TheEconomist

Well guys, you are right. I found the fallacy. Pretty tricky. Say a moving average of 4 values is comprised of: ( 1.4501 + 1.4502 + 1.4503 + 1.4504 )/4 = 1.45025 If it comes a next value, say 1.4510 , it changes: ( 1.4502 + 1.4503 + 1.4504 + 1.4510 )/4 = 1.450475 a growth of 6 pips of last quote produces 2.25 pips in MA gain. Supposing I had trades open on the first quotes. Then results are (supposing no spreads) The first trade of 1.4501 closes at 1.4510 with a profit of 9 pips ; on a quarter of volume, 9/4 = 2.25 = 1.450475 - 1.45025 ; the other open trades have an unrealized value of (8+7+6)/3*(3/4) = 5.25 pips. The trades emulated the MA; however, on the other side, there are the sells, opened at Bid prices, always higher, putting the hedge on loss. I thought that "refreshing" the MA would keep it going like it should, as I knew that once the MA stops , the hedge would be futile. But it is futile from beginning, due to the opposite MA hedge. Average is right. There is no arb potential. There will not be a BOTH SIDES PROFIT if the market heads towards the MA, like in oil arb crossover, for example. They will absorb each other and produce losses with spreads. This is the real trick: to have a hedge like it was put some time before, but without endangering the portfolio, neither being futile. Later edit: this sci-fi hedge would create a profit at the very time we were thinking to trade. Crossover won't be needed, just a close is enough to get profits. Perhaps this is the TRUE FALLACY : USING ONE ASSET. We should stick to arb on every step we make.

Thanks for the appreciation, I'm trying to do my best and keep thinking outside the box... Regards, Bogdan

I thought about this system several months ago. I wasn't able to find a suitable implementation, because I tweaked it to the point I completely screw it up, so I'll put it the way I initially thought it. Trader's Fallacy: Hey, I can follow that trend and make cash like a pro! Traders love charts. Perhaps our reason for entering stock , forex or other markets was watching trending markets and simple indicators. The simplest indicator that someone gets in touch with when entering forex or stocks is the Moving Average (MA). Look how beautiful it is: http://img259.imageshack.us/img259/4120/beautifulmatk0.jpg You always used the MA as an indicator and always it tricked you! Damn whipsaws! But the MA is pretty neat. Whatever happens and wherever far the market will go, eventually it will slow down and be catch up by the MA. Haven't you ever dreamed that the MA to be an asset to hedge against the market itself ? Say go short the market, long the MA, and get profit on cometogether? Look at the highest peek, it's so far from that 200 values MA at about 183 pips! But how in the world could we trade that way...?! The value of a MA is roughly the arithmetic average of the last N values (We consider Simple Moving Average). So if we have N trades, done at these prices, this composite asset will have been traded at about the MA value! Well, you may say...how we do that? How do we know if we buy the MA and sell the market or we do the reverse ? I have an answer...Good news and bad news. Starting with the bad news: it's gonna cost. Not a killer cost, but it will be. And the good news: you will be able to trade both ways, without needing to know what will happen and without being whipsawed! Not knowing what will happen It's gonna be either BUY or SELL. But until then, it's gonna be both: You construct the moving average, every bar, until it's ready: on every bar, buy and sell the Nth part of the position you take. Once the moving averages, completely hedged of course, will be ready, on every bar you will close the oldest pair buy/sell and place a new pair of buy/sell, keeping the moving average "assets" in line with the moving average itself. How much it will cost ? Well, you get slipped on both placing and closing trades, and you will pay the spread on both buys and sells. That means that after the first N bars you will have a spread cost of about 2 x spread (for buys and sells) Any other series of N bars came in will produce this cost, plus the slippage costs: ((2 x average slippage) for opens and closes ) x 2 for buys and closes) Now once the basis (market minus moving average) jumps over a fixed , predefined value (say 50 pips), we will trigger the "trade": say the market trades at 1.4503 and the moving average is at 1.4320 . That means, sell the market, buy the moving average. We keep the series of BUYS, since its ready, and close the sells. Nothing happens, we were hedged. We place another sell at about 1.4503 , a huge trade, by the volume equal to the series of closed sells. We continue to "refresh" the moving average, replacing the oldest buy with a new entry on every bar. When the market will join the MA, we will have a fixed profit on the SELL side, some profits made by current BUYS and closed profits of the older BUYS. Of course, the extra spread and slippage costs for maintaing buys will be added. We close all trades and start rebuilding the moving average. Until the moving average is finished, no "trading" will be allowed ; at the moment when this is ready, we barely start up waiting for a new "trade". Problems to think about 1. Enlarged spread Once you take a "trade" you will be exposed to the basis going higher thus generating virtual losses. Take care not touch the margin call! 2. Broker limits You need a broker that allows microlots and an outrageous number of trades. More trades possible, the bigger moving averages used, the higher number of pips caught! 3. Market delay If the market will come together with the MA too late, hedge maintaing costs may be significant! I'm awaiting comments!

