Market Statistics & Probabilities (requests welcome) | Page 4 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Odds and Edge - Probabilities in Day Trading 1 reply

Best book on randomness in markets and market probabilities? 4 replies

Trading and Probabilities 3 replies

enoc's psychology&probabili ties thread 23 replies

Trader Statistics and Broker Statistics 0 replies

Attachments: Market Statistics & Probabilities (requests welcome)
Exit Attachments

Market Statistics & Probabilities (requests welcome)

  • #61
  • Jan 6, 2011 5:11pm Jan 6, 2011 5:11pm
I did find it very interesting that the sessions do have very similar pip averages

(according to last 2 years)
when london breaks upwards it averages 54 pips in up movement

and

on the days london beaks down it averages 56 pips down movement


I was impressed with the close stats.. it shows that if london decides to trend up or trend down between 3-6am (EST) it will average 55ish pips either way


I find this interesting myself.


Thanks for the idea
  • #62
  • Jan 6, 2011 5:14pm Jan 6, 2011 5:14pm
also perhaps it gives you some insight into the amount of money flowing through london.. you can in a way possibly judge londons trend.. if london is making an average of 9 pips directional movement per half hour then you would have some insight into the average rate of trend?

as well you would be aware if last hour made 18 pips of directional movement could feel comfortable there is an additional 37 pip move on average during that timeframe

ofcourse it sounds good in theory.. may or may not be that easy
  • #63
  • Jan 6, 2011 5:32pm Jan 6, 2011 5:32pm
Quoting newyear498
also perhaps it gives you some insight into the amount of money flowing through london.. you can in a way possibly judge londons trend.. if london is making an average of 9 pips directional movement per half hour then you would have some insight into the average rate of trend?

as well you would be aware if last hour made 18 pips of directional movement could feel comfortable there is an additional 37 pip move on average during that timeframe

ofcourse it sounds good in theory.. may or may not be that easy

I am not sure how much time you have for this. But an idea I have is to do another study and find on average what percentage of the time the net london breakout (up or down) matches the color of the 2 am EST bar. For example, if the 2 am EST bar is green, then what percentage of the time does london finish net positive.
If we have this information and it turns out that say the answer is 70% of the time and we combine this with the average breakout data from before then we can make a system where you enter the market at the end of the 2 am bar in the direction of that bar and aim for say 30-40 pips.
Does that make sense?
  • #64
  • Jan 6, 2011 6:16pm Jan 6, 2011 6:16pm
Quoting stockey
I am not sure how much time you have for this. But an idea I have is to do another study and find on average what percentage of the time the net london breakout (up or down) matches the color of the 2 am EST bar. For example, if the 2 am EST bar is green, then what percentage of the time does london finish net positive.
If we have this information and it turns out that say the answer is 70% of the time and we combine this with the average breakout data from before then we can make a system where you enter the market at the end of the 2 am...
I think this type of setup has been dreamed up by every trader and no one has yet to make money off it

The research I have made from past and current has left me to conclude that previous candles direction does not give any insight into future performance

the open and close to me only indicates where price existed at 2am (EST) the bar could have been going up and down a million times and when the clock struck 3am the close of the 2am candle got locked in upwards

almost all tests like this result in 50% probabilities because price can get stuck up or down 50% of the time

to me high and low I personally view as the best data with the exception of actual session times open and closes..

thats my concern
  • #65
  • Jan 6, 2011 6:31pm Jan 6, 2011 6:31pm
Quoting newyear498
I think this type of setup has been dreamed up by every trader and no one has yet to make money off it

The research I have made from past and current has left me to conclude that previous candles direction does not give any insight into future performance

the open and close to me only indicates where price existed at 2am (EST) the bar could have been going up and down a million times and when the clock struck 3am the close of the 2am candle got locked in upwards

almost all tests like this result in 50% probabilities because price can get stuck...

So you think this would also be around 50%?
I always get the impression that the 2 am bar kind of "primes" the london traders and gives some indication as to what will happen during london, but I could be wrong.

