Why exactly do EMA crossover systems not work? | Page 12 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Help. EMA Crossover with different EMA cross for exit 0 replies

EMA crossover (ema 14 & 34) 4 replies

Please help code an EMA Crossover 12ema x 24 ema cross 1 reply

Systems, Systems and Systems... 1 reply

EMA Crossover Systems 1 reply

Attachments: Why exactly do EMA crossover systems not work?
Exit Attachments

Why exactly do EMA crossover systems not work?

  • #221
  • Jul 24, 2011 3:42pm Jul 24, 2011 3:42pm
Quoting Troikaone1
You are correct.....as I mentioned in my last post....a computer is unbiased. It can not differentiate certain things when you start talking about where the crosssover is occuring with respect to the trend. It treats all crossovers the same regardless of the circumstance. A human can make a much better judgement when you start introducing multiple variables that a computer cannot interpret.
yes, that is true. But I personally think you can program a lot of different conditions and make a computer differentiate between good an bad signals (of course the data all has to be price/time based).
If you incorporate fundamentals then that's a totally different story, then of course a computer backtest is worthless, since there are no data providers for that yet (at least I have not encountered one).

If I can systematize my decision making process and it is all price/time based, then it is programmable in my opinion.
  • #222
  • Jul 24, 2011 3:46pm Jul 24, 2011 3:46pm
Quoting Custos
yes, that is true. But I personally think you can program a lot of different conditions and make a computer differentiate between good an bad signals (of course the data all has to be price/time based).
If you incorporate fundamentals then that's a totally different story, then of course a computer backtest is worthless, since there are no data providers for that yet (at least I have not encountered one).
True......it would take about a million lines of code for a computer to realize something that I can simply eyeball in 10 seconds. That is awesome. It would take another 10 million lines of code to figure out why a particular crossover didn't work. You can make a backtested crossover system work, you just have to use proper money management I suppose.
  • #223
  • Jul 24, 2011 3:48pm Jul 24, 2011 3:48pm
Quoting Troikaone1
True......it would take about a million lines of code for a computer to realize something that I can simply eyeball in 10 seconds. That is awesome. It would take another 10 million lines of code to figure out why a particular crossover didn't work. You can make a backtested crossover system work, you just have to use proper money management I suppose.
haha, well, programming trading systems, even complex ones doesn't require so many lines, haha. But you have to give it some thought of course.

Well, money management might help make the profitability better, but if you don't exploit a recurring pattern then no money management will save your account, at least that's what I found from my testing. That is why I will never go back to manual eye-balling, cause I cannot know if I am exploiting an edge or if the signals are just random. But after a backtest I can be pretty sure if I exploit an edge or not and then just tune up the profitability by employing some money management & portfolio rules.
  • #224
  • Jul 24, 2011 3:51pm Jul 24, 2011 3:51pm
Quoting Custos
haha, well, programming trading systems, even complex ones doesn't require so many lines, haha. But you have to give it some thought of course.

Well, money management might help make the profitability better, but if you don't exploit a recurring pattern then no money management will save your account, at least that's what I found from my testing.
OK. Again....I would have to see a good system to be a believer.
  • #225
  • Jul 24, 2011 4:01pm Jul 24, 2011 4:01pm
Quoting Pipalicious
When you refer to "interpreting", are we talking about subjective interpretation that cannot be logically defined?
Based on my four years worth of data there appears to be certain types of nuances in price that simply cannot be programmed into a computer's logic. I believe the market makers do this on purpose....in an effort to confuse anyone trying to exploit simple crossovers using simple algorithms.
  • #226
  • Jul 24, 2011 4:09pm Jul 24, 2011 4:09pm
Ok...When using a moving average, whichever it may be, is there any particular reason as to apply it to the open/close (?) as opposed to the high/low (chop/trend/vol.)? Just curious about this...
  • #227
  • Jul 24, 2011 4:30pm Jul 24, 2011 4:30pm
Quoting Pipalicious
Ok...When using a moving average, whichever it may be, is there any particular reason as to apply it to the open/close (?) as opposed to the high/low (chop/trend/vol.)? Just curious about this...

