Has anyone been successful backtesting a system that is consistently profitable?

[demarcog]

naifwonder,

Looks like you are are using Decembers (Z) contract month. Try using AB #F for the continuous contract, as in July that contract was not actively traded which causes gaps in the charts.

[drobertson]

demarcog,

Is there anything out there for the S&P E Mini ?

[demarcog]

drobertson,

Funny you should ask.

The e-mini has been an obsession of mine for many years.

Any system I developed, the baseline for testing it out was the good old S&P emini.

Even when I wasn't system trading, the emini was the only market I traded for several years, and there were some lean ones after 2000 in that baby.

The good new was, what better place to learn how to discretionary trade and system trade ... the TOUGHEST instrument, during the TOUGHEST time period (lowest volatiltiy) in several decades.

That S&P intraday, is the hardest nut to crack for me if ever there was one.

Once I have a system that makes money on the e-mini I feel like I climbed Mt. Everest.

The question is, why trade a system that makes 10 cents a trade in the SPY'ders or a few ticks in the emini when you can trade the same system on the Russell and make alot more?

Unless you are hedging OEX positions or trading monster size or have another reason for trading the emini S&P?


To answer you question directly, I have systems that make money on the e-mini BUT, because that market is so choppy it's better to apply a little discretion and that becomes a slippery slope.

If you are asking about a bit longer term system, that hold for several days is a different story, there are several that do welll.

I'm assuming you are asking about intraday and rignt now I'm focusing on the shorter term development, becasue they make more money and have smaller drawdowns, although the precentage profitable and average trade is lower and smaller then the longer term systems.

The short term systems for the most part usually do well longer term (15-120 minute intervals) but again trade more infrequently.

The problem with the emini S&P is that there are some things that are difficult to put in a EFS strategy. Like the strong time of day factors, like how far has this market moved before I got this signal, like is today a Friday in August where the 1st hours range is 2 handles, like is today the day before a Fed announcement etc.

Bottom line, until it starts moving 25 handles a day like it did for many years, we are kind of stuck with swing trading systems that hold for 2-3 days.

[bandraguy]

With all these posts, just to make a point that there is a system, which works 100 percent.

Now that we know, there is a system out there which works. Where is the file? Can we download the file and start using it.

Is this someone's personal system and don't want to share with trading community or is it for sale?

[naifwonder]

So I backtested the system today.. the results were not as good as I hoped but still decent. It got approximately 60% profitability, which means after losses, yields a 20% profit. Not bad, but I will not settle for anything less than 80%. Just about all of the trades did go into profit at one point or another, but since I had preset profit targets, they were not realized. Therefore, I will be editing the system until I can make the momentum indicator give dynamically generated profit targets. This should lock in more potential profits, be able to take larger profits, and increase the profitability rate. Until I can get the momentum indicator to do this, I do not think it will be logical to use this system nor would it be fair for me to sell it. I guess it was just the month of september that gave such extraordinary results and I am sorry for getting everyone's hopes up.

But on the bright side, as I had said before, if I can get the momentum indicator to give profit targets customized to each trade, than it is likely that I can get this system to a level where it can be traded with confidence.

I know that this system has potential and will be working on it very heavily until this potential is realized. As such, I will keep you guys posted on any further developments. Thomas Alva Edison in regards to when he was making the lightbulb once said these famous words, ""I have not failed. I've just found 10,000 ways that won't work". Technically speaking, 60% profitabilty is still a profitable rate of trading, but not enough for me. I still have a few thousand attemps to go.

And once again, thanks for all the feedback and suggestion guys - it has helped me immensely.

[drobertson]

Demacar, I am seeking a system that can satisfy my initial thirst. I am a newbie and currently getting my wallet *****. I need a little help.

I know you ??? my reason for trading S&P E Mini. It is all that I know at this point. THanks for getting back with me.

[demarcog]

Originally posted by drobertson
Demacar, I am seeking a system that can satisfy my initial thirst. I am a newbie and currently getting my wallet *****. I need a little help.

I know you ??? my reason for trading S&P E Mini. It is all that I know at this point. THanks for getting back with me.

Do we know each other?? email me [email protected] ....

