Long term Profits within Negative Expectancy | Page 16 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

What is Expectancy? 13 replies

System Expectancy - Monte Carlo Simulation 5 replies

Short Term / Medium Term / Long Term? 3 replies

Random Market vs Set Expectancy 8 replies

Expectancy is always zero 41 replies

Attachments: Long term Profits within Negative Expectancy
Exit Attachments

Long term Profits within Negative Expectancy

  • #301
  • Aug 25, 2009 6:31pm Aug 25, 2009 6:31pm
Quoting daytrading
Petra,

Since this is not a coin flip or a roulette wheel, winning X does not have to equate to losing X on a 50/50 basis. I give you an example:

One of my models has a pretty balanced win/loss ratio - i.e. over the course of a month, quarter and year, it comes to more or less 50% winning and 50% losing trades (1200 on avg per year). The model relies on a less frequent but reoccurring phenomenon which delivers days (not indiv. trades) with a 1:4 risk/reward ratio. Since I never really know when these days happen in advance, I had to create a...
DT,

Thank you for your candor. You have set forth parameters that more than overcome any negative expectancy. 1:4 Risk/Reward with a win ratio of anywhere near 50% is exciting. I cannot accuratley calcuate your positive expectancy without the exact pip loss and pip profit points--- but no matter. You have a winning edge. How much data do you have (live and backtest) to give you confidence? I would be happy to run any numbers for you. One number you might be intrested in is the longest string of losers you should expect over 100, 1000, and 10,000 trades. That way, when the bad steak happens, you will be able to withstand it with confidence, knowing it is just a statistical blip and there is no reason to question your system. Congratulations!

Regards,
  • #302
  • Aug 25, 2009 9:00pm Aug 25, 2009 9:00pm
Quoting pipmutt
Sorry, setting a daily target is not the way to trade. The market only offers what it offers, it won't care about your goals.
You are correct, Mate, the market does not care about my goals. That's why I have to tend to my account and quit when I reach my goals.

No one knows if their next trade is going to be a winner or loser. When I hit my goals, I quit and tend to other matters. I can trade the next session.
  • #303
  • Aug 26, 2009 2:08am Aug 26, 2009 2:08am
Quoting gbppip
When I hit my goals, I quit and tend to other matters. I can trade the next session.
Sure, there are more important things in life than trading, it's just a means to an end. I'm surprised by the amount of time some traders spend in front of the screen, like 16 hours a day, wow!!
  • #304
  • Edited 6:29am Aug 26, 2009 5:52am | Edited 6:29am
Quoting FX-Petra
DT,

Thank you for your candor. You have set forth parameters that more than overcome any negative expectancy. 1:4 Risk/Reward with a win ratio of anywhere near 50% is exciting. I cannot accuratley calcuate your positive expectancy without the exact pip loss and pip profit points--- but no matter. You have a winning edge. How much data do you have (live and backtest) to give you confidence? I would be happy to run any numbers for you. One number you might be intrested in is the longest string of losers you should expect over 100, 1000, and...
Hi Petra,

The entire construction is based on maximum %-loss per currency pair (5 pairs traded daily) and maximum %-loss per day. The indiv. positions have the same size across the board and use the same entry and exit criteria day in, day out.

I am (when trading the model) not so much interested in the individual performance of each currency pair, but instead I look at the daily result in percent of the total equity as well (and more importantly) the monthly results (again in percent).

My observations and the data show that although there are frequent losses (which are limited as explained above), a complete loss of all 5 positions in one day is a very rare occurence. On the other side, when all positions are winning (which is rare as well) they outperform (the losers) by a large margin since the winning positions are not limited in the same way by a percentage stop loss. The profits can run unhindered so to speak.

I have about 4 years of data, of which 1.5 years have been tested under live conditions. All the back-tests have been done manually, so they are accurate and not curve fitted since no indicators are used at all. Entries and exits are simply based on a 'time base' calculation.

We are talking over 5000 individual positions, but as I mentioned it is not a string of losses in a specific currency pair that I pay attention to - rather a string of losing days that fall outside a set category (for example more than a week of losing days that hit the full daily loss limits) before I would get concerned about the profitability - since this never happened and if it suddenly did, it would constitute a 'sigma - event'.