HIGH and LOW has advantages, you'll need both gives a nice spread which self adapts to average pip value between high and lows.
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #228
  • Jul 24, 2011 4:33pm Jul 24, 2011 4:33pm
Quoting Pipalicious
Ok...When using a moving average, whichever it may be, is there any particular reason as to apply it to the open/close (?) as opposed to the high/low (chop/trend/vol.)? Just curious about this...
You can apply close, open, median, low, you name it. Heck change the colour too, you can even make rainbows. Unfortunately newbies will always find unique ways to fail while veterans do the opposite.

Cheers~
  • #229
  • Jul 24, 2011 4:34pm Jul 24, 2011 4:34pm
Quoting Turveyd
HIGH and LOW has advantages, you'll need both gives a nice spread which self adapts to average pip value between high and lows.
What did you say exactly?
The Market pays you to be disciplined
  • #230
  • Jul 24, 2011 5:23pm Jul 24, 2011 5:23pm
Quoting Troikaone1
How can you backtest an effective MA crossover system? What does it look like and what does it look for............make me a believer..........lol!!!!
Still not convinced that it is possible to make money with a simple moving average cross-over system? Well here is a 16-year backtest on 21 different futures contracts.

Test Period 1/1/1983 to 8/31/1999
$75 allowed for commissions and slippage.
One contract per trade

US Bonds $55,940
Treasury Note $18,760
Muni Bonds $44,330
British Pound $281
Deutsche Mark $52,200
Japanese Yen $73,913
Swiss Franc $40,363
Crude Oil $19,530
Heating Oil $10,828
Natural Gas $41,500
Cotton $12,170
Coffee $32,944
Sugar $7,269
Soybeans - $19,995 (LOSS)
Corn $4,363
Wheat -$15,163 (LOSS)
Siver -$595 (LOSS)
Gold -$3,820 (LOSS)
Live Cattle $8,984
Pork Bellies -$17,740 (LOSS)
Live Hogs -$9,680 (LOSS)

Net Profit : $355,638
Maximum drawdown : $61,425

Rules :
Buy on the open when the 13-day moving average of closes crosses from below the 39-day moving average and yesterday's close is greater than the close of 40 days back. Do the reverse for short trade.

Calculate the average true range (ATR) for the past 20 days.

Use 5 times the ATR as the initial protective stop.

Every time we move 5 ATR in the profit place a trailing stop five ATR from the high of the day that penetrated that point.
  • #231
  • Edited 5:47pm Jul 24, 2011 5:33pm | Edited 5:47pm
Quoting Mr Breakout
Still not convinced that it is possible to make money with a simple moving average cross-over system? Well here is a 16-year backtest on 21 different futures contracts.

Test Period 1/1/1983 to 8/31/1999
$75 allowed for commissions and slippage.
One contract per trade

US Bonds $55,940
Treasury Note $18,760
Muni Bonds $44,330
British Pound $281
Deutsche Mark $52,200
Japanese Yen $73,913
Swiss Franc $40,363
Crude Oil $19,530
Heating Oil $10,828
Natural Gas $41,500
Cotton $12,170
Coffee...
I never said it was impossible......I simply said that I have never seen one before. I'm sure the fat cats use them and they probably don't care about a 120 pip stop loss. Thanks for the data. I would have to see what basis they use.....again.....they probably have no regard for stop losses. I also do not believe it is anywhere near as reliable or profitable as my human generated analysis over the same period . I simply do not trust "curve fitted" data. Is that curve fitted or actual results? Make me a believer....I'm not trying to be cynical...but every losing automated trading system has positive "curve fitted" data as a basis.....otherwise Tradestation would be out of business. Right? For me to be a believer, I would have to see the system play out on a live basis (I am not asking that of you of course); then I will be a believer.

P.S. - I do see the rules
  • #232
  • Jul 24, 2011 6:03pm Jul 24, 2011 6:03pm
Quoting Troikaone1
I simply do not trust "curve fitted" data.
Troikaone1, it is mathematically and statistically IMPOSSIBLE to curve-fit over 16 years of data on 21 different futures contract and with such simple trading rules!

I-M-P-O-S-S-I-B-L-E, no matter how hard you try!
  • #233
  • Jul 24, 2011 6:06pm Jul 24, 2011 6:06pm
Quoting Mr Breakout
Troikaone1, it is mathematically and statistically IMPOSSIBLE to curve-fit over 16 years of data on 21 different futures contract and with such simple trading rules!

