DislikedAnyone else got trouble logging in? or have my curiousity banned me once again,,,Ignored
2
which fibonn was vegas talking about? 1 reply
Changing Your Self Talk By Talking to Yourself 2 replies
My left brain starts talking to me 7 replies
talking to myself(smile) 8 replies
DislikedAnyone else got trouble logging in? or have my curiousity banned me once again,,,Ignored
DislikedLooked at bit interesting for a moment. It's unfortunate that Microsoft got involved with this project. As usual they are going to ruin everything with their stupid corporate policy. Already you can see they forced the developers to put too many limitations and "safeguards". Not to mention the insufficient CPU power. They are doing this on purpose! “Chat GPT is at capacity right now” yeah sure. What a joke! They just got a billion investment. Guess who is calling the shots now..... Should at least give OpenAI enough CPU power and set it free with...Ignored
DislikedExciting results.. turns out self learning AI is possible and the longer AI trades/learns the better results it provide. how I tested this: 1. Made 45000 H1 candles training data 2. Split data by 100 H1 candles so AI trained first iterations with 1000 H1 candles, then forwardtest H1 100 candles (~4 trading days) 3. After every 100 H1 candles retrain AI model - add new candles to training data so training data every iteration enlarges by 100 Here full AI self-lerning iterations audit: {file} Here some insights.. In the beginning, we can notice bad...Ignored
DislikedExciting results.. turns out self learning AI is possible and the longer AI trades/learns the better results it provide. how I tested this: 1. Made 45000 H1 candles training data 2. Split data by 100 H1 candles so AI trained first iterations with 1000 H1 candles, then forwardtest H1 100 candles (~4 trading days) 3. After every 100 H1 candles retrain AI model - add new candles to training data so training data every iteration enlarges by 100 Here full AI self-lerning iterations audit: {file} Here some insights.. In the beginning, we can notice bad...Ignored
DislikedThe model now categorizes the data and you say that some data is sufficiently directional for you. Then when you add equal size SL and TP to use in conjunction with the model - you essentially start working with volatility only. So the way you implement your model makes it predict sufficient directionality of volatility. Is that what you intended to do?Ignored
QuoteDislikedIt is not accurate to say that the AI predicts volatility. The AI categorizes the data into one of three categories (SELL, WAIT, or BUY), based on the patterns it learned from the training data. The addition of equal size SL and TP to use in conjunction with the model can make it predict the sufficient directionality of volatility, but this is not the primary goal of the AI.
QuoteDislikedIn essence, the AI has learned the behavior of the financial asset's price movements based on the patterns present in the training data. Whether or not the financial asset's price moves in unknown patterns, the AI will make predictions based on the patterns it has learned from the training data.
DislikedIn the meantime, backtest further in the past? You only did 1 year of hourlies. Should be pretty easy now, just load older data, right?Ignored
QuoteDislikedThe AI will predict the most frequently occurring answer in the training data for that particular pattern. If both "BUY" and "SELL" occur with equal frequency in the training data for that pattern, then the AI may predict either "BUY" or "SELL" randomly. However, the decision tree algorithm is deterministic, so the same inputs will always lead to the same outputs.
QuoteDislikedIn this case, you could consider using a different machine learning algorithm, such as a Random Forest Regressor or a Support Vector Machine (SVM) Regressor, which are more capable of handling non-linear relationships between features and target variables. Another approach could be to try using a neural network, such as a Multi-Layer Perceptron (MLP) or a Convolutional Neural Network (CNN), which are often used for pattern recognition and can handle complex relationships between input and output.
Dislikedhere 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 prediction? randomly select? answer: {quote} and here OpenAI suggests to use different algoritms... {quote} seems to need to play around with those too... a lot of analysis needs to be done here.Ignored
DislikedAnd here 1st signal for gbpusd next week: {image} So if gap is up then trend up will establishes..Ignored