Coin-Toss Trading Strategy | Page 4 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

TEB: Random Entry - Coin Toss Great with R.R.R. 105 replies

Weighted Coin Toss Trading 118 replies

Toss a coin? 9 replies

coin toss entry question 24 replies

Attachments: Coin-Toss Trading Strategy
Exit Attachments

Coin-Toss Trading Strategy

  • #61
  • Feb 10, 2008 11:42pm Feb 10, 2008 11:42pm
Quoting Paris Hilton
The Daily Stochastics Oscillator System in on the Trading Systems forum and you will find a thread there stating Daily Stochastics Oscillator System #2. The take profit is 120 pips and the stop loss is 75 pips. v1.4 EA has been back tested for the full year of 2007 and has a 66% win rate with never losing more than 4 times in a row in that year. It should work as good as anything and better than most with binary equation. I doubt the luck of a coin toss can do any better. We know that binary equation is not martingale or anti-martingale. It's a diffent money management system but it is a progression. A coin toss with money management sounds very iffy and alot like gambling. So maybe this is better. Paris
Sounds good. I can't say that I liked using the BES all that much although it was a good experience. It really needs the right trading strategy to be successful, otherwise the drawdown can be pretty severe. So far I haven't really found any trading strategies that I would feel comfortable using it on, but I am testing two at the current time, so we will see.

With the right strategy it has a lot of potential.

Cheers,

Damon
"Keep your eyes on the helpers" - Mr. Rogers
  • #62
  • Feb 12, 2008 2:39am Feb 12, 2008 2:39am
Quoting damonl
If you use your imagination, there is a very simple way to reduce your risk.
Damon, there are three ways that I know of to decrease risk:
1) More accurate entries and exits.
2) Decrease position size (but this decreases return in like proportion).
3) Trading uncorrelated simultaneous positions (a general principle, whose application in forex might be limited due to correlation between pairs)

Any others that you care to share with us?

David
  • #63
  • Feb 12, 2008 2:45am Feb 12, 2008 2:45am
Quoting billbss
No.
The statement isn't subjective. The statement is a cardinal rule of sound money management. A person may not be able to determine that their chance of a win has increased. In that case they would continue risking the 2% (or whatever).
This is 100% correct. Adjustment in bet (position) size can only be justified in an event where expectancy can be shown to likewise greater. (That is how the Blackjack pros in the 1970s/80s beat the casinos: placing larger bets when the card count was highly positive, i.e. higher number of favorable cards remaining to be dealt). If (and I stress IF) it can be shown that wins and losses tend to "cluster" together, i.e. a winning sequence improves the probability that the next trade will also be a win, only then is an increase in position size justified. Position sizing systems (e.g. martingale, anti-martingale) do not, in themselves, alter expectancy. Any reliable authority on games involving probability will confirm all of this.

If exit (as opposed to position size) is what is meant by MM, then I submit the following theories for consideration:

a) The principle of cutting losses quickly, and letting profits run, will complement a trend following entry method. This will generally result in a win rate of less than 50%, but there will be occasional large wins to compensate. The more the market "trends", the greater this effectiveness of this approach.

b) Using back-testing to optimize TPs and trailing SL values amounts to curve-fitting in much the same way that optimizing indicators used for entry does. To what extent, if any, this can provide a long term edge is open to debate (and has been debated regularly on the forum).

c) Whatever "randomness" might exist with entries also exists with exits, since the market cares nothing as to whether an individual trader is entering or exiting. If "high probability" entry is about finding long term inefficiencies, then exits can potentially improve overall expectancy by likewise exploiting inefficiencies. If random entries/calculated exits can be shown to be profitable, then calculated entries/random exits must necessarily be equally so......

d) ...... albeit with one proviso. Exit doubles as a function of risk management (capital preservation). However, I see no reason why SLs can’t first be set according to technical considerations, and then position sizes set to control the amount at risk.

Must admit that I haven’t seen the BES or the Daily Stochastics.

David
  • #64
  • Feb 12, 2008 6:24am Feb 12, 2008 6:24am
Daily Stochastics is simply an entry system posted on these forums.

http://www.forexfactory.com/showthread.php?t=69151

Binary Equation System MIGHT be an Exit/MM system, granted, I don't know enough about it to make any certain conclusions.


