EA opens correct on round numbers as you can see in myfxbook: image But there are some seconds till published in website and pushed signals to subscribers. Therefore might be 1 or 2 pips difference. Anyway each broker can have different quotes..

Notable numbers strategy

Here some update on TensorFlow testing vs Python. Training data EURUSD from year 2000 H1 candles contains entries with fixed SL/TP with risk/reward 1:1 so good result would be at least 51% but better 52-53% just in case. After epoch 7 results for ...

AI talking

And here breakdown how each asset performed from our portfolio of 10 assets: image The biggest loser is EURCAD.. according to the analysis here, it is ok to have 3 losing deals in a row. This trades with risk/reward 55/65 therefore no wonder of ...

Notable numbers strategy

Nice run-up for notable numbers strategy: image 74% profitable deals: image in 6 months 85% in profits: image Is anyone of the subscribers making deals with signals? Maybe someone built an EA and get similar results? Keep in mind.. according ...

Notable numbers strategy

I did myself neural network in python with TensorFlow. But spent not much time therefore maybe can't expect results. Maybe you talk about this TensorFlow python solution for stock price prediction? I implemented it, made it running with Python 3.11 ...

AI talking

It is needed min 3 years track record to attract investors money. For example here 1 trader account operating more than 4 years now attracted 16mil usd capital. As we can see it had 6 months in drawdown. image 3 to 5 years trading to get ...

AI talking

Won't use python AIs.. no stable results found. I will use the simple neurons which is possible to implement directly in the MQL5 expert advisor and use optimizations to find neuron weights.

AI talking

Here how AI drawdowns look.. 80 days sit in DD.. that's real trading image with 100 usd capital DD 30% - real life too. here example of one of my live accounts trading >3 years now, was 2x sitting in ~6 months DD. That's real life when ...

AI talking

Finally found the error - it was CSV file export for training data not correct. It included the "current candle" already in prompt data therefore AI was so easy to "learn" to spot that pre-made entry. Fixed. And now all look much more realistic... ...

AI talking

Yep...

AI talking

Current week. Compared with demo trades made this week and noticed that demo now in dd -8usd but all those 4 trys on same week ended up in profits. 1 week is too small timespan to conclude something.

AI talking

Did that.. for days till 16 February included it produced various results not seen at demo account. Each time AI is executed it produces different signals. file file file file When try enough times I bet will get some loosing too.. AI1 I wont ...

AI talking

Here are historical stats for AI1 regarding drawdowns time: image As we can see it is ok to be in DD for 1 entire week.

AI talking

AI1 results so far: image Was 8 losing serie.. max in tests was 6.. not good. Total dd is still acording to tests. AI2 detected errors fixed yesterday therefore no enough correct data. For 1 day 66% profitable..

AI talking

Strategy keeps working: image

Notable numbers strategy

Here AI2 (with Random Forest Regression algorithm) forward tests results (prepared signals CSV with Python and then executed in MT5 with real ticks): image file It is what it is... cant comment.. I know I know.. Demo ...

AI talking

First AI version traded on Demo account allready 38 hours. Did 17 trades. 8 profitable. Was losing serie 5 in a row. Win serie 7 in a row. Noticed that Python script execution usually takes 3-4 seconds. Was once 18 seconds and 9 seconds. Maybe on a ...

AI talking

made first comparision DecisionTreeRegressor vs RandomForestRegressor image image huge difference in correct predictions %.. and 2x fewer deals which is promising.. will prepare additional demo account and run in parallel those 2 AI models..

AI talking

AI itself

AI talking

here I continue discussion with OpenAI regarding AI trading.. my question: but what will AI predict if in training data sometimes an pattern signals BUY and sometimes SELL as ideal answer? if using DecisionTreeRegressor then it will return what ...

AI talking