Good point about checking by mail the swaps! At least you don't wire the money just to see the bank fee lost and no arb opportunities. Now about FXDD experience and swap volatility : For one of my friends, when we tested, SwapFinder recommended and a traded a 7-currency ring (sorry I no longer have the picture). Then I found out that, if I run the script over 20 minutes, I get other results! It seemed there were just a few "versions" but only one version would exist at that split of a second when swaps are wired. And if it's not the one you traded, you won't feel too good! Regards, Bogdan

Hi, Andi It would be nice to start designing the Bond Arbitrage. In our terms, so we could understand it. But, for this I need lots of data... And I have nothing now! Regards, Bogdan

Sure, I'm thinking at this too, especially for the next investment cycle. Regards, Bogdan

I heard they aren't so smart. Many can't even see the hedge in Swap Arbitrage. This could work. However, results will decrease a lot, because of the higher margin requirement on CFDs. Look here: Required margin is like : 1/3 (fx) + 2/3 (cfd) = 1 ; "working capital" = 1/3 = 0.33 ; results are considered 100% here; If you use two cfds: 1/6 (fx) + 2/6 (cfd) + 2/6 (cfd) = 5/6 = 0.83 ; "working capital" = (1/6)/0.83 = 0.20 ; results at 60% (0.20/0.33*100); If you use two fx: 1/3 (fx) + 1/3 (fx) + 2/3 (cfd) = 4/3 =1.33 ; "working capital" = (1/3)/1.33 = 0.2481 ; results at 75% (0.2481/0.33*100); Better use two fx rates and a cfd, but see the impact of swaps! *Please note I chosed the fractions to simplify explanation: the totals, as 1, 0.83, 1.33, don't necesarily mean margin used will go up, or down, unless you allow that; Given your margin used to be constant, use the numerator and denominator only to find out the internal margin allocations. Regards, Bogdan

I heard they aren't so smart. Many can't even see the hedge in Swap Arbitrage. This could work. However, results will decrease a lot, because of the higher margin requirement on CFDs. Look here: Required margin is like : 1/3 (fx) + 2/3 (cfd) = 1 ; "working capital" = 1/3 = 0.33 ; results are considered 100% here; If you use two cfds: 1/6 (fx) + 2/6 (cfd) + 2/6 (cfd) = 5/6 = 0.83 ; "working capital" = (1/6)/0.83 = 0.20 ; results at 60% (0.20/0.33*100); If you use two fx: 1/3 (fx) + 1/3 (fx) + 2/3 (cfd) = 4/3 =1.33 ; "working capital" = (1/3)/1.33 = 0.2481 ; results at 75% (0.2481/0.33*100); Better use two fx rates and a cfd, but see the impact of swaps! Regards, Bogdan

