Doing some Basic Data mining

Was going back to the basic's and doing some Data Mining on the EUR/USD 1hr. Just some simple things to see if I could calculate it in MT4.
The result of this test was somewhat confusing

I was checking to see on the hourly charts the percentage a green bar is followed by a green bar and a red bar is followed by a red bar.... Why just bored I guess.

We would all expect hey green is normally followed by green and red normally followed by red (or at least I would) but that doesn't seem to be the case.

I believe this information just shattered my world what do you all make of it??

Oh and the Duration was from 2012.11.27 to current

//----

hey max.... roughly speaking, on every pair on every timeframe, the results will be the same.... there are some notable exceptions such as the gbpjpy....

realistically speaking when looking at a very large amount of bars,

the % green bars will be almost == to the % red bars.....

when using the total number of green bars as the divisor, a green bar will be followed by a green bar roughly 50% of the time, +/-2%....

when using the total number of red bars as the divisor, a red bar is followed by a red bar those same %'s.....
//------

when using the total number of bars as the divisor,

a green bar will be followed by a green bar roughly 25% of the time, +/-4%.... the reduction is due to divisor now be twice as much....

a red bar will be followed by a red bar roughly 25% of the time, +/-4%....
//----

the number of bars under test is obviously very important.... and again, those results will be roughly the same on every pair on every timeframe..... +/- a few%....
//-----

don't stop there...... expand your data mining.....

what is the % of times an up bar follows qqe crossing above the 50.... on every pair.... on every timeframe.... and average pips...

how about rsi crossing the 50.... or macd crossing 0.... or tsr color change..... or cci crossing 0 .... or... or....

you will be surprised by the results.....h

H
Your right this is just all very odd I was up until 2am crunching different numbers then realized my family had went to bed and it was 2am.

For example the RSI cross you mention its even the opposite of what we would expect for example the hourly attached EURUSD.

Some of the other timeframes where I have more bars it gets to about 49% but still I would think we would see.

I'll have to write something a bit more complicated when I get some time to see the swing averages after some of indicator strategies we see. To check if they are even viable or if they a counter to what we would expect too

Here's the same thing with avg close pip difference.... If we could somehow use the 1% probability to our favor with the avg closes to profit is maybe something to look into ?? Still just Monkeying around with it but it's just so interesting. Complete opposite of what I would expect lol

Might not be as interesting if your not a big nerd like me :nerd: but just playing the odds wonder if I could make something like this profitable lol.

But with Spread + Commission is a 10 pip move with 1% in our favor going to be profitable on average with 10 pip profit...... Maybe... Wish I could think of a formula for this or if it would just be quicker to write the backtest

Punched them into the spreadsheet real quick easier then coming up with a real formula
Looks like even with a 5 percent advantage a 10 pip average move would not be sufficient to cover round trip commission and spread

//---

hey max.... teamsky tripled their account with only a 6 pip profit target in 2007 metaquotes contest trading only the eurusd.... the eurusd has about 100 pips atr a day..... so a lot of money can be made with a small target....
//-----

very few are familiar with the real purpose and use of gettickcount()..... this will make it all too clear.....

using your largest monitor, on only one pair at a time, count the number of signals on every timeframe 5 and up for a minimum of 50 different indicators.... calculate the average length of each signal.... calculate the average +/- price excursion during the signal..... calculate the length of the last signal... denote whether it was a buy or sell.... calculate the length in bars of the current signal.....

this tends to give a rough idea of how long in bars and +/- pips the average signal lasts.... and gives a hint of a pending signal change.... i have a video on this very subject where after makayla placed a sell i asked her "now what"..... the type of data mining your doing answers that question.....

this must be done only a single pair at a time... yet still expect it to take upwards of 10 minutes to complete.... so do it during non trading hours or a separate platform not used for trading.... the platform will be non responsive until it completes ...... gettickcount().....

compare the results of the eur pairs with each other..... or the jpy pairs.... or usd..... you'll be surprised.....

it's an unbelievable amount of truly mind burning code but is worth it..... export the results to excel for ease of comparison.....

the charts below will mean very little to anyone other than me, but are some brief examples.....h
//----