I know all about loosing money......the main purpose of all this is to have us put our heads together and come up with something we can make money on....


My long winded point in the last post was if you are getting ***** trading the S&P emini maybe it's not just all your fault, and what do you mean it's "all I know"'

Change the symbol, everything else stays the same...try it for a few days...

........try the Russell 2k.....

.and if you really want to trade something that trends try the oil etf.....OIH but that's another conversation

[theplumber]

Originally posted by naifwonder
So I backtested the system today.. the results were not as good as I hoped but still decent. It got approximately 60% profitability, which means after losses, yields a 20% profit. Not bad, but I will not settle for anything less than 80%. Just about all of the trades did go into profit at one point or another, but since I had preset profit targets, they were not realized. Therefore, I will be editing the system until I can make the momentum indicator give dynamically generated profit targets. This should lock in more potential profits, be able to take larger profits, and increase the profitability rate. Until I can get the momentum indicator to do this, I do not think it will be logical to use this system nor would it be fair for me to sell it. I guess it was just the month of september that gave such extraordinary results and I am sorry for getting everyone's hopes up.

But on the bright side, as I had said before, if I can get the momentum indicator to give profit targets customized to each trade, than it is likely that I can get this system to a level where it can be traded with confidence.

I know that this system has potential and will be working on it very heavily until this potential is realized. As such, I will keep you guys posted on any further developments. Thomas Alva Edison in regards to when he was making the lightbulb once said these famous words, ""I have not failed. I've just found 10,000 ways that won't work". Technically speaking, 60% profitabilty is still a profitable rate of trading, but not enough for me. I still have a few thousand attemps to go.

And once again, thanks for all the feedback and suggestion guys - it has helped me immensely.

Have you ever traded ES in a live account with money? I'm not being sarcastic as it will pertain to my point on system trading the Eminis. Es has a habit of 1 you being in the end of the que even if you enter with what you think is a fast broker and software and 2 ES has price revisitation like no other. When you enter an order do you hit the bid or ask or wait for price to come to you? In the real world of system trading you will encounter situations where the system entered an order but that price was thick and your order didn't get hit so have you programmed hitting the bid/ask price? This has to be thought of when trading ES. Then if you do hit price ES has a habit of price revisitation, does your system cancel if price moves .75 away from price or do you sit there hoping for it to come back? Sometimes it will and sometimes it won't but these factors can play havoc on any system. Point 1 is the human factor. Not all brokers are really direct. They have internal checks and sometimes a book for their own traders/brokers. Do they advertize this? Hell no but it happens and sometimes you think your somewhere in the middle of a thick price but your instead at the end because your broker held it internally until price went against your position and they took it to flip. Anyways just some of the games people play that can create problems for any system trading ES.

[naifwonder]

Yea I took that into consideration. The system isn't auto traded yet so I don't have to worry it automatically voiding out orders. If you look at the bar layout, even if I were to just leave the entry order sitting there and it didn't fill for a while, I would still get a proper entry and get my profit. If I didn't get the entry, there is no reason to force it. I would simply just not take the trade.

And yes, I do trade on a live account with money (I trade for a living). I have traded the ES, but I prefer to trade the ER2.

I also reworked the system recently and I increased the success rate to 87.5% but I am still trying for the 90% goal. I made it so it does not count a profit target reached unless the current price exceeds the target price by at least one tick. This guarantees the trade getting filled. For example, if I am long and my profit target is 730.20, this system would not count the trade successful unless the price actually hits 730.30. More profit would be good so I am still working on a way to at least increase the quanitity of trades by 40% or so while still maintaining the success rate. However, I will not trade in quality for quanitity. Anyway, the link for the results is below:

Click here to see the complete results for the new system

[demarcog]

Originally posted by naifwonder
Yea I took that into consideration. The system isn't auto traded yet so I don't have to worry it automatically voiding out orders. If you look at the bar layout, even if I were to just leave the entry order sitting there and it didn't fill for a while, I would still get a proper entry and get my profit. If I didn't get the entry, there is no reason to force it. I would simply just not take the trade.

And yes, I do trade on a live account with money (I trade for a living). I have traded the ES, but I prefer to trade the ER2.