SOBs! Now you can clearly see the evil promoted by NFA. Nobody likes hedgers and arbs. And once that the entire retail market has been taught that trading must be normal (to be read: speculative, without knowing the final result) , it will be hard for someone to teach them give up this approach! What we should do is simply to migrate from MetaTrader and FX brokers to real futures brokers. But damn we need cash for that! But with real futures brokers you would not find the conditions to do this arbitrage... Sure, Jim, but we could use the either the Spot-to-Futures B - the missalignment version - (if possible) or Oil Arbitrage with them! This is an open request to all my readers: anyone traded municipal bonds ? If yes, with which broker, account sizes, margins, platforms, etc... Regards, Bogdan

Well, a yield of 41% surely beats the hell of the Barclay's Hedge Fund Index http://img222.imageshack.us/img222/4685/tradereportoct26ou2.jpg This result was at a DiscountedBasis value of about 2 points. I have to mention that starting from morning of Oct 25, right after Globex respawn, the DiscountedBasis indicator, which was working at -5...-3 for about two weeks jumped around 0 and is averagely positive, meaning that the 6J future (direct quote) is slightly underpriced, compared with last week, when it was a bit overpriced (compared to the theoretical value). That means there is still room for average length arbitrages, shorter than a month. The strategy risks are practically extinct. But the moral hazard risks?.... Don't ask!

SOBs! Now you can clearly see the evil promoted by NFA. Nobody likes hedgers and arbs. And once that the entire retail market has been taught that trading must be normal (to be read: speculative, without knowing the final result) , it will be hard for someone to teach them give up this approach! What we should do is simply to migrate from MetaTrader and FX brokers to real futures brokers. But damn we need cash for that!

Salut, Fata de cele relatate de tine, consider ca sunt necesare cateva lamuriri... Consider ca analiza tehnica trebuie completata prin cea fundamentala. Miscarile ample ale pietei sunt determinate de factori macroeconomici care pot fi identificati si cuantificati numai prin analiza fundamentala. Sigur, puteti spune ca intr-o strategie intraday nu trebuie sa cunosti trendul macro al perechii valutare atat timp cat STO si RSI arata supracumparare... deschizi pozitia si in cateva ore o inchizi pe profit (ceea ce NE si doresc) sau in SL... Interesul meu pentru analiza fundamentala apare in momentul in care pozitia imi merge in pierdere... daca directia pierzatoare este in trendul macro al perechii valutare in mod cert voi inkide acea pozitie. Daca totusi directia pierzatoare este contrara trendului macro, pot astepta (functie de riscuri si moneymanagement) ca acea pozitie sa revina pe profit. Revenind la serviciile in speta... Avand in vedere ca analiza fundamentala necesita mult mai mult timp, timp pe care nimeni nu il are, va intrebam daca ati gasit un srviciu care sa imi indice (eventual pe un TF saptamanal) o directie probabila a trendului, bazandu`se pe analiza fundamentala a unei perechi valutare. Nu vreau sa-mi spuna unde sa intru, nivel SL etc., vreau o simpla sageata care sa indice (cu o probabilitate convenabila) sus, jos sau level. Asta nu inseamna ca voi urma orbeste indicatiile lor. Ar fi insa o completare la AT-ul pe care il pot face si singur. Stima, Necre Daca vrei niste liniute, poti sa folosesti DashBoard FX de la fxuniversal... Asa, un serviciu de semnale pentru MT este la www.forex-signal.ch . Scrie in germana, dar poti sa trimiti mail si in romana. Executia este automata din MetaTrader , dar semnalele vin de la un trader "discretionar". Daca ai vreo problema, vorbesti cu ...moi . Eu am scris 85% din expertul de executie client. Bogdan

Seems Windsor Brokers is quite a pain! At least with Orion you don't even need to ask for the swap free regime. It's by default. However, the margin is pretty big. They seem to operate in the middle east. That's it...it doesn't seem too safe, but the swap free regime was invented in the middle east! Choose wisely: look carefully at the broker, reviews & reports from clients, simplicity of access to swap free regime, leverage and futures margins. I can't tell you which one to choose. I seem to be working well with FXIndia..for now. Later edit: as you see, my fee at 1.5 lots is $30...so it would be $15 for a lot. Regards, Bogdan