Just out of curiosity, what kind of probability do you think is a good starting point for building a strategy? 60%, 70%, or more? and why?
  • #66
  • Jan 6, 2011 6:40pm Jan 6, 2011 6:40pm
(10 years)
London's direction today will be London's direction tomorrow = 47.73%

NewYork's direction today will be NewYork's direction tomorrow = 47.82%
  • #67
  • Jan 6, 2011 6:54pm Jan 6, 2011 6:54pm
Quoting stockey
So you think this would also be around 50%?
I always get the impression that the 2 am bar kind of "primes" the london traders and gives some indication as to what will happen during london, but I could be wrong.

Just out of curiosity, what kind of probability do you think is a good starting point for building a strategy? 60%, 70%, or more? and why?
2am may prime things.. but the actual condition of open to close is a shitty indicator IMO, I prefer higher highs and higher lows .. if 2am makes higher high and higher low then 1am then we know 2am was an up move.


Well every probability represents a different factor so its a tough question

but you need to consider your risk to reward.. if you can have a 1:1 risk to reward and a 55% stat you can make money (don't forget spread tho)

and keeping 1:1 RR with 70-80-90% stats.. just gets better




If you want my real opinions then we get into talking about the chance of really getting a 1:1 RR with a 70% probability of hitting TP.. it just doesn't thrill me that much

I find the hedge funds have been finding ways to make money off random price action.. and I feel they are random.. I am currently interested in narrowing down my probabilities to get me in the center of all the movement and profiting from the fluctuations

but hey like they say.. two people using 1 system can have opposite results, whatever works for you.. just look at the contrarians

to complex a question to really answer I suppose but the RR and probabilities would go hand and hand is probably simplest opinion I can give
  • #68
  • Jan 6, 2011 7:30pm Jan 6, 2011 7:30pm
Quoting stockey
what kind of probability do you think is a good starting point for building a strategy? 60%, 70%, or more? and why?
I'll throw my 2 pips into this. The question isn't single trade gain/loss probability, it's overall strategy gain, and it thus depends on the strategy.

For example, you can be assured there are positions that were taken into this current E/U move at around the 3375 level in the early US hours of the 4th. Those positions (not counting increases in size of the positions as the move evolved) have been over 400 ahead in the last hour. But to establish those initial shorts, they might have taken as many as three 20-30 tick failed shorts in the late hours of the 3rd on into the early hours of the 4th. Those strategies "know" that entry is going to be a costly business. They are willing to accept 3 out of 4 trades being *losses*, along with the drawdown that goes with that, in order to profit the several hundred ticks that comes when the move does occur.

So pure "probabilities" is illusionary. The per-trade gain/loss profile depends on the relative size and nature (trailing, min profit, etc) of the stop strategy, relative to the trade exit strategy (the most simple of which is the pre-defined 'TP').

A key element is to find a set of rules: whether they are technical or fundamental, following or fading, that you can define and set in your mind. Then, if you can convince yourself that you are comfortable with the *overall*, not "this trade, that trade" risk profile those rules are going to give you, and assure yourself that you and your account size can accept that risk, and drill those rules into your thinking, comparing the market's behavior to them alone, and follow them without emotional overrides, you can be successful.
  • #69
  • Jan 6, 2011 8:33pm Jan 6, 2011 8:33pm
Quoting OldQuant
I'll throw my 2 pips into this....
OQ,

Its the points you bring up that force me for better or worse but I chose to do it since I believe its for better.. to trade by algorithmic models so I can avoid the emotions and follow things to a tee
  • #70
  • Jan 6, 2011 8:59pm Jan 6, 2011 8:59pm
I have only read the first few posts but, newyear, you are a god-sent. This is a thread I've always wanted to exist and you are a brilliant man for doing so.

I will continue to read the rest and look forward to more posts.
  • #71
  • Jan 6, 2011 9:26pm Jan 6, 2011 9:26pm
Quoting vx13
I have only read the first few posts but, newyear, you are a god-sent. This is a thread I've always wanted to exist and you are a brilliant man for doing so.