I hope that helps.

regards
daytrading
Enter Signature
  • #305
  • Aug 26, 2009 12:34pm Aug 26, 2009 12:34pm
Quoting daytrading
Hi Petra,

The entire construction is based on maximum %-loss per currency pair (5 pairs traded daily) and maximum %-loss per day. The indiv. positions have the same size across the board and use the same entry and exit criteria day in, day out.

I am (when trading the model) not so much interested in the individual performance of each currency pair, but instead I look at the daily result in percent of the total equity as well (and more importantly) the monthly results (again in percent).

My observations and the data show that although there...
For the record, with a 50% win ratio you should expect At Least a string of losses equal to 10 in a row within the first 1000 trades. From a practical standpoint I would make sure I was prepared for a 13 or 14 sequence loss... this would not be unexpected.
  • #306
  • Aug 26, 2009 12:49pm Aug 26, 2009 12:49pm
Quoting FX-Petra
For the record, with a 50% win ratio you should expect At Least a string of losses equal to 10 in a row within the first 1000 trades. From a practical standpoint I would make sure I was prepared for a 13 or 14 sequence loss... this would not be unexpected.
Hi Petra,

I have gathered that much.

Just so you have a little insight how I collect the data, I have attached the worst and the best month data for the year 2007 (percentages relative to equity on daily basis in the left pane, cumulative percentages relative to equity in the right pane).

The numbers in the fields are pips achieved per day (spread not included - I have a calculation that allows to cater for it by subtracting 5% off the annual return)

The year ended up +32% before cost of 5%, so +27% net return.

I also file distribution of trades based on percentage brackets per month plus other data (but that would go to far for the thread).

regards
daytrading
Attached File(s)
File Type: doc ORT March 2007.doc   95 KB | 543 downloads
File Type: doc ORT November 2007.doc   95 KB | 489 downloads
Enter Signature
  • #307
  • Aug 30, 2009 2:28am Aug 30, 2009 2:28am
Quoting mikkom
Money management only keeps risk of ruin minimized (which is important of course) but more important is to enter so that your initial stop is not hit before exit and exit the trade at the moment when price is as much positive as possible. There is no need to predict direction as long as these two are statistically on better side on a big number of trades. Might sound simple but that's it at least for me. There are points in price where price is statistically biased to go to one direction - See this pic http://fxlabs.oanda.com/products/his...erhistory.html...
Great thread. Does anyone use COT, a weekly report which shows open interest for futures contracts?

http://www.dailymarkets.com/commitment-of-traders/

Quoting mikkom
About stop hunts, yes they do happen. I have a friend in a tier 1 bank and here is a post from former tier 1 bank guy http://www.forexfactory.com/showpost...96&postcount=8 (fti is not just a random trader, read his thread - he reveals his history at some point) Believe it or not but ask yourself, if the revenue of your company [I'm not sure how many people here have run companies] would depend on sweeping stops levels, would you do it? Then ask again in what business banks are?

And once again, here is a link to scientific...
Stop-hunting is a regular occurrence. A lot of FX traders complain about stop-hunting, perhaps their stops are placed in the obvious areas, which are regularly hunted and taken out by the big money.
  • #308
  • Aug 31, 2009 8:54am Aug 31, 2009 8:54am
Well, it seems all the comments about this topic have been posted. I will try to sum up what we have learned and see if there are any conclusions that can be drawn:

The core of the problem seems to be that same old vexing problem: Is the market random? There is a yes crowd and a no crowd.

If the market is random then I think we can conclude that long term winning traders will be very few. If we cannot predict market direction correctly to a statistically significant degree, then all we are left with is our money management and the volitility of the market. These forces alone are not enough to overcome negative expectancy. The bell curve does allow for a small number of successful traders beyond the 10,000 trade mark.... but those traders have been, at least to some extent, very lucky. If they continue to trade the number of survivors will continue to be reduced.

If the market is not random then there is money to be made. Most traders fall into this category. Please remember that I am talking about short term daytrading. Some have submitted suggestions that, on a longer term basis, the market is not random at all. I tend to agree with those ideas. However, many have also suggested that the market is non random on short term intraday basis. I have seen no evidence, either from my reasearch or from posters comments, that this assumption has any basis in fact.

I encourage all traders to analyize the reasoning behind their decision to enter a trade. I submit that most trades are entered on the basis of hope, instinct, belief, or frustration. What research is there, when price hits that fib line, that it will bounce? If you cannot answer that question, then you are entering a trade with only money management and volitility to protect you and turn you to profit. Over the long term these tools alone will not be enough. You must be able say that, after research, for example, that when price hits the 38.2% fib it will bounce 10 pips 60% of the time. Now that is market knowledge that you could take, along with good money management, and become successful.

