Indicators are like playing roulette - page 4
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
PUT UP OR SHUT UP. Walander.mt4stats.com
while this is a bad site for me to not salute EA's, one has to understand that one needs to LEARN to trade long before they learn to code --- once you find a satisfactory methodology, THEN you attempt to codify it, not the other way around.
Personally, short of "big blue" or a team of Crays, i dont think these little EA's can challenge a decent manual trader who understands the workings of the market, but if one allows for how the market works, which takes more time than you can imagine unless you have a VERY GOOD teacher, you can come close.
you must understand that the market is NOT as random as some might state, but works within support and resistance, trends, alternating time zones that reverse the direction but not necessarily the trend and a few hundred other general rules that most newer traders simply dont even know exist !
if these things did not exist, trading and EA's would be simple and profitable, but we deal with stop losses that get hit, even though the trade WAS correct, interday price reversals that you can set a clock to and the ability to recognize support and resistance areas and trends within trends within trends.
ALL of this is learnable, but how much can be transferred to a small EA remains for me to see happen --- some day, like in chess, forex will be "tradeable" successfully by computer alone, but i believe its a while off yet ---
essentially you should design an EA to work in ONE type of market and it would probably be up to you to determine WHEN youre in that market or not.
the top winners in the EA contests are TUNED only for the contest requirements and i remember one winner that simply rode a long term run (cad if i remember correctly) buying the dips and moving onwards through compounding, BUT the creator HAD TO FIND THE CURRENCY PAIR TO PLAY, not the EA !
i personally feel that most EA/s would work better if they simply allowed MUCH higher drawdowns, as any trend aligned movement in forex has its ups and downs which tend to be where the EA stutters and loses --- take a good look when testing and see how many lost positions because of small SL's actually go on to recover fully and produce profit.
Forex is an instrument that seeks to move from major support to resistance and then back again --- unfortunately, in the middle of the movement, it looks for all the world as if its gotten drunk and starts weaving and bobbing, moving up and down, BUT INVARIABLY GETS BACK TO IS SUPPORT if it left RESISTANCE ---- you can prove this to yourself any day of the week by simply LOOKING AND WATCHING.
then, believe it or not, there are indicators that CAN be of major assistance with trend development, but they have to be teamed with the remainder of the knowledge to use correctly.
as i write this i begin to wonder at some of my thoughts, because if you can clear up the drawdown issues, once the price starts its trend move, it has to pay off (as long as the drawdowns dont break you first) and its beginning to become scary !
enjoy and trade well
mpWhen you play roulette the ball can go Red or Black, there's no way to know ecxactly where it's gonna go. Even if the ball fell 15 times in a row on red, it doesn't mean that the next time there's more chances of it going black, there's alway 50% chances of it going either way. Indicators are based on past history and there's no way to know if the pip is gonna go up or down based on passed history because well, it's in the passed and indicators cannot foresee the future. IMHO indicators are there to calculate the odds of future movement. The trick is to find an indicator that can calculate those odds the best. I'm starting to think that trading forex based on the news would give me better odds and I wonder if people trading forex with indicators or EA ONLY can be are successful or one has to trade using the news as well ? Please start the discussion
And this is ours common mistake, believing that we can trust indicator ! We have to trust the price movement and only that. believing that you can be more clever than the market mouvment and knowing where he will go like you can be in is head is just the way to loose, simply because we are too small.
We are the small fish which just follows the big ones. We have to learn to follow the market and not trying to imagine where he will go.
NO T/P and NO S/L
It fits into a theoretical model very easily. I dont particularly like Igors equal SL/TP constraint on the test, I understand why it wants it there, but in practical terms why burden yourself with unrealistic constraints that are not required.
In simple terms, you can compare the performance of trading systems, by comparing the distributions of gains and losses. Adding some indicator either does nothin, or it leads to a change in the distribition of gains and losses, either for better or worse.
Actually removing the stop and target constraints is quite sensible during testing cos it lets you test the thing you wanna test, without the results being undulty influenced by external factors. Otherwise you end up with Igors rather good horse race analogy.Hi Zupcon,
Lets TRY to continue this debate about PREDICTING FUTURE DATA.