The plumber makes some very valid points. Stuff Happens...trading no doubt about it, especially during fast market conditions.

(I'm curious who he executes his trades through and what size he is trading becasue my experience with that market trading through Tradestation Brokers several years ago and RediPlus recently trading smaller size is that 90 percent of the time a market order was filled instantly....and frequently you can buy at the bid and sell at the offer if you are willing to wait a few seconds, if your not filled you go market and pay the 1 tick slippage.

Outside of FOREX, the emini S&P is the most liquid market there is, the spread is 1 tick and the size is on average several hundred contracts bid and offered.

He raises some very important questions which we could go on about but what he is talking about is "Slippage".

If there is a 1 tick slippage of 12.50 then I double it. If a stock like the SPY or IWM is 1-2 cents wide, then specify 5 cents. That will give you a worst case scenerio, for those off the wall times when you get an incredibly poor fill.

I also reworked the system recently and I increased the success rate to 87.5% but I am still trying for the 90% goal. I made it so it does not count a profit target reached unless the current price exceeds the target price by at least one tick. This guarantees the trade getting filled. For example, if I am long and my profit target is 730.20, this system would not count the trade successful unless the price actually hits 730.30. More profit would be good so I am still working on a way to at least increase the quanitity of trades by 40% or so while still maintaining the success rate. However, I will not trade in quality for quanitity. Anyway, the link for the results is below:

Click here to see the complete results for the new system

Would you rather have a system that is 100 percent profitable and trades once a week for an average trade of +$100. Or a system that trades 5 times a week, is correct 50 percent of the time for an average trade of +$50 a day or $250 a week.

I guess it's personal preference but to me the second system makes 2.5 times more money then the first system and I would consider that a better quality system?

My point is you definately can achieve very high percentage profitable systems, I have a binder with a bunch (not 100 percent but 75-90) but they traded so infrequently that I felt the missed opportunity cost in not being in the market was a real practical issue. Do I sit there all week waiting for the next trade?

My criteria is high returns, and LOW drawdowns, I don't want to see you banging your head against the wall trying to achieve perfection....when you account balance or clients only care about the bottom line and how much pain (drawdowns) were endured.

[theplumber]

What I was basically saying was are your trades executed at the bid or ask( or market order for lack of explaining) or do you join the bid or ask ( which would be a limit order).A market will be executed but a limit will sit on internal broker books until they release it. A computer backtest is no where near real world for ES, I know because I tried it. Worked great for about 3 weeks and then the erratic behavior of ES price action made multiple hiccups in execution and stops being hit. Price revisitation makes for tight stops to be hit even when your system is right. This can't be coded into a backtest so all backtests are worthless for ES. So my only advice would be to trade 1 ES contract real world and that will be your testing, win or lose.

[demarcog]

Originally posted by theplumber
[B]What I was basically saying was are your trades executed at the bid or ask( or market order for lack of explaining) or do you join the bid or ask ( which would be a limit order).

Most strategies that I have developed base trade entry decisions on bar(-1) and use MARKET-THISBAR, which will report in the strategy backtester, the opening price of the current bar as the trade price. Then specifying a 1 tick slippage will provide a very realistic representation of an actual trade in the ES. A quick look at time and sales for the ES #F will illustrate that 99 percent of the time the spread is 1 tick wide by several hundred contracts thick. So unless you are trading more contracts then currently offerred or bid, you are going to get filled on market orders at the offer for purchases and at the bid for sales.

I also have strategies that I backtest using limit and stop orders,
and while there is no 100 percent gurarantee that becasue a trade occurs at a particuliar price that you will necessarily get filled at that price (plus then 1 tick slippage) it is close to 100 percent.

Again, given the massive liquidity of that market virtually guarantees a fill....I've traded it thousands of times and never waited longer then a few seconds for a fill.

A market will be executed but a limit will sit on internal broker books until they release it.

The Emini S&P contract trades electronically on the CME. Whenever an limit order is placed you should immediately see that order reflected in the current Time and Sales BID/OFFER SIZE fields.

Not sure what you mean exactly by sitting on "internal brokers books". Whether it be a limit order on a listed stock, nasdaq stock, or Emini S&P, in general the broker is legally obligated to post your limit order immediately and having trading full time for several years that's what I usually see whenever I place an limit order, market orders are filled usually within a second or two and limit orders are reflected electronically in level two type quote displays or time and sales information.....almost immediately.