As for the original goal, I've been running the numbers while coding this.

A chance of heads or tails is 50%. What is the chance that either flip will be correct? Another 50%? If that's the case, then the MM needs to be setup for a 25% accuracy system.

Either that or I just need to go to bed.
  • #65
  • Feb 12, 2008 6:50am Feb 12, 2008 6:50am
Quoting Ronald Raygun
Daily Stochastics is simply an entry system posted on these forums.

http://www.forexfactory.com/showthread.php?t=69151

Binary Equation System MIGHT be an Exit/MM system, granted, I don't know enough about it to make any certain conclusions.


As for the original goal, I've been running the numbers while coding this.

A chance of heads or tails is 50%. What is the chance that either flip will be correct? Another 50%? If that's the case, then the MM needs to be setup for a 25% accuracy system.

Either that or I just need to go to bed.
Ronald, many thanks for the link.

Re the coin toss, I'm not quite sure what you mean. The probability of tossing a head is always 50%, even if the previous toss was a head also. The probability of tossing two successive heads is 25%. The probability of tossing at least one head in two tosses is 75%.

David
  • #66
  • Feb 12, 2008 7:17am Feb 12, 2008 7:17am
Clearly I was not awake.

Anyway, the chance of a head or tails if 50%.

The chance of a head(buy) making a profit is 50%
The chance of a tail(sell) making a profit is 50%

.5 x .5 = .25 or 25%

Do you see my logic?
  • #67
  • Edited 10:41am Feb 12, 2008 8:20am | Edited 10:41am
Quoting Ronald Raygun
Clearly I was not awake.

Anyway, the chance of a head or tails if 50%.

The chance of a head(buy) making a profit is 50%
The chance of a tail(sell) making a profit is 50%

.5 x .5 = .25 or 25%

Do you see my logic?
The chance of it making a profit may or may not be 50%. It depends on your SL and TP.
Whatever it is, you will always wind up break even, over time, minus the effect of the spread, if you are relying on a coin toss. The spread will gradually eat away your account.

There is no "creative" MM method that will prevent this from happening. Most of them will hasten the end. You cannot money manage a loser into a winner.
  • #68
  • Edited 2:02am Feb 13, 2008 12:36am | Edited 2:02am
Quoting billbss
The chance of it making a profit may or may not be 50%. It depends on your SL and TP.
Whatever it is, you will always wind up break even, over time, minus the effect of the spread, if you are relying on a coin toss. The spread will gradually eat away your account.

There is no "creative" MM method that will prevent this from happening. Most of them will hasten the end. You cannot money manage a loser into a winner.
Spot on, IMHO, Bill.

Everything else being equal, a wider TP (relative to SL) means good RR (win size), but low win rate. A wider SL vice versa. There's a widely held belief that cutting losses quickly and letting profits run is enough to give one a discernable edge. I've yet to either prove it or disprove it.

It's always good to see people trying to think innovatively, like many of the posters in this thread. When I first started, I likewise thought I had a good chance at conquering the markets by way of mathematical sleight-of-hand. But alas, it's not that simple. If there was a clear-cut "Holy Grail" trick or technique, everybody would be using it, and eventually there wouldn't be a market, since each transaction must ultimately involve a willing buyer and a willing seller.

There's an excellent post by Plutonite here:
http://www.forexfactory.com/showthread.php?p=1810347
Here's a section of it:

"....... It is a MIRACLE that we are making ANY correct statements about a market that was at one point thought to be inherently chaotic and foolish to participate in. Remember that it took several academic papers in the Journal of Finance to prove that the market is not a random walk model.

In fact, the market (and any similar exchange) is inherently chaotic, due to the large number of unrelated participants whose actions both commercial and speculative cannot be previously determined, and whose collective effect, weighed by volume, is even harder to predict. It's a crazy game we are playing if you think about it. Just because you have charts that seem to tell a picture in hindsight does not mean that anything makes sense at all. This is an unknown function with a built in random factor, and we're trying to ride it. F*cking crazy.

