This topic has been reported for inappropriate content
- This topic has 455 replies, 41 voices, and was last updated 8 years, 10 months ago by
.
- Posts
Hi Anti
Just a small mistake.. don’t call your self using that word..
actually.. at first glance.. I think the same thing as you too..
tankhang
Hi bharath4x
Normally.. (during news hour)
when the spread is widening.. there is something wrong with the market..
the ‘high volatility’ may cause the price become extreme..
and the indicator value will become extreme and less reliable..
It is better not entering the market at that moment..
tankhang
Hi bharath4x Normally.. (during news hour) when the spread is widening.. there is something wrong with the market.. the ‘high volatility’ may cause the price become extreme.. and the indicator value will become extreme and less reliable.. It is better not entering the market at that moment.. tankhang
Hi tankhang,
Thanks for your reply.
My Point is, we would be placing pending (Limit/Stop) orders X pips below or above current price and there will also be other pending orders waiting for the price to reach.
Price needs to go either up or down, if so, are we not going to miss an opportunity/ies.
Please correct me, if i’m wrong.
Thanks!
This is our MAIN indie, so it deserve its place here, untill concluded and we can agree on stable indicator… G.
What about the one posted by @Lowphat here which is already part of the v.1.2 ?
When forward testing I didn’t notice any repainting, and peak through areas match with simplex indicator so the logic should be also good.
Anybody agrees or disagrees?
V.
Agree:
Anybody agrees or disagrees?
The one @Lowphat has posted is giving the same signals as my current WIP version. So as long as we’re using iCustom to test why not take that one?
It should be possible to derive a function for direct EA use from only one of our approaches, no matter which one. The logic of the algorithm demands to keep track of current and previous swing state as well as highest low and lowest high. That’s all, sounds simple. In an indicator, you can do so declaring additional buffers. Costs some memory, but who cares, right? In an EA, you got to find another way. That’s what I’m currently working on.
Never had the idea that such a simple logic could be such a nuisance ;-} I wouldn’t mind if somebody came up with a coder’s stroke of genius and posted it over the holidays.
Maybe I just need a break from the terminal and my coding editor.
Yesterday evening I re-read Ron Black’s article about that Clear Method. We’ve had some talk about 4-bit vs. 3-bit logic here. The EL code published in Black’s article is working on a pure 2-bit logic, so there’s no undefined swing state. This means, that you have to start with an assumption about the initial swing direction. This assumption will be approved or refuted during the following bars. After the first change of swing direction, it does not matter anyway what your initial assumption was.
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
No, you’re not:
I’m an idiot.
I’ve read so many good posts of yours, I’m sure you’re not!
s. [/quote］Agree with simplex … if Anti can understand R good, and he is and idiot, then what about other people that do not understand R ?
@simplex, can we create Dot or Arrow OBJECTs on Chart Window (not at separate window), so we can see wether the indicator repaint or not at some points better?
Attached… G.
Thanks a lot G. Wishing you and your family, all penguins and their families, a Merry Christmas and very happy, healthy and succesful new year 2016 …
No, you’re not:
I’m an idiot.
I’ve read so many good posts of yours, I’m sure you’re not!
s. [/quote］ Agree with simplex … if Anti can understand R good, and he is and idiot, then what about other people that do not understand R ? @simplex, can we create Dot or Arrow OBJECTs on Chart Window (not at separate window), so we can see wether the indicator repaint or not at some points better? ClearMethodSignals_01 indicator (post #10824) does exactly that.
“x” denotes swing point (repaints while bar active, static after bar close).
Arrow denotes area breakout, same behaviour (repaints while bar active, static after bar close).
G.
Wish you merry Christmas, Happy Holidays and a Happy New Year!! G.
Thank you @GG53. Happy holiday season to you too and to all the members.
V.
ClearMethodSignals_01 indicator (post #10824) does exactly that. “x” denotes swing point (repaints while bar active, static after bar close). Arrow denotes area breakout, same behaviour (repaints while bar active, static after bar close). G.
You are right G. Sorry, not tested it before sent my post. Thanks for your indicators
I’ve noticed meta-trader back-tester bug. When there is a gap from Friday to Monday. An order is opened on Monday, but the open price is taken from Friday and it results as TP.
I’ve noticed meta-trader back-tester bug. When there is a gap from Friday to Monday. An order is opened on Monday, but the open price is taken from Friday and it results as TP.
Yes, Strategy tester in MT4 is not reliable for backtesting.
Sometimes we had to download chart data from Metaquotes.com which will not sync with broker’s data.
Sometimes subsequent runs of same EA with same input parameters provide different outputs.
Please advice, if there are other efficient way to backtest our EA’s.
Thanks!
I read somewhere Metaquote data is the median price of _many_ brokers ticks data. If youoptimise you EA with this data you could have the EA running on MetaQuote Demo server and copy the trades with a trade copier to another MetaTrader instance running with your desired live broker. I know people do it this way but it’s not that smooth.
Bette to download 99,9 tick data with Tick story. I think this data is pretty reliable to optimize EAs with. Don’t for get to change settings for data before downloading like spread, minstop levels, swap etc..
I think the difference with downloaded ticks data and Meta quotes tick data is huge. Many times when I tested my experts in the strategy tester it could be profitable with meta quotes data and then when switching to the data downloaded from tick story its not profitable at all. I believe volume[] data is not included in this tho so could be trouble running indicators using it?
Is there another/better way testing besides going demo or live?
MT4 BackTests relies on actual data (if exist…) of Open, High, Low, Close (OHLC) of each bar.
