Random Price Movement | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Is random really random? 231 replies

random price movement - reasons why your system was wrong 5 replies

My feeling on price movements - they are random 13 replies

GBP/USD daily pip movement vs. EUR/USD daily pip movement 3 replies

Price action pattern, and random thoughts 6 replies

Attachments: Random Price Movement
Exit Attachments

Random Price Movement

  • #21
  • Aug 30, 2006 1:08pm Aug 30, 2006 1:08pm
Very interesting thread here. As a new forex trader, I've been so overwhelmed with all the information available about trading that my head began to spin. I've decided that for my own trading strategy, I would give priority to developing a good money management plan. I'm of the opinion that someone with good technical analysis skills and a poor money management plan has a higher probability of failure than someone with a strong money management plan. Can profitability in trading exist simply with a sound money management strategy and a coin toss?

I'd like to run an experiment with a demo account someday using a coin and a simple money management plan to see how it pans out. I believe in keeping things simple, so maybe something like the following:

Toss coin.
Heads=long 1 lot
Tails=short 1 lot
Risk=20 pips
Reward=80 pips

With the odds of a head or tail coming up at 50%, and my money management risk/reward ratio at 4:1, I would have to be wrong 4 times to every 1 time I was right, just to be even. If I was right 4 times, I would have to be wrong 16 times to break back to even. Shouldn't I be profitable in the long run with this simple money management plan and a coin toss?

I'm still very new to trading but it appears to me that a good technical analyst should do better than a simple coin toss and will improve the odds of a trade moving in his favor. I realize I've simplified this and haven't taken into account spreads and interest costs, but you get the basic idea. I'm sure something like this has been discussed and debated already but I haven't stumbled across it yet in this "sea" of information. I'd love to hear everyone's thoughts.
Hemico If you're not having any fun, you're doing it all wrong!
  • #22
  • Aug 30, 2006 1:49pm Aug 30, 2006 1:49pm
Quoting hemico
Very interesting thread here. As a new forex trader, I've been so overwhelmed with all the information available about trading that my head began to spin. I've decided that for my own trading strategy, I would give priority to developing a good money management plan. I'm of the opinion that someone with good technical analysis skills and a poor money management plan has a higher probability of failure than someone with a strong money management plan. Can profitability in trading exist simply with a sound money management strategy and a coin toss?

I'd like to run an experiment with a demo account someday using a coin and a simple money management plan to see how it pans out. I believe in keeping things simple, so maybe something like the following:

Toss coin.
Heads=long 1 lot
Tails=short 1 lot
Risk=20 pips
Reward=80 pips

With the odds of a head or tail coming up at 50%, and my money management risk/reward ratio at 4:1, I would have to be wrong 4 times to every 1 time I was right, just to be even. If I was right 4 times, I would have to be wrong 16 times to break back to even. Shouldn't I be profitable in the long run with this simple money management plan and a coin toss?

I'm still very new to trading but it appears to me that a good technical analyst should do better than a simple coin toss and will improve the odds of a trade moving in his favor. I realize I've simplified this and haven't taken into account spreads and interest costs, but you get the basic idea. I'm sure something like this has been discussed and debated already but I haven't stumbled across it yet in this "sea" of information. I'd love to hear everyone's thoughts.
Just for haha's I looked at the GBP/USD from 7/31/06 through today (hardly a complete data set). I used a quarter to determine long or short (heads long, tails short). I bought the opening price on the 3am EST bar and used your stoploss/takeprofit. Positions were held until win or lose or until the next 3am bar. No move to B/E. I did account for spread.

Out of 23 trades 5 were winners. Net pips = 20.

I have a lot on my plate at the moment but if I get a chance I'll code a strategy in Tradestation so I can change the variables and run it on larger data set.

Kind of funny to look at but I will agree with you that R:R is very important. I would go so far to say that many folks would do better by focusing a little less on trying to post crazy win % and more on improving their R:R.
Secret Identity: Bill Sullivan
  • #23
  • Aug 30, 2006 1:57pm Aug 30, 2006 1:57pm
  •  McD
  • | Joined Aug 2006 | Status: -Pecunia non olet- | 24 Posts
Quoting smjones
I got to thinking about it and that brings up an interesting point. There was no bias here, because it was random and yet here is the pattern. What caused it? Well it would be easy to say in terms of a market that there were more sellers than buyers or more tail than heads, but That really on a deeper level is not true. What there actually was, was congregation of tails to heads in a very small data set. This really tells nothing of trend at all, because there is no bias in this case it is random.

