Archives

# Michael

Michael has written 39 posts for RaidenWorks

### Confusion Matrix Confusion

I was debating with myself whether this was too basic to bring up but after writing it, I’m glad I did. Referencing scikit-learn’s confusion matrix documentation, it’s not immediately apparent what convention the axes follow. When the confusion_matrix function is called, we get the following output: So which value corresponds to True Negative, True Positive, … Continue reading

### Plotly makes great interactive visualisations

Data visualisation with Plotly R library Continue reading

### MQL4: Get the highest price of a set number of periods, candles, bars

High[x] returns the high price of the period at shift x. https://docs.mql4.com/predefined/high iHighest() returns the shift of where the highest price is observed for the defined periods. https://docs.mql4.com/series/ihighest Combine them to obtain the highest price for the defined periods: double highestPrice = High[iHighest(Symbol(), 0, MODE_HIGH, 10, 1)]; In this example, the last parameter being “1” … Continue reading

### MQL5: CopyTime() vs SeriesInfoInteger()

To obtain a period’s open time, we can use either CopyTime() https://www.mql5.com/en/docs/series/copytime or SymbolInfoInteger() https://www.mql5.com/en/docs/marketinformation/symbolinfointeger I’ve attached a script testing the speed of each showing that SymbolInfoInteger() is faster than CopyTime() by 50%, even after optimisation.     Just throwing it out there.

### MQL5: Getting your EA to manage its own orders with others around

‘Selecting’ a Position “copies data about a position into the program environment, and further calls of PositionGetDouble(), PositionGetInteger() and PositionGetString() return the earlier copied data. This means that the position itself may no longer exist (or its volume, direction, etc. has changed), but data of this position still can be obtained. To ensure receipt of … Continue reading