EVERYTHING that is missing from the above is Randomized, which means all ticks within that data are randomized.
The MAX accuracy that you can get is getting perfect history of M1 OHLC of that test period and TimeFrame. Even that is hard to get.
On top of that, you will NOT get variable Spreads, but constant one throughout the whole test period.
VOLUME data is broker depended, and only relates to OHLC.
So… if your trade entries are in mid-bar, or trade exits in mid-bar (or both…) – your BackTests are questionable, to say the least.
BackTests are good for Algo Logic checking and debugging, but not much more than that.
G.
Ok, hope that this is working now. At least I did not manage my new Clear Method or ‘Peak & Trough’ indicator to display anything unstable.
Main difference as compared to my previous approaches: I allowed for a certain amount of redundant calculation instead of shifting historic values around in static arrays. Makes life easier from time to time if you’re throwing textbook coding paradigms over board, when it makes sense.
This time the whole Clear Method algorithm is contained in one single, self-initializing function getClear. So inside function
OnCalculate(), and later also in an EA, the complete logic can be reduced to one single line of code:CLEAR_STATE clearState = getClear(symb, tf, shift);Please note that the function will return a custom enumerated integer of type CLEAR_STATE. See source code for details, and link for conceptual information in case you’re not accustomed to using enumerations.
Following remarks on the algorithm:
- Calculation starts with an arbitrary assumption of status CLEAR_BULL, and with a lookback period determined by CLEAR_INIT_DEPTH.
- Initialize data for first bar within lookback period, then check all following bars for a change of clear status. If at least one change has been detected, leave loop at index 0 and return clear status of latest bar 0.
- If no change has been detected, increase lookback period and start all over.
- We want at least one change of status, because we cannot know whether our initial assumption CLEAR_BULL was correct.
If calling the indicator by iCustom for backtest, you can choose buffer index 2 (not drawn on chart) for a direct access to output +1 for upswing and -1 for downswing. Buffers 0 and 1 are for visualization in green and red histograms.
Not much more to say. Have fun testing!
@VlanFx: any question welcome regarding EA integration, but should be pretty simple!s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
on sx_ClearMethodSimple_v1_2.mq4 the signal from swing low to high and swinghigh to low on my test works perfect. nice job
“I believe the very best money is made at the market turns. Everyone says you get killed trying to pick tops and bottoms and you make all
your money by playing the trend in the middle. Well for twelve years I have been missing the meat in the middle but I have made a lot of
money at tops and bottoms.”
– Paul Tudor JonesHere is version v.1.3 of our Penguin EA
Added changes were made based on @GG53 ideas in post here and here
Still missing:
– FFCal – avoid trading before/after news. Emonts and jaguar sent me some version of ffCal library, but I couldn’t get them to work. See here.– Limit orders – functions for placing LimitOrders is included in the EA, but after talking with GG they are not being called yet.
Known EA limitations
– for partial orders closing logic to work properly, TrailingStop parameter must be bigger than PartialTP parameter (take profit value at which part of the order is closed). We can discuss more about this in the coder’s chat thread.V.
I just started to look at the topic of news calendar integration.
Still missing: – FFCal – avoid trading before/after news.
A lot of rubbish code seems to be underway in WWW regarding news … indicators that don’t display anything, or don’t even load, or jamming my screen with millions of alerts.
I found these to work reliably on my system: here and here, regarding FF calendar and DailyFX calendar respectively. Appear to be very close to Derk Wehler’s original, while ‘special’ adoptions seem to cause problems more frequently.
I’m starting to work through the code now, preparing for my mods to come.
So my question: what are our requirements precisely for news calendar integration in our EA?
- a) functionally
- b) technically regarding programming interface
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
I just started to look at the topic of news calendar integration.
Still missing: – FFCal – avoid trading before/after news.
A lot of rubbish code seems to be underway in WWW regarding news … indicators that don’t display anything, or don’t even load, or jamming my screen with millions of alerts.
I found these to work reliably on my system: here and here, regarding FF calendar and DailyFX calendar respectively. Appear to be very close to Derk Wehler’s original, while ‘special’ adoptions seem to cause problems more frequently. I’m starting to work through the code now, preparing for my mods to come. So my question: what are our requirements precisely for news calendar integration in our EA?- a) functionally
- b) technically regarding programming interface
s.
Functionality.
We want to call that function (?) with user params like BeforeTrade/AfterTrade (minutes), Medium/High impact, Currency pair relevance.
G.
Ok:
Functionality. We want to call that function (?) with user params like BeforeTrade/AfterTrade (minutes), Medium/High impact, Currency pair relevance.
BeforeTrade/AfterTrade (minutes) means number of minutes to next event and number of minutes after previous event?
A Function? Would be nice. The indicator has got some nice features, but the overall architecture appears to be more or less monolithic. Will take a while to convert this into a set of functions that make sense.
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
Hi,
Today I gave VlanFx an update of the check news procedure/routine.
I use this procedure in a db I use, see attachment.
The call is very simple:
int result = getNewsInfo(“EURUSD”, newsMinutesBefore, newsMinutesAfter);
result = 0, no news
result = 1, high impact news
result = 2, medium impact news
result = 3, low impact news
Ofcourse VlanFx has to confirm, but i think the item is solved and will be implemented in next release.
Br
Huub
Sorry wrong picture
Hello friends,
Let me read the whole thread first…
Best regards and happy new years…
Nothing has ever motivated me more than this...
s. [/quote］
I found these to work reliably on my system: 
- You must be logged in to reply to this topic.