I will continue to read the rest and look forward to more posts.
the feelings are mutual.. without all of you the wheels would stop grinding and new paths would not be explored
  • #72
  • Jan 6, 2011 9:40pm Jan 6, 2011 9:40pm
Extremely interested in this thread. I will post again in the upcoming days.

Thanks.
  • #73
  • Jan 6, 2011 10:02pm Jan 6, 2011 10:02pm
I made an error when I enter the last data..

the new york averages were wrong.. they have been corrected.. see main post for correct averages


thanks
  • #74
  • Jan 7, 2011 12:01pm Jan 7, 2011 12:01pm
Nice thread..... I spent a good amount of time on various stats.
It is realy funny how your typical market thoughts falls into oblivion
after a "few" math proven statistics.....after a load of these tests,
you will bassicaly start to question everything learned previously.
You will realize that 99% of things you thought you knew is basicaly
crap wich totaly falls appart on simple math tests.

The simple truth is, market is random 80% of the time.
The other 20% is everything but random.....and it is called a TREND.

Use your math on trends.....not on same random events!

....but, if you want to test various things, I will help you to save some time.
The one and only thing that I proved to myself to be an edge is a gap.
Test those on weekly TF.....hit rate is bigger than any other stat I tested.
Some pairs % are high as 8:1 for gap to be filled

Regards
  • #75
  • Jan 7, 2011 12:13pm Jan 7, 2011 12:13pm
Quoting theFXButcher
The simple truth is, market is random 80% of the time.
The other 20% is everything but random.....and it is called a TREND.
lol....
try again...
after couple of years...
  • #76
  • Jan 7, 2011 12:15pm Jan 7, 2011 12:15pm
Realy.....

than enlighten me....i will shread to pieces everything you post!
  • #77
  • Jan 7, 2011 12:30pm Jan 7, 2011 12:30pm
Quoting theFXButcher
Realy.....

than enlighten me....i will shread to pieces everything you post!
price moves a path of least resistance....
there are two forces that push it....
liquidity size and market orders....

when price moves against size influence...
it will eventually hit the resistance point...
commonly known as overbought/oversold level...
trend is just one peace of disequilibrium in the market....

you know what you're doing and have the right info,
or stare at "random noise" most of the time

everything else is not free
  • #78
  • Jan 7, 2011 12:32pm Jan 7, 2011 12:32pm
So, you fall into Tape reading category guys.....good luck, you will
need it in uncentralized market!

This is a quote from your thread......

Yep, thx
I've had a felling this was the case with spot trading. I've been trying to figure DOM but it just doesn't look reliable, or at least it looks that some automated analysis is needed to forcast with some degree of certainty.

LOL......yes Rikers......fx is an uncentralized market

Again.....Trends are the only thing that is not random event!!!
  • #79
  • Jan 7, 2011 12:37pm Jan 7, 2011 12:37pm
Quoting theFXButcher
It is realy funny how your typical market thoughts falls into oblivion
after a "few" math proven statistics.....after a load of these tests,
you will bassicaly start to question everything learned previously.
You will realize that 99% of things you thought you knew is basicaly
crap wich totaly falls appart on simple math tests.
Most true.

I don't agree with the rest if what you said however. Just because you master trend trading doesn't mean other styles (such as tape reading/orderflow, mean reversion algorithms, etc.) can't work. You need to get over that typical mindset.
  • #80
  • Jan 7, 2011 12:38pm Jan 7, 2011 12:38pm
Quoting theFXButcher
So, you fall into Tape reading category guys.....good luck, you will
need it in uncentralized market!
I've posted this in an another thread
forex is not uncentralized as you think
most of the liquidity is centralized trough interbank
especially for the majors...
this is why you don't have many arbitrage points
when dealing with different brokers...

most of the arbitrage points is with crooked brokers that you won't make money anyhow

I don't trade tape...don't know why you would think that
0 traders viewing now