Differences

This shows you the differences between two versions of the page.

--- forex:probability [2022/12/15 11:14] – peter
+++ forex:probability [2022/12/15 11:26] (current) – peter
@@ Line 1: / Line 1: @@
 ====== Forex - Probability ======
+Trading is thinking in probabilities and finding setups that make money.
+----
+===== Positive Expectancy =====
+  * You may not make money on the trade right now, or even the next one, but if it makes money over the long run (has positive expectancy).
+  * You could have a system which is 35% accurate which still makes money (and a lot of it) over time.
+  * Get comfortable with uncertainty, and comfortable with losses (which are inevitable), then you will find yourself making good decisions more often than not.
+----
+===== Probability =====
 <code>
@@ Line 44: / Line 61: @@
 <WRAP info>
-**NOTE: This is slightly less than the original value of 6/9.
+**NOTE**: This is slightly less than the original value of 6/9.
   * This method is called Laplace smoothing.
@@ Line 51: / Line 68: @@
+----
+===== Smoothing using Weights =====
+<code c++>
+double value=price[i+center], // Price value at the center.
+       max=_Point;            // Maximum deviation.
+for(int j=0; j<period; j++)   //Calculate price deviations from the central one and the max deviation.
+{
+  weight[j]=MathAbs(value-price[i+j]);
+  max=MathMax(max, weight[j]);
+}
+double width=(period+1)*max/period,  // Correct the maximum deviation from the center so that there are no zeros at the ends.
+sum=0,
+denom=0;
+for(int j=0; j<period; j++)  // Calculate weight ratios for each price.
+{
+  if (Smoothing==Linear)  // Linear smoothing.
+    weight[j]=1-weight[j]/width;
+  if (Smoothing==Quadratic)  // Quadratic smoothing.
+    weight[j]=1-MathPow(weight[j]/width, 2);
+  if (Smoothing==Exponential)  // Exponential smoothing.
+    weight[j]=MathExp(-weight[j]/width);
+  sum=sum+weight[j]*price[i+j];
+  denom=denom+weight[j];
+}
+buffer[i]=sum/denom;  // Indicator value.
+</code>
+----
+===== References =====
+https://www.mql5.com/en/articles/11627