I-M-P-O-S-S-I-B-L-E !
contender for the dumbest thing you have said on this forum so far.

you probably think you can't curve-fit some retarded astrological nonsense system either.
Congratulations ScalaFX for coming out of the closet!
  • #234
  • Jul 24, 2011 6:12pm Jul 24, 2011 6:12pm
Quoting nubcake

you probably think you can't curve-fit some retarded astrological nonsense system either.
We dare you to show us one single profitable astrological system backtested with the same trading rules on 16 years of data on 21 different futures contracts.
  • #235
  • Jul 24, 2011 6:16pm Jul 24, 2011 6:16pm
Quoting Mr Breakout
We dare you to show us one single profitable astrological system backtested with the same trading rules on 16 years of data on 21 different futures contracts.
we? you and what army?

why would i EVER bother trying to prove anything to you. it would just be wasted on you.

lame response in 3....2....
Congratulations ScalaFX for coming out of the closet!
  • #236
  • Jul 24, 2011 6:30pm Jul 24, 2011 6:30pm
Quoting nubcake
why would i EVER bother trying to prove anything to you. ..
Traduction : "I, the Great Nubcake, can say any nonsense I please, without backing up my childish claims with any proof of any kind. Those who want a proof or dare to question my ideas are simply morons. Whatever I say is true because I said so".
  • #237
  • Jul 24, 2011 6:35pm Jul 24, 2011 6:35pm
Quoting Mr Breakout
Traduction : "I, the Great Nubcake, can say any nonsense I please, without backing up my childish claims with any proof of any kind. Those who want a proof or dare to question my ideas are simply morons. Whatever I say is true because I said so".
lol. yeah you got me, haw haw. your backtest is marvelous. all hail your curve-fit. i have some tea-leaves to read...

I'm sure the next thing you will tell us is that you don't actually trade this marvelously backtested super system for reasons x, y, and z.
Congratulations ScalaFX for coming out of the closet!
  • #238
  • Jul 24, 2011 6:58pm Jul 24, 2011 6:58pm
Quoting nubcake
lol. yeah you got me, haw haw. your backtest is marvelous. all hail your curve-fit. i have some tea-leaves to read...

I'm sure the next thing you will tell us is that you don't actually trade this marvelously backtested super system for reasons x, y, and z.
The backtest above revealed that this simple moving average cross over system generated a profit of $355,638 over 198 months, or less than $1800 a month. This is hardly a "marvelously backtested super system". Of course the profits quickly exceed the million dollar mark if we trade more and more contracts as the equity curve grows, instead of using just 1 contract.

In any case, I just wanted to show our fellow traders that simple moving average systems DO work and make money in the long run, assuming the traders use a diversified portfolio similar to the one on the backtest.
  • #239
  • Jul 24, 2011 7:11pm Jul 24, 2011 7:11pm
Quoting Mr Breakout
This is hardly a "marvelously backtested super system".
that was the point i was being subtle about.

Quote
In any case, I just wanted to show our fellow traders that simple moving average systems DO work and make money in the long run, assuming the traders use a diversified portfolio similar to the one on the backtest.

sure, any system will work if you give it enough friggin years and the right ruleset. similarly i'm sure someone with enough time and effort and belief can spin a ruleset based on astrology.

but does any of it make sense? trade when this crosses here and the tide is high but only when it's tuesday and between 3 and 3.45pm... and so on. if it doesn't make sense, then it's a curve-fit. if it makes sense then it suggests a possible basis in the workings of the market and something to really sink your teeth into.

why don't ema crossovers work? why SHOULD they work? what reality are they based on? you hear all the time muppets like turveyd say how "the 'big boys' trade this MA bounce or crossover or moon cycle"... it's all bullshit they pulled out of their rears just to sound like they know something when they know sweet f a.
Congratulations ScalaFX for coming out of the closet!
  • #240
  • Jul 24, 2011 7:31pm Jul 24, 2011 7:31pm
Quoting nubcake
sure, any system will work if you give it enough friggin years and the right ruleset.
Actually it is the exact opposite. There is no way on earth that a losing system is going to make money "if you give it enough friggin years and the right ruleset".

It is like saying that with enough time each casino player is going to bankrupt the casino. Sure you can make money in a short period of time, but as time goes by the Law of large numbers takes over and if your trading system has no statistical edge to begin with you will ALWAYS end up in the poor house, even if you are Bill Gates, you can bet the farm on that.
0 traders viewing now