Finally, a bit of common sense. Take a look to the left and right of this page. Notice the many adds of bucket shops soliciting your business. Almost all of these market makers trade against you by taking the opposite side of your trade. Do you believe they are advertising to lose money? Are they trading against you "hoping" or "wishing" that they will be successful? No, they know they have a edge with positive expectancy. They now that the vast, vast, vast majority of their customers are trading with negative expectancy. All they need to do to make your account balance theirs is time and money.... and they have plenty of both.

I wish the best to all and kindest regards,
  • #309
  • Aug 31, 2009 9:57am Aug 31, 2009 9:57am
Quoting FX-Petra
Well, it seems all the comments about this topic have been posted. I will try to sum up what we have learned and see if there are any conclusions that can be drawn:

The core of the problem seems to be that same old vexing problem: Is the market random? There is a yes crowd and a no crowd.

If the market is random then I think we can conclude that long term winning traders will be very few. If we cannot predict market direction correctly to a statistically significant degree, then all we are left with is our money management and the volitility...
Hi FX-Petra:

The actual problem is there is no agreement on the definition of random.

If you can't say with 100% certainty an event is going to happen, then it is a random event.

The sun will shine tomorrow. Sounds like a certainty, doesn't it? What if the sun became a black hole, no more sunshine. Sooner or latter the sun will become a black hole. So the sun shining is a random event. Even though the probability is high, it is still a random event.

You will wake up tomorrow. Is that certain? NO. So that too is a random event.

Absurd examples? Not really. They get to the core of the problem - the definition of random. Yes, the market makes patterns that sometimes repeat and can be exploited for profit. That does not prove the market is not random.
If you only trade long, then you can't always be wrong.
  • #310
  • Aug 31, 2009 9:58am Aug 31, 2009 9:58am
Quoting FX-Petra
Finally, a bit of common sense. Take a look to the left and right of this page. Notice the many adds of bucket shops soliciting your business. Almost all of these market makers trade against you by taking the opposite side of your trade. Do you believe they are advertising to lose money? Are they trading against you "hoping" or "wishing" that they will be successful? No, they know they have a edge with positive expectancy. They now that the vast, vast, vast majority of their customers are trading with negative expectancy. All they need to do...
HI FX-Petra:

And if you beat them often enough, they will not let you trade with them.
If you only trade long, then you can't always be wrong.
  • #311
  • Sep 1, 2009 2:51am Sep 1, 2009 2:51am
Random?

http://stockcharts.com/charts/histor.../djia1900s.png
  • #312
  • Sep 1, 2009 2:59am Sep 1, 2009 2:59am
Quoting TickJob
Random?
That picture also tells quite well why Buffett is held in such a high esteem - in a world where a market has a bias a man with a clear bias is a billionaire (he was born in 1930).

Forex is different than stocks though.
  • #313
  • Sep 2, 2009 8:02pm Sep 2, 2009 8:02pm
Quoting FX-Petra
Well, it seems all the comments about this topic have been posted. I will try to sum up what we have learned and see if there are any conclusions that can be drawn:

The core of the problem seems to be that same old vexing problem: Is the market random? There is a yes crowd and a no crowd.

If the market is random then I think we can conclude that long term winning traders will be very few. If we cannot predict market direction correctly to a statistically significant degree, then all we are left with is our money management and the volitility...
I think there are a couple of obvious givens in all this. There appears to be overwhelming evidence that trading for a living does not work for a statistically significant percentage of the population. That right there is telling us something. So what is wrong?

There is only one thing left IMHO and that is to discover some sort of new tool that will give a good probability for finding the next rare and dramatic moves in the markets for any scale that your looking at, 5min, daily, weekly, yearly whatever. It's all 80/20 with scalability.


If you argue with the first paragraph then think about the following.
It's obvious due to all the trading packages, charting software, new inventions to find an edge, books on the shelves and most importantly of all the people who try and just plain right out fail. The only people right now that are guaranteed a profit is DUKE and DUKE who get the commisions.
  • #314
  • Oct 21, 2009 7:24am Oct 21, 2009 7:24am
Quoting TickJob
Random?
If you took inflation out of the equation...

http://www.stock-market-insight.com/images/uscpilog.gif
  • #315
  • Oct 21, 2009 7:26am Oct 21, 2009 7:26am
Quoting FX-Petra
Well, it seems all the comments about this topic have been posted. I will try to sum up what we have learned and see if there are any conclusions that can be drawn:

The core of the problem seems to be that same old vexing problem: Is the market random? There is a yes crowd and a no crowd.

