Trianglar arb...almost there

When looking at fx arb situation what’s the best way to know where to begin the process from cursory glance, overvalued or undervalued pair? I can usually get it after banging around the bid/asks a few different ways, but not directly. Think I understand the triangle approach, but getting hung up on what currency pair to begin with. Thks for help on this, you guys have been great.

I also do that using a few scenario… Less memorization is better for me :stuck_out_tongue:

you look to see which currency is over or undervalued in the market, then you position yourself to sell the overvalue currency and buy the undervalue one. buy low, sell high

that is the part i’ve been struggling with. given set of exchange rates with common 3rd currency what’s the best way to identify if they’re under/overvalued?

if you have /EUR and Yen/ and are asked to compute Yen/EUR then all you do is multiply the 2. (the dollars cancel and your left with YEN/EUR) If your given bid-ask spreads with /Yen instead of Yen/, you have to take the reciprocal of both bid and ask (1//Yen) In this case you have to adjust the bid-ask spread and give the bid the lower of the two reciprocals (since thats what the market makers will be buying it for) and the give ask the higher of the two reciprocals (thats what the market maker will be selling it for.) after that just multiply the currencies to ahve the dollars canceled. I found the following easier for finding the profit of a 3 party arbitrage: if you have 3 currencies ,Euro and Yen then you will be buying into 2 of the 3 since you will be starting with one of the currencies. For example if the question asks what is the profit in dollars, your options are to buy into the euro or yen. without finding which one is overprices/undervalued start converting. first scenario) You have dollars (use a value of 1) you buy into the EURo, use the EURo to buy into Yen then use the yen to buy into the dollars if the 1 dollar ended up more than 1 then you make the difference (for example 1.002-1= .002) but thats only for 1 dollar so you have to multiply the difference by the principle given in the hypothetical question. However if you $1 is less than 1 then there is no arbitrage move on to the next scenario 2) You have dollars (use a value of 1 again) you buy into the yen first this time, use the yen to buy into the Euro and then use the Euro to buy into the dollars. if your dollar is more than 1 then you make the difference. Usually if scenario 1 is greater than 1 you dont move to the next scenario.

argh

http://www.analystforum.com/phorums/read.php?12,1258317,1258317#msg-1258317 Try this. I have this way of thinking about it and I have never got it wrong since.

Usually they are 3 curencies, so in the worst case, which is what I do instead of trying to figure out who is overvalued and undervalued, there will be two trials only. If /Euro, /Pound, and Euro/Pound are the rates, then start with $1 million. With $1 million, you can only exchange it into either Euro or or Pound, so try the Euro. If you do that you have no other choice other than to convert the Euro to Pound then back to dollars. If you end up with more than $1 million, you have arbitrage. If that doesn’t work, you only have one more try, which is to convert the $1 million to Pound, which then forces you to go to Euros and back to dollars. If you end up with more than $1 million, you have arbitrage. If not, stop. I don’t think you need more than two trials, do you agree?