Beside the test that S/L and T/P are equal to proof that one can not come up with the consistent hitrate higher then 55% I did also other tests.
Let me explain some other apraoch I did (that is not in the horse race analogy)
This aproach was not using any S/L nore T/P.
When one starts of from the idea that he wants to predict future data then automaticly you arrive at trying to predict the very first bar. Reason for that is that it would not make a lot of sence to say: I can not predict the first bar with a reliability bigger then 50% but I can do that for the 4th bar or the 11th bar.
At first one would think that predicting only 1 bar is not very intersting but if you think in terms of doing that on a daily or a weekly chart then one should realize that in a daily or a weekly bar there is a lot of potential.
If you take whatever kind of indicator or combination of indicators or set of rules (to talk easy again lets call it again a "system" ) and you make the test in such a way that when the conditons of the "system" are met then it takes in a position on the opening of the next bar and closes the position at the closing of that very same bar.
If the system is still in place on the closing of that bar one takes in again the same postion on the opening of the next bar and close that position on the closing of that bar. So that means that it will open and close on every single bar a position. Or long or short depending on the direction that the "system" shows.
It can be that a certain bar will have a range of only 1 pip between opening and closing and some will have for ex. 400 pips between opening and closing. So the amount of pips one would loos or would win is now completaly irrelevant and of no influence in this aproach. If one looks over a longer period of time in a random way the risk and reward are the same in an aproach like this. Again is to see if we can find a "system" that can produce a hitrate that would be systematicly higher then 55% in predicting the direction of the next bar.
The reason or filosophy behind this setup is that if a system shows that the trend is up then the system will open and close a long position on every bar but because the trend is up it should so calledcatch more long bars then down bars.
Well on this setup (without T/P and without S/L) the very same conclusion can be made.
With all possible combinations of indicators simple or complicated one will find with a "system" with a certain set of parameters over some period and over a certain pair, a hitrate that is clearly more then 55% (remember more then 55% we accept as proof of predicting future data).
But if one places this very same setup (with exact same parameters) on other pairs and over other periods then the avg hitrate falls back to 50% so no more proof that a system can PREDICT FUTURE DATA. Or that data comes in the same way as red or black on a roulette table.
I also added a spread sheet that shows you what the result WOULD BE if one is indeed cappable of predicting a single next bar with a hitrate that would be 55%.
There is a setting to fill in the hitrate. It is on 0.45. That is because in the spreadsheet you need to fill in the negative hitrate. So 0.45 is taking a 55% positive hitrate in the calculation. If you would place is on 0.60it is going to take only a 40% positive hitrate in calculation.
So I suggest to leave the number on 0.45 (that is making a calculation of a postive hitrate of 55%)
If you now click the "new sequence" button the spreadshet is going to take in longs and shorts in a completaly random way but according a hitrate of 55% on the opening of a bar and close it on the close of that very same bar. So no S/L or T/P are used.
One would think that you can't do that because for all the same you go long on a bar and it is actually a down bar that would loos as much as 500pips. So that this way the risks would be way to big.
But if you click 10 times on that button you gonna see that 8 out of 10 equity curves make clearly profit. The data was taken from a weekly chart on the eur/usd between 1990 and 2009 ( make sure that your macro setting in excel are correct otherwise clicking the "New Sequence" button will not work).
This shows also clearly the power of having "only" a 55% hitrate. By power I mean that 55% clearly proofs that this is a more then suffient edge on the normal 50% and on that moment one can say he is predicting future data. But sad enough no real system can systematicly do this.
If your interest is still there I gonna explain an other aproach in a next posting.
Friendly regards...iGoR
It fits into a theoretical model very easily. I dont particularly like Igors equal SL/TP constraint on the test, I understand why it wants it there, but in practical terms why burden yourself with unrealistic constraints that are not required. .............
Zupcon,
If you would have any suggestions or methods on how to proof predicting future data, I would happely like to hear about it very much.
But to stay on the same level or talking the same language on giving proof of prediction we need to come to higher then 55% and the risk and reward need to stay NATURAL. Meaning one can not start to use small synthetic T/P's combined with big or no stops.
ex. Reports PrizmaL - Automated Trading Championship 2008
Here you see the results of the guy who ended up 2nd in the EA metaquote championship.
