Thanks Medisoft. Hey timehopper I am still on.
Iām carrying a trade on AUDCHF/NZDCHF since Friday, and still canāt close it. The problem with it seems to be with the spread. I opened it on Friday near to end of trading hours, and I got a HUGE spreads, about 11 pip on AUDCHF and about 16 pips on NZDCHF. The accumulated loss on that trade is about 60 pips, and it is varying from 50 to 70 pips, without exceeding that range.
Iām looking for a way to rebalance positions on this type of trades to close them with profitā¦ Do you have some suggestions?
I decided to add more cash to the system, because with only 10 usd, it cannot balance the positions. The minimum cash amount to balance the positions with a mini account with 0.01 lots for minimum is 200 USD, so I increased the size of the account to that value, and now I can balance the positions 
This isnāt a lot of help with the existing trade, but the only long term solution I can think of would be to trade the AUDNZD. Then you have only one set of transactions costs that cannot fluctuateā¦
The problem with AUDNZD is with that method I canāt balance the size on AUD and NZD differently. And I found that there are times when a well balanced AUDCHF/NZDCHF works better than AUDNZD.
The problem is on my real account that had only 10 dollars (well, 13 dollars with this week profits hehehehe) and my risk didnāt permit enough lot size to balance the positions. On a demo account with plenty of virtual cash this problem didnāt happened
Another thing Iām now doing, if the combined spread of the two pairs exceeds 11 pips, then I didnāt take that trade. I need that the spread of the pairs be small enough, maybe 5 pips, to allow a trade and to win the trade easy
With 200 USD on a mini account I can do some balance, enough to avoid this problem
The 99.7 rule only applies when the underlying data is normal. Market data is not normal but has fat tails (heavy-tailed), indicating that larger standard deviation moves happen as a rule rather than as the exception. Based on the vast amount of evidence regarding the nature of market data it is not appropriate to assume normality. Here is an excerpt from wikipededia:
In statistics, the 68-95-99.7 rule, or three-sigma rule, or empirical rule, states that for a normal distribution, nearly all values lie within 3 standard deviations of the mean.
And then this excerpt with highlights:
Quantities that grow exponentially, such as prices, incomes or populations, are often skewed to the right, and hence may be better described by other distributions, such as the log-normal distribution or Pareto distributionā¦ As a result, statistical inference using a normal distribution is not robust to the presence of outliers (data that is unexpectedly far from the mean, due to exceptional circumstances, observational error, etc.). When outliers are expected, data may be better described using a heavy-tailed distribution such as the Studentās t-distribution.
The inference is that if the underlying data is not normal (market data) then you can expect a much greater deviation from the mean as a matter of course. How much greater? It is fairly trivial to measure the actual excursion beyond the 3rd (or 2nd/1st) deviations and compare them with the expected 68-95-99.7 Rule to determine how close to normal the underlying data appears, and then to adjust expectations accordingly for risk management purposed / position sizing.
How many periods do you use to calculate the standard deviation?
Did anyone try to trade that wild move yesterday? I did just for shts n grins. Closed it this morning for a 1% loss. Thatās the only loss Iāve had with this so far. Itās a high win rate strategy. Just have to be careful you donāt end up with lots of small wins and a few big losses.
I was at my screens listening to the squawk when the the BOE news was announced. I then sat and watched the resulting charts move. It was really hard to avoid the temptation to enter, but definitely the right decisionā¦ 
I was already in before the big move. Still Waiting to ADD to the position as soon as the GBP starts a decent downward move. Iām personally looking for a move below the 1.60000 range. just a question but am I the only one who trades with real money?
Anyway, I will post the outcome of this trades if anyone wants me toā¦
on a side note my aud/cad/usd trade closed at a wonderful profit.
So everything I said was right. most of your post was a paraphrase of what Iāve said. (you even quoted me saying āunder normal situationsā
Thanksā¦ of course you can expect some deviations to occur out side on 3sd. that would fall under the .3% of the time. LIKE THE RULE STATES And the SD is always changing other wise prices would occur within a set range all the time. The event yesterday was clearly not normal.
