Trading

In this thread from time to time I will add some knowledge. For me it's just a way to organize/store data and make it accessible to others.

Intra-day Trading Cycles
https://www.forexfactory.com/showthread.php?t=728185

Mark Douglas Trading Psychology Workshop

If I remember correctly then these are simple statistics on EURUSD from D1 H/L/O/C data.

Video Trading Journal
Link: https://www.forexfactory.com/showthread.php?t=835092

Why you should not use Martingale position sizing technique!
Here is the math:

First lets establish some rules. Lets say that possibility to win in each separate trade is 50%. Now There should be limit to how far you are willing to go with martingale, but for the sake of example it will be low, that is , 4 losing trades in a row to start position sizing from lowest again. There will be 3750 trades made.
So if we are only willing to take loss 4 times in a row then
There will be 2000 trades where position size for each trade is X, 1000 trades where position size is 2X, 500 trades with position size is 4X , and 250 with position size 8X.

Now from those trades half of them will be losing ones, this means that you will lose
1000 trades with X size , 500 trades with 2X size , 250 trades with 4X size and 125 trades with 8X size.
In this case you will get out beak even

Now here is the most important part

This win ratio is made from winning and losing trades that are unequally distributed in the trading sample, thus there are periods where you win more times in a row and periods where you lose more times in a row. If you have large enough sample size as in this example , that is 3750 trades and only increase your position size maximum 4 times then that will not impact the results.

Now here is the thing for smaller trading samples sizes , even 50% win ratio strategy might not produce 50% win ratio in 10 trades sample, it might be 70, 80, 40, 30% , most likely it will not be 50% .
This also means that when you are using Martingale, different position sizes must be separated into groups to calculate win ratio for each group separately.
Now if we use Martingale then in perfect scenario win/loss ratio should look somewhere like this.

1. 500 trades total (size X)-250 won;250lost
2. 250 trades total(size 2X)- 125won;125lost
3. 125 trades total(size 4X) – 62won;62lost
4. 62 trades total(size 8X) – 31 won ;31lost
5. 31 trades total(size 16X) - 15won;15lost

In reality starting from 4th point where you have 62 trades total you should start to worry, you might not end up with the same win ratio as your strategy showed. As As amount of trade decreases problems start showing up. Starting from this point 50/50 win ratio could look like 55/45, if position size is increased then difference might get even larger – 30/70; further if size is increased it becomes even more larger 10/90 could even be possible(even with 50/50 win ratio strategy).

Solution
Instead of using martingale it is better to keep constant position size ,thus improving your overall results.

Be consistent and do not use martingale because it makes your system inconsistent.

What to look for in strategies where you wait for consecutive losses to increase your edge in following trade?

1.Formula for coefficient - Minimum Coef(MK)
MK=Win% / Loss%+1

Example, win/loss probability is 50/50.
MK=50/50+1=2
This coefficient will be used to look for hidden edges in strategies.

Theoretically when dividing possibilities when consecutive loss increases ,possibility will decrease MK times. Example,
Win/loss ratio is 50/50. Thus MK is 2.
Possibility of 1 consecutive loss is 50%.
Possibility of 2 consecutive losses is 25%(0.50^2)
Possibility of 3 consecutive losses is 0.50^3 = 12.5%.

As you can see when consecutive loss count increases by 1 , chance of that event occurring decreases MK times.
In practice these coefficients may differ. If it is larger than theoretical coef. then you have an edge. In other words , if possibility of X consecutive loss is smaller than mathematically calculated then this is ,as I called it, "hidden edge".

This idea is not easy to grasp, easiest example would be explaining through grid trading (when tp distance is = to sl distance and both equal to distance between grids).

Theoretically these could be the possibilities
Possibility of 1 consecutive loss is 50%.
Possibility of 2 consecutive losses is 25%
Possibility of 3 consecutive losses is 12.5%
Possibility of 4 consecutive losses is 6.25%

But lets say that from back-testing you have calculated that these are possibilities for this strategy
Possibility of 1 consecutive loss is 50%.
Possibility of 2 consecutive losses is 25%
Possibility of 3 consecutive losses is 12.5%
Possibility of 4 consecutive losses is 2.10%

As you can see there is a difference for possibility of 4 consecutive losses in row.
Theoretically showing 6.25%, but in reality showing 2.1% possibility of occurring. This means that if you only open position after waiting for 3 signals to fail ,then you will have an edge, in this case with
(6.25-2.1)/6.25= 66.4% winning chance.

I made a pool in FF, unfortunately it didn't get large enough member interaction.

Overall it seems that most of the retail traders are using "Small" time frames when setting SL levels, that is 15M and 1H.
Pool had only 13 votes so results are not accurate.