It makes no sense at all to look at small data sets in any market and try to determine a cause for movement... Only with a larger sample can even a trend be seen. And that is probably due to a self fulfilling prophecy.


Hi smjones !

First thanks to Phil for this interesting contribution !

BUT I have to make some remarks:

Already about 15 years ago I prgrammed a little program that did just the same (no EXEL,pure DOS). As far as I can recall you had to do the following in order to achieve
these amazing similarity to "real" chart-movement:

* produce a random number (random walk,but restricted in its movement by certain rules;as
Professor Mayes pointed out in his XLS-file)
* AND (and this is my point) one has to RESTRICT THE
DIRECTION of the movement.Otherwise there is no
"chart-like" behaviour.
This would be done by tossing another coin and the simple rule "that the new price follows in ca.67% the
movement of the previous price". Only in 33% there
would be the CHANCE of change in direction.

This would mean that there IS A CERTAIN BIAS programmed.

So it is not as easy as "toss a coin and place a trade"
(BTW there is alos a third possibility: moving sideways).


Sorry to disappoint you,smjones !

But on the other hand,there should be a similar bias in
the real market (a natural thing).

And that should be the real reason why one can produce
"random" numbers that look very similar to chart-movements.


But,still it is amazing and well worth being mentioned !

Thanks,
McD
  • #24
  • Aug 30, 2006 2:09pm Aug 30, 2006 2:09pm
Quoting hemico
Very interesting thread here. As a new forex trader, I've been so overwhelmed with all the information available about trading that my head began to spin. I've decided that for my own trading strategy, I would give priority to developing a good money management plan. I'm of the opinion that someone with good technical analysis skills and a poor money management plan has a higher probability of failure than someone with a strong money management plan. Can profitability in trading exist simply with a sound money management strategy and a coin toss?

I'd like to run an experiment with a demo account someday using a coin and a simple money management plan to see how it pans out. I believe in keeping things simple, so maybe something like the following:

Toss coin.
Heads=long 1 lot
Tails=short 1 lot
Risk=20 pips
Reward=80 pips

With the odds of a head or tail coming up at 50%, and my money management risk/reward ratio at 4:1, I would have to be wrong 4 times to every 1 time I was right, just to be even. If I was right 4 times, I would have to be wrong 16 times to break back to even. Shouldn't I be profitable in the long run with this simple money management plan and a coin toss?

I'm still very new to trading but it appears to me that a good technical analyst should do better than a simple coin toss and will improve the odds of a trade moving in his favor. I realize I've simplified this and haven't taken into account spreads and interest costs, but you get the basic idea. I'm sure something like this has been discussed and debated already but I haven't stumbled across it yet in this "sea" of information. I'd love to hear everyone's thoughts.
hemico,

I wish that I had focused more on money management and less on technical analysis when I started trading. Money management is not sexy. Telling your buddies that you are right 80% will bring you a lot more praise. One thing I would caution you against though: Simply saying that you want 4:1 reward doesn't make it so. I see a lot of people "tell" the market how much they want to make and it simply doesn't work that way. The more accurate your system, the less reward you need to live on. 4:1 is great and if you can get it, that's great. But to be honest, I know very very few people who come remotely close to that.

Phil
  • #25
  • Aug 30, 2006 2:53pm Aug 30, 2006 2:53pm
While I agree with Phil that you can't simply pull a number out of your hindquarters.....

Actually, I agree with Phil on a lot of things. I've found it's a wise thing to do as he's one of the smartest traders I've ever met. Enough of the

For a particular entry you can test the effectiveness of a particular exit strategy say like a trailing stop or ema cross. You can test a static exit exactly the same way. That information can be useful in determining 1) targets 2) reasonable R:R.

I use static targets as a guide. I use additional TA to determine acutal exit. I do have some setups that target 4:1. I don't enter those setups unless I feel I have a reasonable expectation that 4:1 can be reached.
Secret Identity: Bill Sullivan
  • #26
  • Aug 30, 2006 2:57pm Aug 30, 2006 2:57pm
Quoting philmcgrew
The more accurate your system, the less reward you need to live on.
Phil
I'm not sure if I understand your position with regard to randomness in market prices. How would one hope to attain any level of precision if price action can be explained by random movements?
  • #27
  • Aug 30, 2006 3:09pm Aug 30, 2006 3:09pm
I disagree. I have also written a random coin toss generator, but for the example here it is not necessary. All one needs to do is plot out the toss of a coin, Heads or tails. If heads graph = +1 if tails graph = -1 and then take a cum(toss(n))

This would cause a running total of the + and - 1s to a chart and it is random and therefore has no bias. It is simply a cumulative total. This produces the same patterns we have seen here.

I am never disappointed. Sometimes I am wrong, but never disappointed. As a trader, I do not have that luxury.

Thanks Scott


Quoting McD
Hi smjones !

First thanks to Phil for this interesting contribution !

BUT I have to make some remarks:

Already about 15 years ago I prgrammed a little program that did just the same (no EXEL,pure DOS). As far as I can recall you had to do the following in order to achieve
these amazing similarity to "real" chart-movement:

* produce a random number (random walk,but restricted in its movement by certain rules;as
Professor Mayes pointed out in his XLS-file)
* AND (and this is my point) one has to RESTRICT THE
DIRECTION of the movement.Otherwise there is no
"chart-like" behaviour.
This would be done by tossing another coin and the simple rule "that the new price follows in ca.67% the
movement of the previous price". Only in 33% there
would be the CHANCE of change in direction.

This would mean that there IS A CERTAIN BIAS programmed.

So it is not as easy as "toss a coin and place a trade"
(BTW there is alos a third possibility: moving sideways).


Sorry to disappoint you,smjones !

But on the other hand,there should be a similar bias in
the real market (a natural thing).

And that should be the real reason why one can produce
"random" numbers that look very similar to chart-movements.


But,still it is amazing and well worth being mentioned !

Thanks,
McD
  • #28
  • Aug 30, 2006 3:10pm Aug 30, 2006 3:10pm
Quoting SisterCurare
I'm not sure if I understand your position with regard to randomness in market prices. How would one hope to attain any level of precision if price action can be explained by random movements?
For the record, I do believe in technical analysis. As stated earlier, we can still count on fear, greed, and a herd mentality. I guess my point is that some movement is purely random and a little luck doesn't hurt. I think when people say stuff like the USD can go up 30 pips and then pat themselves on the back when it happens may have been the unknowing recipient of a random move rather than a skilled technician. I highly recommnd the book called "Fooled by Randomness". I think that the more you are aware that there is an element of randomness in the market the more you should realize that you better have some sound money management practices.

You can hope to attain precision by trading a system that you belive gives you a slight edge and then letting risk management and money management take care of the rest. It's the best we can do.
  • #29
  • Aug 30, 2006 3:22pm Aug 30, 2006 3:22pm
Certainly the idea of randomness in the markets is intruiging, but I just have a difficult time reconciling the influence of randomness on the markets with the forces of supply and demand. If randomness does play a part, I can't help but believe it is a minor one. Where's that article by Darkstar?
  • #30
  • Aug 30, 2006 3:40pm Aug 30, 2006 3:40pm
Quoting SisterCurare
Certainly the idea of randomness in the markets is intruiging, but I just have a difficult time reconciling the influence of randomness on the markets with the forces of supply and demand. If randomness does play a part, I can't help but believe it is a minor one. Where's that article by Darkstar?
Supply and demand is created by traders taking action on their beliefs. I think the question is whether their beleifs are correctly based on technical analysis or if there is a bit of randomness involved. I'm not out to change the world. It takes all types to make a market. If we all thought alike there would be no one to take the other side of our trades. You considered what I had to say and I appreciate that. I hope you are one of those who actully downloaded the spreadsheet and gave it a try. I wish you the best.

Phil
  • #31
  • Aug 30, 2006 4:47pm Aug 30, 2006 4:47pm
Quoting philmcgrew
hemico,

I wish that I had focused more on money management and less on technical analysis when I started trading. Money management is not sexy. Telling your buddies that you are right 80% will bring you a lot more praise. One thing I would caution you against though: Simply saying that you want 4:1 reward doesn't make it so. I see a lot of people "tell" the market how much they want to make and it simply doesn't work that way. The more accurate your system, the less reward you need to live on. 4:1 is great and if you can get it, that's great. But to be honest, I know very very few people who come remotely close to that.

Phil
Thanks Phil. I'm late to the game with the money management focus too. I'm thinking that the 4:1 loss to win ratio is just a way to beat the 50/50 odds of the coin toss. I arbitrarily picked the 80/20 pip numbers. With the ratio of 1 win to 4 losses, I would only have to be right 20% of the time to break even, right? 4 losing trades @ -80 pips, 1 winning trade @ +80 pips. 1 winning trade out of 5 total equals a 20% success rate. So if a coin toss gives me the potential of being right 50% of the time, that, in theory, puts me in the money. I'd love to tell all my friends that I'm right only 20% of the time and still make a profit.

The volatility of a currency pair would certainly need to be considered as well to see if 20 and 80 pips is realistic. I guess I'd like to see the results with these numbers on pairs that move like GBPSD or EURUSD. Who knows...you might be able to stretch it to 30 and 120 if the pair moves enough.

If someone could completely detach themselves emotionaly from a trade, set their stops and limits with a simple money management strategy like this, and walk away until the trade closes, could it be profitable long term? If it is, it would certainly reinforce the importance of focusing on development of a good money management plan. Good technical analysis would be that additional edge people look for to give them better than a 50% chance of being right.

I am the furthest thing from a math wizard, so my math may be incorrect, but am I missing something here? It would be a fun little experiment to see the results of. I'm hoping Captain Piptastic can come through with some numbers. Go Captain Go!
Hemico If you're not having any fun, you're doing it all wrong!
  • #32
  • Aug 30, 2006 6:14pm Aug 30, 2006 6:14pm
Ok. A little birdie sent me some code. The code is an always in strat. Basically if you don't have an open position you flip a coin and that determines long or short at the open of the next bar.

I used a 5min chart on GBP/USD from 7/6/05 to 8/30/06.

I ran it five times because obviously you get diffrent results each time.

I used a 20 pip stop and an 80 pip target as hemico first suggested.

I'm only going to post one of the tests. I'll explain why in the next post

1308 trades btw with a 21.64% win rate.
Attached Image
Secret Identity: Bill Sullivan
  • #33
  • Aug 30, 2006 6:24pm Aug 30, 2006 6:24pm
What the last post doesn't take into account is spread. As soon as you add the spread every test lost money. Everytime it was due to the spread.

So, I ran the optimizer a few times. TS's optimizer (so take this w/ a grain of salt) says optimal settings are 31 pip stop w/ a 152 pip target. that's 4.9:1 R:R. Those settings do yeild a gain (even w/ the spread accounted for).

Keep in mind this is a completely randomn entry so I ran the tests multiple times but they were remarkably close.

I have adjusted the code so that it enters just once per day at 3am but it has a bug somewhere. Honestly it will be a while before I have a chance to look at it much more.

At any rate here is the "winning" test. Ugly but profitable. LOL
Attached Image
Secret Identity: Bill Sullivan
  • #34
  • Aug 30, 2006 7:11pm Aug 30, 2006 7:11pm
Hmmm,

Interesting that Zero falls ~ 50% between Max & Min Equity.
  • #35
  • Aug 31, 2006 12:14am Aug 31, 2006 12:14am
  •  McD
  • | Joined Aug 2006 | Status: -Pecunia non olet- | 24 Posts
Hi Captain Piptastic ,

as I tried to point out in my earlier post, there is not a
50:50 chance of movement.
Therefore, a toss of a coin does not apply here.
The 50:50 chance is more in the amount of value
the price is going up or down.

As you can see there are always two (more or less) random processes to consider:
1. the direction
2. the amount of price change (in that direction)

both have to be considered separately because they do
not have anything in common .

In ca. 67% of the time Markets tend to go in the same
direction as the previous tick.

Just imagine, the direction would in fact change with
a 50:50 chance; there would be a mess on your chart
and nothing anywhere near the reality.


If you backtest anyway, then how about trying to apply your coin toss only in 67% of the time (depending on the
direction of the previous coin toss?).
I do not know the possible outcome of this,but I am
very curious indeed.

Just my 2 cents.

Good luck,
McD
  • #36
  • Aug 31, 2006 12:55am Aug 31, 2006 12:55am
Quoting McD
Hi Captain Piptastic ,

as I tried to point out in my earlier post, there is not a
50:50 chance of movement.
Therefore, a toss of a coin does not apply here.
The 50:50 chance is more in the amount of value
the price is going up or down.

As you can see there are always two (more or less) random processes to consider:
1. the direction
2. the amount of price change (in that direction)

both have to be considered separately because they do
not have anything in common .

In ca. 67% of the time Markets tend to go in the same
direction as the previous tick.