If the market is random then I think we can conclude that long term winning traders will be very few. If we cannot predict market direction correctly to a statistically significant degree, then all we are left with is our money management and the volitility...
Wow, that is the most common sense I have heard for a while on this forum, and for that Sir I salute you!
  • #316
  • Feb 3, 2010 1:51pm Feb 3, 2010 1:51pm
[quote=FX-Petra;2961586]

Long Term Winning in a Negative Expectancy Game


...... If you go long E/U with a 2 pip spread, you must go a full 10 pips higher for a 10 pip gain, but you need only 8 pips lower for a full 10 pip loss. I hope that everyone can see that, from a probability standpoint, forex is negative expectancy.......

------------------------------------------------------------------------------------------------------------------------------------

First of all, a small mistake, when long E/U, assuming 2 pips spread, to win 10 pips profit, the price needs to move up 12 pips, instead of 10 pips, because the bid is 2 pips lower than ask.

therefore, a system declares a 10 pip TP, and 10 pip SL, is actually fighting at a 12 pips distance to TP, while a 8 pips distance to SL.

The ratio is drastically changed from 1:1 to 3:2, this difference, kills most of so called "profitable" system.

Of coz, this makes the situation worse, however, I doubt that "long term" FX trading is a negative expectation game. My guess is not that pessimistic. Profitable systems for retail fx should exist, just very very difficult to find.

1. to overcome spread, i guess you have to trade long-term, or long distance of TP and SL, say 200 pips to TP or SL, so that 2 pips of spread only proportion of 1% or 2% of a trade's P/L, instead of converting TP vs SL from 50% vs 50% to 60% vs 40% explained above.

2. we don't believe that there are profitable systems because we don't have one or see one. I still think that there should be a profitable system, but i don't have proof, maybe because i just wouldn't quit this tragedy.

finally, all your observations are golden, and right, except the point above.

i feel sad for i have not realised your points earlier. sometimes it takes years, surprisingly, to understand these simple facts. However, again, i wouldn't believe your conclusion though.

Still seeking...
  • #317
  • Mar 8, 2010 9:56pm Mar 8, 2010 9:56pm
Quoting FX-Petra
"Random" is a loaded word. Understanding randomness is a topic phd dissertations are built on. I don't care if it is random, I just want to know if it can be predicted correctly more than 52% of the time, give equidistant profit/loss points. If anyone can do that over a large number of trades, let them step forward. I do not believe it can be done.
hi FX-Petra,
check out this book," Encyclopedia of Chart Patterns" by Thomas N. Bulkowski
I think you may find it interesting as you are thinking about probabilities and odds. I have the pdf version is case you are interested
  • #318
  • Apr 12, 2010 8:52pm Apr 12, 2010 8:52pm
Thanks, I have heard of that book before, but never read it. I will give it a look. This thread continues to be read, as I get frequent emails from folks asking me questions or making a comment. If it gives someone a moment of pause before they put down their hard earned money, thinking "How hard can this forex thing be?".... then I am glad. I have been doing some research lately focusing on the first 1000 trades of any system. I will not go into all my findings here, but I will say that your first 1000 trades yeilding profit has almost as much to do with staistical variance as it does any small "edge" a trader might have. That is assuming any retail day trader has an edge (this is a BIG assumption). Oh well, perhaps I will start another thread one day to discuss this new area that I am looking into. Good trades to you and kindest regards,
  • #319
  • Apr 13, 2010 1:25am Apr 13, 2010 1:25am
Here is Mr Bulkowski's web site. A number of statistics about certain patterns are posted there. They're not necessarily forex-specific, however, so interpret them however you see fit.
  • #320
  • Apr 13, 2010 4:04am Apr 13, 2010 4:04am
Quoting hanover
Here is Mr Bulkowski's web site. A number of statistics about certain patterns are posted there. They're not necessarily forex-specific, however, so interpret them however you see fit.
His research did not include the risk reward ratio. Is that means RRR = 1? Probability itself does not tell the whole story.
0 traders viewing now