The only thing you said that was different from what I have been saying is that āDATA MAY BE BETTER DESCRIBED USING A HEAVY-TAILED DISRIBUTIONā Okay sureā¦ if you want to use that method go for it. no argument there, although I wonder how many people reading this post will understand that method. If they do then by all means go for it. let me know how it turns out please. One can throw darts at a chart or use MA crossovers or b.Bands or what have you.
And like I have posted to me this is a great opportunity to enter more into the position. I have said that my money management is to allow for a 1000 pip spread to blow my account out. not likely to happen. as a trader using stat arb for a very long time I havenāt loss. Period. Real money, no demo. Iāve come close. a pip or less in some cases. very few home runs of 100 pips or more. and there are others who have not lost using the overlay version of stat arb on this very post. bet small and wait awhile is the key. But lets assume yesterday led to a loss because the position taken was too big. which should be the only case since we WANT big deviations to occur. the bigger the better. OKAY one lost out of many winsā¦ so whatās the win loss percentage in that case? 1 loss out of what 10? 20? 30? I can live with that. Bears make money, Bulls make money, Pigs get slaughtered. My position was small just like always, and it still open. the only way I will loss on this trade is if they never converge again.
Tell me what are the odds of that?
Would love to see how it is working for you with real money! Iām playing demo still!
sureā¦ apr. 17th @1657 eastern short GBP/USD
apr. 17th @ 1658 e. long EUR/USD
currently as of this posting -133.6 pips gbp
-1 pip eur
I think what he was trying to say is that the data is not normally distributed. This is why there are more moves that go to 3, 4, 5+ SD. Historically, you can see that there are many moves that move this far, and if the data were normally distributed, the chances of this would be so small, it should not be this common. I havenāt analysed the data (although I am sure that would be interesting and revealing), it was just a gut feel looking at the chart, but it looks like it occurs more than 0.3% of the time.
Another way to look at this is, I could randomly open any position with absolutely no strategy, and providing my position is small enough, I can simple wait until it becomes profitable. Is that not basically what you are doing?
take the last 50 spreads (the spread only) and measure the standard deviation OF THE SPREAD. if itās over 3 sd and you had pulled the trigger what would have been the outcome of those trades? on whatever time frame you use. Im lost as to how come it works for me and others but not for him?
and as strange as it sounds yes one could in theory enter a small trade randomly with no strategy and simply wait. there is a trader sitting directly across from me at the moment who almost does just that. She trades with no stop and her positions are small RELATIVE TO HER ACCOUNT SIZE. to my horror she routinely has trades go against her by 100 sometimes 200 pips or more using a 4 hour charts. oddly enough she rarely suffers a loss. thatās only because of a thing called MEAN REVERSION. More often than not the price action will revert to the mean at some point. That is proven by people much smarter than me. And then she takes a profit. But somehow it never works for me. Go figure.
But to answer your question yes and no. i use a small position in case something unexpected in the market happens like yesterday or if the spread continues to widen. my drawdowns are only a problem if it continues to widen, in which case I add more to the position. If my position was too large to start with I would not be able to add more in that case. Someone on this post had that problem yesterday. Once the spread starts to close my drawdown drops rapidly because one leg of the trade will start to become positive and offset the drawdown of the other leg. In most of my trades one side will be a loser but the other will be a winner by larger than the loser and the cost to put on the trade itself. you have to wait for that to occur. but there is always a target exit in mind so Im not just simply waiting for some unknown event. But iām waiting none the less.
What mean period are you using for calculating the standard deviation?
