Raiden wrote:
Yes, exactly, and you can see how much this extra is from the net effective rate of 0.91484 USDCHF while you could have just gotten it at 0.91373. You are paying an extra 10.1 pips.
Thank you and Jeuro for getting it so quickly. No offense to others as it took me a while to figure it out when I was working it out myself.
.........
And this needs to be cleared up. If you sell 1 lot EURUSD, buy 1 lot EURCHF, sell 1 lot USDCHF. Your effective cross is a smaller buy USDCHF position, not EURCHF.
With the current random entry, the system will lose eventually when the USDCHF trends in the opposite direction of net USDCHF trade. And negative swap will eat up the position gradually too.
And even with a selective entry (overbought/oversold analysis), there's no point to pay the extra 10+ pip spread rather than just take the direct USDCHF trade.
Now, if you truly hedge the triangle trade with the proper lotsizes, with a random entry with no analysis of divergence, you will be stuck in the position forever and not be able to exit in a profit.
If you do make a divergence analysis, which does occur occasionally, the 1-2 arbitrage pips will be swallowed up by the slippage. Take note that you also have to cover the total spread of all 3 positions. The risk of a requote also throws the whole system out the window when it happens.
Take it from a guy who's worked this to the bone.
Raiden,
Many, many thanks for your reply; you've explained everything very clearly and concisely. Confirmed all my suspicions about hedging; it's always seemed sensible to me that no 'edge' is possible without performing some kind of analysis, and no return possible without taking commensurate risk. If there was an easy shortcut to making money at forex (for the retail trader), surely it would be more widely known; and of course the market by its very nature is quick to close 'loopholes' or inefficiencies.
I could tell from the spreadsheet image in your first post that you'd put a significant amount of research into resolving these triangular setups. It's always been a policy of mine to evaluate problems by first breaking them down atomically into their simplest components. The result of spending too may years as a software developer, I guess!
David