A computer backtest is no where near real world for ES, I know because I tried it.

I have a backtested strategy result that I'm looking at nowfor example that reports a buy trade at 10/20/06 at 14:05 at a price of 137300, at the time based on time and sales the market is: 137275 bid 137300 offerred 424x236.

I consider the bid, offer, size and trade information in Time and Sales to be as "real world" as it gets for me. A market order at that time would 99 percent of the time be filled at the current offer (and the 1 tick slippage for good measure).

Worked great for about 3 weeks and then the erratic behavior of ES price action made multiple hiccups in execution and stops being hit. Price revisitation makes for tight stops to be hit even when your system is right.

Not sure what you mean by multiple hiccups in execution, but stops getting hit are definately an issue for systems. Never heard the term "price revisitation" but if you are getting stopped out and the "system is right" maybe experiment with looser stops. Is it possible this is strategy specific and not indicative of some flaw in the accuracy of backtesting the ES?

all backtests are worthless for ES.

Not sure upon what you base this conclusion, but I would respectfully disagree and interested in hearing what specific issues you base this conclusion on.

I would encourage anyone seriously considering risking their hard earned and possibly irreplaceable trading capital to take full advantage of the backtesting facility, as it significantly less expensive then real trading.

I think that the more liquid the market, the more valid the backtested results, not the other way around. The Emini S&P contract alone is a bigger market in nominal dollars that all the combined trades on the NYSE on an average day. It can trade a million contracts a day times the current prices of 65K per contract is 65 billion dollars a day. The NYSE does 1.5 billion on an average share price of perhaps $25. If there is ever a market that the backtested prices reported are realistic it is the emini S&P.

If anyone has any doubts abouts whether a trade price reported in the backtester is realistic, simply open up a time sales window and see what the "real market" was for that time period.

The strategy object used by the backtester when coded according to the techniques in the tutorials is sufficiently accurate and IMHO an invaluable tool.

I would welcome anyone who has had a different experience or disagrees to please feel free to reply as we all could benefit...

[naifwonder]

Hey, hows it goin guys? Sorry I haven't been around for so long, been busy making more systems and adjusting current ones. After some more tinkering, I've created one more system that I will now be using. 5 trades in 5 months for 23 points worth of profit (that's after commission). I figure if I keep making systems like this, after I get 3-5, I should be getting about 10 trades a month wich is roughly 1 trade every other day. That would solve the inactivity problem for me.


I created one more system but the accuracy is at 70 percent, which I cannot personally tolerate. The profit targets are at 8.5 points and the stop losses are at about 2.1 points and all trades are intraday. It seems that the failed trades have been a result of sudden market reversals due to bad news or something of the sort and I have to figure out a technical way of equating that out. Until than, the system will be left on the shelf.

Like I said before, 1 good trade is better than 2 good and 1 bad.

Just wanted to address something though. It is my fault I did not make myself more clear Doji so I apologize:

In the case of a 1 tick slippage value for each trade, this would result in a -$600 for Naif's results - netting $2400 (about) for his 120 day test. Still very respectible.

I already have slippage accounted for in my system, well actually almost completely accounted for. The system enters the trade if the price is one tick below the entry and exits if it is one tick above the exit by means of limit orders (Opposite for short trades). That virtually guarantees the entry and exit since I am offering a better price than the market price. I've made a few touch ups on the system though since than. Stop losses are the only time where slippage may really be a problem but even if I were to tack on 1 point for slippage for those stop loss orders, I would still be in good shape. The improved version nets about 50.25 points in profits (that's using 3 ticks for slippage which is quite a lot).

Anyway, the new system results are below:

[xxjohnxx]

Does anyone know of a Bollinger band strategy using z score. That I can use to back test the E-mini Russell. I would like it to show buy and sell signals on the chart too.

Thank You In Advance

[sams]

Hello,

did anyone made a (back)test for trading an e-mini future when S+P and NQ and YM and Russell 2000 goes the same direction or e.g. MACD gives for all the futures a buy or sell signal?

regards sams