If you look to his hitrate he has a 93% hitrate witch looks fantastic at first sight.
But if you look to his risk and reward or his avg winning trade and his avg loosing trade then you see that the avg loss is 7x times! bigger then his avg winning trade. So he is FORCING his hitrate up by using small TP's and big S/L's. The hitrate is much higher then 55% but is no proof what so ever that he can predict price moves with an accuracy of more then 55%.
Even more if his hitrate would drop to 87% (witch still sounds as an extreme hitrate) his system is already losing money.
ex. 87 winning trades x 306$= +26.622$......13 losing trades x -2188$ = -28.444$....... -28.444 + 26.622$ = -1822$ loss
Or a bit like steinitz (that you knew also very well). He had also a hitrate of 99% by using no S/L what so ever but his floating losses were burning every account on the long term.
So as said the reward and the risk must stay NATURAL. One can not use any synthetic influences by placing T/P's.
Hope to hear some suggestions from you.
Friendly regards...iGoR
Actually I posted this thread in the general discussion forum so we can talk about anything as far as I know.
I have no other "job" halfasleep, but THIS is not the place to discuss anything but indicators -- joeschmoe created a thread for those to ask questions and throw stones, so PLEASE can we stick to the subject of indicators here ? mp
is 60% any good?
hi guys!
i have back tested several systems. which are simple and easy to follow they have a 1:1 risk reward ratio. and there thers hit rate is always usually about 60% - 63%
my question is; is that any good? and what money management strategys would u recommond to increase the returns
thanks
When you play roulette the ball can go Red or Black, there's no way to know ecxactly where it's gonna go. Even if the ball fell 15 times in a row on red, it doesn't mean that the next time there's more chances of it going black, there's alway 50% chances of it going either way. Indicators are based on past history and there's no way to know if the pip is gonna go up or down based on passed history because well, it's in the passed and indicators cannot foresee the future. IMHO indicators are there to calculate the odds of future movement. The trick is to find an indicator that can calculate those odds the best. I'm starting to think that trading forex based on the news would give me better odds and I wonder if people trading forex with indicators or EA ONLY can be are successful or one has to trade using the news as well ? Please start the discussion
The chances of winning of a indicator based system is same as flipping a coin. Traders wannabes keep chasing these indicators and think some magic combination will lead them to the road of riches.
I've explained this in lot of detail here.
MP -- Indicators are simply price action on a horizontal plane
While its understandable that everyone thinks they might re-invent the wheel, studies seem to bear out that it doesnt need much re-designing to continue serving its purpose, but of course there is always someone out there with a better idea !
INDICATORS are used because indicators WORK --- simply stated, if they didnt work, no one would use them !
Indicators, along with candle shape and price action are simply the same things, but using different forms of representation.
THIS is the unfortunate thing that modern day adherents of "this" or "that" appear to miss --- EVERYTHING is simply a way of showing how price moves or even more simply, FOLLOWING how price moves.
So the next time you wish to put down indicators, understand that you have to put down anything you may be using also, since EVERY MOVE OF AN INDICATOR is based on price movement, as is EVERY MOVE of anything else.
enjoy and trade well
mp
The chances of winning of a indicator based system is same as flipping a coin. Traders wannabes keep chasing these indicators and think some magic combination will lead them to the road of riches. I've explained this in lot of detail here.
prediction
p-r-e-d-i-c-t-i-o-n is like observe a newly acquired indicator -- to see there is any universal sign
sometime, it is just a pure luck that we put two indicators together, draw something on top, then we got a sign that we could bet on --- even give when to do SL or HEDGE lock
then we realise it only work in some of the days -- when the markets do not have much NEWS breakout after our bet
will prediction work , I don't know -- but I am into indicator with some form of prediction -- for price level (to sell or to buy ) already
hard to find those, as most indicator are into reflecting the current historical data -- I can only tell , when I make a consistent profit (not quite there yet)
PUT UP OR SHUT UP. Walander.mt4stats.com
Werander,
You must like mp, her measurments, body, and bongo-bongo.
Bongo