I had thought about doing this, but figured that in order to hold a position open indefinitely, regardless of the equity drawdown, the position would have to be so small, that the profit would also be equally small. Waiting 2 months for a trade to return me Ā£200 profit isnāt going to earn me a livingā¦
Taking it a stage further, the only way I see being able to earn a living doing this would be to open up say 10 trades a week, so that over a rolling period there were 10 trades closing out each week. These positions would potentially have been open for a long time and each would return Ā£200. However, the margin requirements for all these trades would be huge so the underlying account would have to be absolutely massiveā¦
TO BE CLEARā¦
There are lots of systems out there that work. Thatās how they became a system in the first place. using a heavy tailed distribution works, if you understand it. using m/a, momentum, Rsi , whatever will all work for some. I once posted that almost all systems work if you stick to the rules of that system whatever they are. So one could find other methods that work for themā¦ And that is a good thing in my book. I donāt want to seem like Iām against that. Im simply trying to show that any deviation to this system is no longer this system. if it works for you AWESOME, BUT IT IS NOT THIS SYSTEM ANYMORE. this method as intended is a high probability, low risk system. But with low risk comes low reward. Will there be a time when this system doesnāt work. SURE. the reasons could be many. human error, bad position sizing, missing you exit opportunity (although there is a workaround to that but itās not available all the time and can be a bit complex) or the underlying currencies could default. (think Russia and Long term capital management) other than those you really canāt fail provided you wait long enough. I do the math by hand to avoid errors, I use a 1000 pip spread as my blowout and depending on my schedule I vary the time frame in which I trade so I donāt miss the exits. If a country defaults and you had no idea it was coming you should not be trading.
If the profits come too slow for your taste, you should not use this method as time is a key factor. If you canāt handle the drawdowns adjust you size or choose a different method.
I have read alot of the methods and systems that can be found on this website over the last year or so. There are some awesome ideals and theories on here. I only choose to make postings on this method because itās very close to how I trade now and have been for years. Thatās why my comments cannot be found elsewhere on this site.
I asked on several occasions what happened to Kelton. He has not made a response other than the few that began this post. I think I have figured it out. My guess is he is on a boat somewhere relaxing because he has a high probability system that WORKS AS IS. At this point I would like to follow his example and no longer make any further post regarding anything other that what his system was. My only posting henceforth will be the position I have taken and the results when they close with one update in while the trade is still open. For those who have question they would like me to answer concerning Keltons method please send me an Email and I will be more than happy to address you in that way. It is getting kinda boring with the whole this is better than that or donāt use this math use that, or what about this onetime random event over here. If you canāt handle one time events or one loss out of 20 you SHOULD NOT BE TRADING FOREX.
THERE IS NO HOLY GRAIL IN FOREX. IF YOU WANT A 100% PROFIT, NO RISK , NO WAITING TRADING SYSTEM, CLOSE YOUR FOREX ACCOUNT, OPEN SEVERAL COMMODITIES ACCOUNTS OR FIND A BROKER YOU WILL LET YOU BUY AND SELL ACROSS ALL MARKETS AND USE PURE ARBITRAGE. For those who donāt know itās when you notice a price difference in the same commodity on two different markets. You buy the low priced one and sell it on the market where itās worth more instantly. AS LONG AS YOU BEAT THE COST OF THE TRADES itās instant profit zero risk . Alas this cannot happen in fx so I use itās cousin STAT/ ARB.
ciao
Lets say the 10min time frame. scroll back to be able to see the last 50 times there was a deviation. measure the deviation in pips each time it occurred. use the 50 data points you just go and do the rest of the math to get the mean, use all of that data to get the SD of the spead.
āsmall RELATIVE TO HER ACCOUNT SIZEā Iām going to go out on a limb here and say that her account is different in size than yours. She too is a full time trader as is everyone else in this room. Suffice it to say if sheās doing it full time sheās making full time money.
I only responded because you asked before I made my last post. Much success to you in all your trading my friend.
Keltonās last post was March 28th, my guess is K is wiped out or still waiting for āMEAN REVERSIONā on his March 28th & 29th trades? LOL
Take a look at a daily EUR/GBP chart, highlight March 28th through April 8th or 9th, think about how long youād wait for āMEAN REVERSIONā on that one? :56: