must "guess" where and when price goes within that one minute. Do you want to be guessing whether your trading will be successful?

Now you might be asking why do I care what happens within 1 minute. That's not a very long time. True - if you are trading a method with an average holding period of more than 30 minutes you don't need tick data. But if you hold your trade an average of 2-5 minutes (like the BuyZone) then it is ever so crucial that you KNOW what happened - exactly. You cannot test it otherwise. You cannot place a limit order and see if will be filled. It cannot be done!

Lastly, I will just say that I find it sad but somewhat amusing to think that some guy puts 5 minutes worth or work into backtesting something and suddenly he thinks he's an expert on the method. On the other hand Craig put at least 30 or 40 hours (or more perhaps) into hand testing the results and challenging whether the inputs were satisfactory. And thinking about why the BuyZone works. Ultimately the computer is just a tool. YOU are the key to your trading success. And trading is hard work - not easy money contrary to popular belief.

Hope that clarifies my points.

Edward

Posted: Sat Dec 15, 2007 3:13 am
by Ken_S
Hmmm. This appears to be a lively yet worthwhile exchange. Not trying to discredit anyone, just trying to get things clarified. Avery has graciously shared so much IP, and I read people asking for more ideas from him, but little posts about realtime results using the many ideas on this site.

Not sure if you think I'm the 5 minutes of work guy, but I too have spent 100+ hours working on the buy zone. Spent $1000 automating it for equities, another $1000 debugging the automation program, and many hours of coding, backtesting, and screentime.
I don't understand your comments about TS being a slave to the 1 minute bar. I do all my analysis on a 1 tick chart. Are you saying the ticks on a historical chart in TS are not in the correct sequence within the minute? If they are in the correct sequence, then please explain why TS and 1 tick charts shouldn't be used for backtesting.
One last point, for now, you say you always use limit orders when trading the buyzone. If the buyzone trigger price is 764.60 and the trade price is 764.50...then the next trade price is 764.60, do you enter a limit order in for 764.60, or a limit order at current ask(which may be 765.60, or higher, or lower)?
Best Regards,
Ken

Posted: Sat Dec 15, 2007 5:32 am
by TheRumpledOne
Backtesting never makes you any money, only trading does.

The only thing I care about when I backtest is after the entry is whether or not there is a chance to exit with a profit.

With the Buy Zone, I know STATISTICALLY that I have a 4 out of 5 or 9 out of 10 chance of making money trading the stocks or currencies that I trade.

Funny thing about the BUY ZONE is the less trading knowledge you have, the faster you'll learn and the better you'll trade the BUY ZONE.

It's also amusing how many people want to "warn" others about trading it. I'm catching so much flak over this simple, profitable method.

I equate the BUY ZONE to the bumble bee: in theory a bumble bee is too heavy and it's wings are too small for it to fly, BUT BUMBLE BEES FLY. I guess bumble bees just don't get it!!

Posted: Sat Dec 15, 2007 5:25 pm
by bh_trade
Edward:

I don't understand what you are saying about Tradestation. It has been my experience that TS assumes best case fills (i.e. a print = a fill) when in actuality, more often than not price has to tick *through* you to be filled, especially in highly iquid index futures. So basically, TS overstates profitability, not understates, especially if the logic is using limit orders, which the BuyZone clearly would be.

Posted: Sat Dec 15, 2007 9:39 pm
by eudamonia
Ken,

Sorry to say but you have wasted a lot of time backtesting in TradeStation and a lot of money automating something not worth automating (for short-time frame methods).

Yes the ticks are not timestamped. And TradeStation doesn't allow a resolution in backtesting of less than 1 minute for this reason. Don't believe me? Check the TradeStation wanted items thread and look up "Time stamped tick data" - you'll see many users who have asked for this feature - TradeStation has not responded positively to this since the ENTIRE architecture of TradeStation would have to be altered to allow it (i.e. bar driven data).

I use something that you can't do in most back testing programs (even NeoTicker). I use a STOP LIMIT ORDER. It ensures that I get filled at the price I want 95% of the time. If my BuyZone is at 765.60 that is the price I get 95% of the time. The other 5% of the time my order gets converted to a market order (i.e. no volume at my price) and I give up 1 tick.

bh,

Agreed limit orders are "filled" when a price prints. However, there is no STOP LIMIT ORDER in TS. So all of your stops are at market (i.e. 35% of the time you LOSE 1 tick). Also because of the difficulty in coding it most people are not using a limit entry order they are using a market order (losing another tick every trade).

Edward