At the same time, there seem to be recurring patterns in the market activity that repeat themselves, and which do form both a statistical and an intuitive basis to make time series propositions (aka trades), but that is a topic for another discussion. ......."

IMHO, that sums it up. Somehow the trader must find a phenomenon, or a pattern, amongst the chaos, that offers a potentially demonstrable edge, and then devise a method of entries and exits that systematically exploits it.

The function of position size can only be simple. In a positive expectancy situation, the key is to keep bet size small (and consistent with expectancy), allowing the account to survive a series of losses, and giving expectancy the opportunity to ultimately prevail. In a negative expectancy situation, the best hope mathematically is to place one's entire bankroll on a single lucky shot, and then quit while (in reality, if) one is ahead.

IMHO, scaling into, and/or out of, trades appears sophisticated, but each component trade still has to deliver, on balance, in its own right. Otherwise it is worthless. For example, it might seem smart to take partial profit and move SL to breakeven, at a predetermined point, guaranteeing a winner. But that means that only half of the trade has the potential to achieve a "home run", while the whole position carries the risk of an early stop out. Again, everything else being equal, the innovative MM delivers no long term mathematical benefit: to decrease risk, we are decreasing potential return. I have yet to see a "cunning" MM technique allowing return to be increased without affecting risk, or risk reduced without compromising return, in like proportion. (But I try to keep an open mind, and I'd be delighted if somebody could show me one ).

I assume that "coin toss" refers to random entries (or exits). Costs make trading a negative sum game, so any methodological edge must be sufficient to overcome costs. By relegating one part of the entry/exit equation to a random activity, one is only making a difficult task even more so.

David

Footnote: just in case anybody thinks this post is unduly pessimistic, I've highlighted Plutonite's point "there seem to be recurring patterns in the market activity that repeat themselves". Find one of those, and you are on the way to joining the ranks of the consistently profitable traders who post in this forum.
  • #69
  • Feb 13, 2008 2:00am Feb 13, 2008 2:00am
Quoting billbss

There is no "creative" MM method that will prevent this from happening. Most of them will hasten the end. You cannot money manage a loser into a winner.
Exactly. The coin-toss strategy is a loser. Because you have 50% chance of winning - BUT because of the spread, it reduces your winning %, ever so slightly. So in the long run, you're doomed.
  • #70
  • Mar 11, 2008 9:52am Mar 11, 2008 9:52am
I did coin toss trades a while back......

It work. It profits.......
  • #71
  • Edited 6:44pm Mar 11, 2008 9:53am | Edited 6:44pm
Quoting hanover
The coin toss entry and MM (..........

Anybody who believes that, for example, if a coin toss has produced an excessively high number of heads is now more likely to produce a tail, is suffering from "Gambler's Fallacy". Explained in this thread: http://www.forexfactory.com/showthread.php?p=1786938
Billbss is correct - the coin has no memory. Each toss is an independent event.
David
Then how come it is believed that its a 50% change in the beginning
  • #72
  • Mar 11, 2008 1:13pm Mar 11, 2008 1:13pm
Quoting hanover

It's always good to see people trying to think innovatively, like many of the posters in this thread. When I first started, I likewise thought I had a good chance at conquering the markets by way of mathematical sleight-of-hand. But alas, it's not that simple. If there was a clear-cut "Holy Grail" trick or technique, everybody would be using it, and eventually there wouldn't be a market, since each transaction must ultimately involve a willing buyer and a willing seller....

I assume that "coin toss" refers to random entries (or exits). Costs make trading a negative sum game, so any methodological edge must be sufficient to overcome costs. By relegating one part of the entry/exit equation to a random activity, one is only making a difficult task even more so....

David

I appreciate the intelligence and lengthy market research it took to know the truth behind your post.

There are a couple of things needed to be addressed in the random coin flip trade concept.

 

  1. The first: the currency pairs do not rise and fall daily in a 50/50 distribution.
  2. The second: the fact that the currency pairs are not likely to have a distribution of anywhere near 50/50 on random entries with long hold times and fixed stops on the random entry(regardless of size). Much greater loss percentages, but contained losses. Surviving random profit trades' total profits tend to exceed total random loss trades' losses.