You are both wrong and right. You are wrong, because he cannot choose. The cfds are simply mirrors of the Globex futures. Once the Globex futures switch, the cfds do it too. You are also right, you must be wary on the close moment. It may be possible that he closes after, wiping all your profit. But if he does that, other arbitrageurs that know this will happen will launched reverse trades right before it happens and get their money doubled. A good trading policy is: 1. close safely and early on december 13, may be even 8 hours before Globex shutdown; 2. don't attempt to instantly double your cash. Regards, Bogdan

I have mistaken my calculations...considered a JPY pip of 10, and a day result of $16, but the real pip is more than 8, so the output is about $12-$14 . Bogdan, Thanks for your answer. AFX uses different Futures contracts set up with no expiry date so it is hard to calculate how close they converge with the Spot prices. Have a look at this Post #890 As far as I know, SpaceVision has only standard lot sizes of 100 k. Regards, Mr. H Slightly different...probably it includes a premium (later edit: I heard they have variable spread on futures, and the broker is libanese). It doesn't mean it won't expire. It copies the real futures, even if slightly different. If the broker doesn't closes the futures at expiry, it can be doublecrossed big time. For example you could open the trades reversed a few minutes before Globex futures expiry, the futures jump, and double the money after an hour. I don't recommend this to anyone, the broker would automatically shut this down, because his real futures are closed. Regards, Bogdan

Well, I don't even wanna talk about this subject too much. FXIndia is just an IB. The operating center is Masterforex , owned by BeaverheadFinancial Now, FXOpen says they don't have any connection with Beaverhead, although , if you look on their website, a past note says that the company has been renamed from Beaverhead to Poltek, and also warn customers of a scam email originating from Beaverhead. What is Beaverhead then ?....Gee, I don't wanna know. I just wanna make money and that's all. A friend had an account there, made cash and withdrew. Doesn't look too normal, but it seems to behave so. As long as they respect me as customer, I don't have any problem with them. Other brokers to check up may be AFX (I just heard about) and SpaceVision. SpaceVision seems similar to Crown, displays itself as swiss, but I heard it's middle east based. Regards, Bogdan

Bogdan, Thanks for your answers. Yes, I noticed that the growth is not consistent day-by-day but by the Futures expiry, profit can be realized. I have not looked at Futures.MQ4 yet but my idea is for example: Do the Calculation manually then Set up the following in a script / EA: Sell X lots of USD/CHF Spot at Market BID and Sell Y Lots of CHF/USD Futures at Market BID Once all orders filled then Close all open positions at a specific exit date and time. Sorry Bogdan, I don't know what you mean by "However, it would be interesting to find out how much that fixed fee is. Because, if it's fixed, may be a way around it." Regards, Mr. H You're right. If you want to trade using similar risk parameters like me, consider a margin usage of 50%. Suppose your broker allows max leverage of 200, calculus would follow this: X = ( Account equity * 200 * 50% )/100.000 Z units of bought CHF = X*100.000*(USD/CHF bid); Y=Z/125.000 As for the closing moment, that one should be on december 13. The day before the one you can see in WHC's MetaTrader (don't know, perhaps you have it on Windsor Brokers Metatrader too). Now let's settle out that daily fee problem. If the daily fee is fixed, it means that the cost of maintaining the position is for the entire period to expiry, no matter which forex/futures pair you use, be it the best or the least profitable, and also, no matter how the interests rate change until that time. Now, what "flat" exactly means, is something I can just suppose. Say this would behappening to me: My strategy should make, in average , $16 a day, as you see, I sold 1.6 lots of USDJPY, and, at that time, there were aproximately 70 pips left to go (after cutting a few pips as error and a few pips for the commission and spread) for about 70 days left to expiry. So, one pip a day. Suppose I'd have to pay $5 a day, fixed fee. I should stand up to it pretty easily. Even if I'd trade double, and make $32 a day, it would be no difference, the "fixed" thing would still be $5 a day. I've traded the best thing I had available. Supposed I would have chosen to trade EURUSD/6E, with, let's roll the dice, 40 pips to win / 92 days to expiry. At the same 1.6 lots EURUSD, it would have meant 40/92 x $16 = $6.88 a day. Dangerously closed to the $5 daily costs. Almost nothing, but the profit would help me get a bit farther from these costs for the next cycle. Now, what if those $5 a day would be "per lot" ?. For my trading, it would mean $5 * 1.6 = $8. This would have wiped out the EURUSD/6E strategy and would have eaten 50% of the USDJPY/6J strategy output. Even more, the $8 cost would double in the next trading cyle, so it would be at a rate of 50% of the average profit, whereas by being absolutely flat, the $5 cost would become insignifiant over time. You have to find out all the terms linked to the flat fee if you want to use this trategy. Regards, Bogdan