Just imagine, the direction would in fact change with
a 50:50 chance; there would be a mess on your chart
and nothing anywhere near the reality.


If you backtest anyway, then how about trying to apply your coin toss only in 67% of the time (depending on the
direction of the previous coin toss?).
I do not know the possible outcome of this,but I am
very curious indeed.

Just my 2 cents.

Good luck,
McD
Somehow you've confused and combined two separate ideas being presented in this thread.

Here is how the thread has progressed so far:

Phil posted an excel sheet that attempts to show price patterns can be created from randomn coin tosses. Some folks responded with questions and one person asked about the law of large numbers.

I responed to the law of large numbers question (correctly even). Then heimco asked a semi related question about risk to reward and randomn entry.

I responed with a small test and then a series of larger tests. I did this because I think risk to reward is a neato concept. And I'm really only here for the entertainment. The tests were entertaining.

This separate R:R/coin flip question really should have been a separate thread. But being the son of a pirate I tend to be involved in these sort of hijackings.

BTW applying the toss only 67% of the time (randomnly anyway) would only yeild a smaller data set. It won't change the results.
Secret Identity: Bill Sullivan
  • #37
  • Aug 31, 2006 1:31am Aug 31, 2006 1:31am
  •  McD
  • | Joined Aug 2006 | Status: -Pecunia non olet- | 24 Posts
Captain,

I totally agree: this is pure entertainment.
But there's always a bit to learn in it (while having fun).

I don't want to change the thread's focus,but I think that
we have to consider the (possible change) of direction
as well, as this has (taking Phil's excel file) an underlying
bias of about 67%.

Well,this would yield a smaller dataset.But would the
quality (of the prediction) change to better also?

My point is,if we make the assumption that real markets
have an "inbuild" bias to follow the direction of the previous tick (that's why Phil's spreadsheet can produce
this amazing result) than we should consider that also
in backtesting (even for entertainment).

So I don't see this as a different approach;more as a detail that should/could be considered.

Just an idea,
thanks
McD
  • #38
  • Aug 31, 2006 4:07am Aug 31, 2006 4:07am
Quoting McD
Captain,

I totally agree: this is pure entertainment.
But there's always a bit to learn in it (while having fun).

I don't want to change the thread's focus,but I think that
we have to consider the (possible change) of direction
as well, as this has (taking Phil's excel file) an underlying
bias of about 67%.

Well,this would yield a smaller dataset.But would the
quality (of the prediction) change to better also?

My point is,if we make the assumption that real markets
have an "inbuild" bias to follow the direction of the previous tick (that's why Phil's spreadsheet can produce
this amazing result) than we should consider that also
in backtesting (even for entertainment).

So I don't see this as a different approach;more as a detail that should/could be considered.

Just an idea,
thanks
McD
No.

This is hemico's original question:

Can profitability in trading exist simply with a sound money management strategy and a coin toss?

It has nothing to do with the excel file. We could have used a 6 sided die and said even's long, odds short.

We aren't concerned with direction, bias or anything else.

Sure, consider all the possibilities. But consider reading all the posts carefully first.
Secret Identity: Bill Sullivan
  • #39
  • Aug 31, 2006 4:30am Aug 31, 2006 4:30am
Quoting McD
Captain,

I totally agree: this is pure entertainment.
But there's always a bit to learn in it (while having fun).

I don't want to change the thread's focus,but I think that
we have to consider the (possible change) of direction
as well, as this has (taking Phil's excel file) an underlying
bias of about 67%.

McD
McD

This experiment is about as random as you can get. Can you provide your results that have lead you to conclude the market goes sideways 67% of the time? Also include the pairs you tested and the time frames you looked at.

For the record, this is not my spreadsheet. It was written by a professor of finance. I emailed him and thanked him for using it. He said that he teaches technical analysis to his students and then a week later shows them this spreadsheet. In his words: "it blows their minds".
  • #40
  • Aug 31, 2006 7:33am Aug 31, 2006 7:33am
Quoting FXopportunist
I have seen the coin flip deal applied to money management in many books and articles. I have never seen anyone use it for evaluating technical analysis. If we couldn't see patterns in price movement, and market was truly random, the whole game would be money management and trading would be ruled by accountants. Sorry in advance to the accountants, but, Y A W N ! !
If market is truly random, there is no way you can make money. Does not matter psychics, money management or whatever. What is more, you could lose money becasue of brokers spread.
0 traders viewing now