Regardless of markets tested the facts above remain constant.

Peter

  • #73
  • Mar 11, 2008 1:20pm Mar 11, 2008 1:20pm
Quoting PeterCrowns
I appreciate the intelligence and lengthy market research it took to know the truth behind your post.

There are a couple of things needed to be addressed in the random coin flip trade concept.

  1. The first: the currency pairs do not rise and fall daily in a 50/50 distribution.
  2. The second: the fact that the currency pairs are not likely to have a distribution of anywhere near 50/50 on random entries with long hold times and fixed stops on the random entry(regardless of size). Much greater loss percentages, but contained losses. Surviving random profit trades' total profits tend to exceed total random loss trades' losses.

Regardless of markets tested the facts above remain constant.

Peter

Does anyone have a clue what he just said?
If you do, please fill me in
  • #74
  • Mar 11, 2008 1:55pm Mar 11, 2008 1:55pm
Something bout stop losses I guess.

Quoting billbss
Does anyone have a clue what he just said?
If you do, please fill me in
  • #75
  • Edited 6:56pm Mar 11, 2008 6:41pm | Edited 6:56pm
Quoting billbss
Does anyone have a clue what he just said?
If you do, please fill me in
Let Me try

  1. The first: the currency pairs do not rise and fall daily in a 50/50 distribution.

Price dont always go up and down evenly .....some days it just go down some just up... for weeks.

  1. The second: the fact that the currency pairs are not likely to have a distribution of anywhere near 50/50 on random entries with long hold times and fixed stops on the random entry(regardless of size). Much greater loss percentages, but contained losses. Surviving random profit trades' total profits tend to exceed total random loss trades' losses.

the prospect of a position hold after an entry to achieve its target is less then it reach its StopLoss therefore has greater loss percentage ( small $$ losses because of SL), While the trades that survived ( Not Hitting SL ) tend to exceed the total losses because of its greater profit on each trades , Basically it says Let the porfit Run....

  1. Regardless of markets tested the facts above remain constant.

It doesnt matter where you test it....... It always give same result.Peter

Regardless which market you tr it with, its the same so I guess many small losses are covered by several big winners,Syafi

  • #76
  • Mar 11, 2008 6:50pm Mar 11, 2008 6:50pm
I really wonder.................

Coin Toss strategy were suppose to be the simplest of strategy..but from the debate written here, it has turn into full scale economic paper . with Great high Class English (which I enjoy reading).

How to keep it simple ?
  • #77
  • Mar 11, 2008 7:50pm Mar 11, 2008 7:50pm
Quoting Syafi
I really wonder.................

Coin Toss strategy were suppose to be the simplest of strategy..but from the debate written here, it has turn into full scale economic paper . with Great high Class English (which I enjoy reading).

How to keep it simple ?
The coin toss strategy can be made simpler by just writing out a check to your broker everyday and not messing with tossing the coin.
  • #78
  • Mar 11, 2008 8:21pm Mar 11, 2008 8:21pm
instead , I plan to ask my broker to send me wire every week by tossing te coin
  • #79
  • Mar 11, 2008 10:46pm Mar 11, 2008 10:46pm
Once I finish my study of MQL4 this will be the first EA that I'll develop, I'm planning to optimize 2 parameters for each pair: the SL = TP, and a parameter that will generate the exit signal when the trade is in profit above TP.

I'll post here if I can get to a conclusion about this approach.
  • #80
  • Nov 15, 2008 4:43pm Nov 15, 2008 4:43pm
I don't think you need to make this to hard just use a bar pivot set up the random come in to weather the new bar open over or under the pivot all you need then is the mm/risk side of thing ie the hard part

double high = iHigh(NULL,0,1);
double low = iLow(NULL,0,1);
double close = iClose(NULL,0,1);
double P = (high+low+close)/3;
double Buy1_1 = iOpen(NULL, 0, Current + 0);
double Buy1_2 = P;
double Sell1_1 = iOpen(NULL, 0, Current + 0);
double Sell1_2 = P;

bool Long = Buy1_1 < Buy1_2;
bool Short = Sell1_1 > Sell1_2;

use this or some thing similar

great thread raygun

good luck
0 traders viewing now