You're right Andi, I have kinda forgot posting results... http://img81.imageshack.us/img81/7180/tradestatus1710xf6.jpg After a jump last week from around $150 - $250 now it's consolidating in $300 - $350 area. The time it happened the DiscountedBasis indicator jumped over 0 ; generally it's negative (-1 to -7, average -3 to -5), except for the unhedged hour when Globex is shut down.

So far, so good... http://img73.imageshack.us/img73/8026/mytradingoct102007dl8.jpg

It seems when I inserted the MaxPasses parameter in ReliableOrderSend and ReliableOrderClose they became really broken. Last correction solves the issue, however there is still a bug: Examining the corrected codes of the functions: copy from here and replace in BcLib.mq4! int ReliableOrderSend(string symbol,int cmd,double volume,double price,int slippage,double stoploss,double takeprofit, string comment="",int magic=0,datetime expiration=0,color arrow_color=CLR_NONE,int MaxPasses=0) { int Gle=ERR_TRADE_CONTEXT_BUSY; string ErrType=""; string ErrText=""; string ErrExplanation=""; int passes=0; int res=-1; while (Gle==ERR_TRADE_CONTEXT_BUSY||Gle==ERR_REQUOTE||Gle==ERR_INVALID_PRICE||Gle==ERR_PRICE_CHANGED||Gle==ERR_OFF_QUOTES) { if (Gle==ERR_REQUOTE||Gle==ERR_INVALID_PRICE||Gle==ERR_PRICE_CHANGED||Gle==ERR_OFF_QUOTES||passes==0) { if (passes!=0) RefreshRates(); if (price==0.0) //if (passes!=0||price==0) price=MarketInfo(symbol,PriceOpenMode(cmd)); }//if (Gle==ERR_REQUOTE) res=OrderSend(symbol,cmd,volume,price,slippage,stoploss,takeprofit,comment,magic,expiration,arrow_color); Gle=GetLastError(); TranslateError(Gle,ErrType,ErrText,ErrExplanation); if (Gle!=ERR_NO_ERROR) Print("ReliableOrderSend error : ",Gle," : ",ErrText); passes=passes+1; if (MaxPasses!=0) { if (passes>=MaxPasses) break; } //ADD ONLY THIS LINE if (Gle==ERR_REQUOTE||Gle==ERR_INVALID_PRICE||Gle==ERR_PRICE_CHANGED||Gle==ERR_OFF_QUOTES) { price=0.0; } }//while (Gle==ERR_TRADE_CONTEXT_BUSY||Gle==ERR_REQUOTE) return(res); } bool ReliableOrderClose(int ticket, double lots, double price, int slippage, color Color=CLR_NONE,int MaxPasses=0) { int Gle=ERR_TRADE_CONTEXT_BUSY; string ErrType=""; string ErrText=""; string ErrExplanation=""; int passes=0; bool res; int otype; double olots; string osymbol; res=OrderSelect(ticket,SELECT_BY_TICKET,MODE_TRADES); osymbol=OrderSymbol(); otype=OrderType(); olots=OrderLots(); if (lots==0) lots=olots; if (res==True) { while (Gle==ERR_TRADE_CONTEXT_BUSY||Gle==ERR_REQUOTE||Gle==ERR_INVALID_PRICE||Gle==ERR_PRICE_CHANGED||Gle==ERR_OFF_QUOTES) { if (Gle==ERR_REQUOTE||Gle==ERR_INVALID_PRICE||Gle==ERR_PRICE_CHANGED||Gle==ERR_OFF_QUOTES||passes==0) { if (passes!=0) RefreshRates(); if (price==0.0) //if (passes!=0||price==0) price=MarketInfo(osymbol,PriceCloseMode(otype)); }//if (Gle==ERR_REQUOTE) res=OrderClose(ticket,lots,price,slippage,Color); Gle=GetLastError(); TranslateError(Gle,ErrType,ErrText,ErrExplanation); if (Gle!=ERR_NO_ERROR) Print("ReliableOrderClose error : ",Gle," : ",ErrText); passes=passes+1; if (MaxPasses!=0) { if (passes>=MaxPasses) break; } //ADD ONLY THIS LINE if (Gle==ERR_REQUOTE||Gle==ERR_INVALID_PRICE||Gle==ERR_PRICE_CHANGED||Gle==ERR_OFF_QUOTES) { price=0.0; } }//while (Gle==ERR_TRADE_CONTEXT_BUSY||Gle==ERR_REQUOTE) } return(res); } As you can see, the 0 for entry price is correctly handled. As long as a bad error occurs (such as ERR_OFF_QUOTES occurs) and there is no price sent to OrderSend() the price is reread by MarketInfo. However, if a regular error comes (such as ERR_PRICE_CHANGED) then function enters a possible endless loop, as the MarketInfo price inquiry will not be executed. This is why, before while cycle ends, price will be reset in both ReliableOrderSend() and ReliableOrderClose(), making possible its read in the next cycle execution. This is the only modification to be done. In the same time, given the fact we reset the price before the while ends, the first execution is ok in both cases with a 0 price and a non-zero given price. (Don't look at the old closing while comments, they are really old, when functions didn't handle all these errors!) Later edit: price must be reset only in some error conditions, not in all! P.S. I need some input folks! Can't see all the bugs lurking the shadows! At least for BcLib core modules... P.S. no 2 My trading account floats around 3% ROE (started from late Aug. 27).

Last week our fellow TradeLover from the forum (who is also a contestant in MetaQuotes Automated Trading Championship 2007 - let's give him an for that) made a shocking and very disturbing discovery: a tick is not quite a pip in forex (we even argued for that, but he was right - never take anything for granted) Running a simple script like this: int start() { Print("tick size= ",MarketInfo(Symbol(),MODE_TICKSIZE), " - tick value= ",MarketInfo(Symbol(),MODE_TICKVALUE)); return(0); } what he got after some hours was this: 22:40:03 ticksize GBPCHF,M1: tick size= 0.0001 - tick value= 8.53 22:40:13 ticksize GBPCHF,M1: tick size= 0.0001 - tick value= 8.53 22:40:20 ticksize GBPCHF,M1: tick size= 0.0001 - tick value= 8.53 22:40:24 ticksize GBPCHF,M1: tick size= 0.0001 - tick value= 8.53 22:40:27 ticksize GBPCHF,M1: tick size= 0.0001 - tick value= 8.53 22:40:30 ticksize GBPCHF,M1: tick size= 0.0001 - tick value= 8.53 22:40:31 ticksize GBPCHF,M1: tick size= 0.0002 - tick value= 17.07 22:40:32 ticksize GBPCHF,M1: tick size= 0.0001 - tick value= 8.53 22:40:46 ticksize GBPCHF,M1: tick size= 0.0001 - tick value= 8.53 Now what if that tick value is caught in a hedge calculus that calculates lotsizing or results for months ahead ? In order to settle this issue once and for all, I made the RealTickValue() function as follows: double RealTickValue(string contract) { double res; res= MarketInfo(contract,MODE_TICKVALUE) / (MarketInfo(contract,MODE_TICKSIZE)/MarketInfo(contract,MODE_POINT)); return(res); } The function now works as a proxy for MODE_TICKVALUE inquiries. Now both BcLib and BcArb call that first when dealing with MODE_TICKVALUE. Found surprise bugs in ReliableOrderSend and ReliableOrderClose. Don't know how they got there, as they seem to work well. Also corrected past scripts that use MODE_TICKVALUE. Discovered also mistakes in updated SwapFinder, that's why I corrected and updated again, along with SwapButterfly. Inserted also a mechanism for calculating futures using London InterBank rates. It doesn't seem to be too much of use, as indicators are still wrong, but you can correct their parameters, so I added DiscountedFuturesLIB and DiscountedBasisLIB indicators. Corrected bugs in Futures script. Now it's set up with a MarginUsage of 50% and a SwapFree regime set to True by default. P.S. I will keep you posted with my arbitrage trades working (Spot-to-Futures type A)

Well, it could work, or it may not be enough. I see that sometimes it takes a lot of time to push forward, then jumps instantly, so the growth is not necesarilly day-by-day. You can use my futures.mq4 to place the trades (supposing the symbols are the same), or you could learn the calculations and place them yourself. If you do by futures.mq4, test how it works first on a demo. However, it would be interesting to find out how much that fixed fee is. Because, if it's fixed, may be a way around it. Regards, Bogdan

Hmm, I thought I was pretty clear. Let's go a bit underground. The profit in that screen capture shows $335. Say that night, on my time, at 12:45 PM, 15 minutes before Globex "respawn", the profit, unhedged, in my account, is $670. Double. Suppose at least closing the futures is allowed. Then I could close the trades with the "doubled" profit, and reopen after Globex restart. Not only reopen in the same parameters, but also compounding. Upon restart, the spot to futures basis is re-established where it should, and my account will grow, continuously, from $335 to $1000 (say that's the calculated bound). With two "minor" adjustments: 1. it starts from $670 instead of $335 2. it goes more than slightly beyond $1000, because of compounding. Compounding is safe and honest. But closing the futures when the broker can't mimic and hedge our trading is unfair. Even if it would be allowed, it would still leave the broker unhedged , and the $335 difference from old profit of $335 to the new profit of $670 would be from his pocket. DON'T DO IT! It is against our "deontological" code as arbitrage traders to be unfair with the broker whose trading offer we use in order to arbitrage! It's like biting the hand that feeds you. So, it's a NO NO, like "doubling over night" idea trying to use the futures jump on expiry. Let's keep the things as they are. I am not teaching anyone how to cheat on the broker. If the broker sees any arbitrageur trying to cheat by letting him exposed, it is entitled to stop his trading. Yes, there are free meals. No, we don't steal the broker's meal and call that "free meal". I know Juan understands that, this is for others reading this blog that might try to use unfair arbitrage windows based on broker's mistakes of organisation. Regards, Bogdan

Hi there Mr. H, I didn't see a broker that offers NZD/USD, at least for the MT brokers area that I know. Now I won't put the problem as "if it would be allowed to hedge". You have the spot, you have the futures, you can trade both simultaneously, why not? I'm putting the problem like this : "Is it allowed to keep trades for a long period of time ?" This is a thing to ask your broker before starting up this strategy. Regards, Bogdan

Juan, I said that probably in my first post of the blog. We do, in first instance, a regulation arbitrage. We sell the regulated broker and buy the non-regulated one. It is a risk. Aren't bond arbitrageurs doing , perhaps similar things, using the most risky municipal bonds and the most safe they can access ? Of course the high-yield bond has a default risk bigger : this is why it has a higher yield. (And may I say, the spot-to-futures A looks the hell like Bond Arbitrage). This business is not absolutely without risk. There is still risk, yet we arb successfully in the idea the broker's default won't hit us...remember picking nickels in front of a steamroller. I know...you see profit doubled in half of unhedged hour, close all, then place trades back upon Globex reopen. I thought about that. Two things: 1. It may be, but not tested, a "Trade is disabled" error when attempting to close the futures. 2. Even if it may work, the broker will not have the capacity to hedge our close, Globex being shut down. So this would make the broker angry and we don't want that. Regards, Bogdan