I recently tried an experiment to document the idea of buying low on BTC-E and selling high on Mt Gox Arb opportunity.

I'm from the US.

Price on BTC-E was about 20% below Gox price when I initiated the wire.

On Monday I wired $1000 USD to OK PAY  -                          COST $30.00    Total: 30.00
On Friday OK Pay Received my wires                                     COST 12.00      Total:42.00

OKpay deposits disappeared from btc-e for about a week

The following week I sent 987.00 From OK PAY to BTC-E       COST 2.99       Total: 44.99
BTC-E charged Me 2% to deposit                                            COST 19.74     Total: 64.73


So now I have 967.26 in BTC-E (which is effectively 934.27 when you account for the 30 wire fee and 2.99 ok pay transfer fee)

I buy bitcoin at 127.878 for a 2% fee                                      Cost  19.35      Total:  84.08

Now I transfer these coins to Mt Gox.                                      Cost 1.25         Total: 85.32
Price at the time of transfer 130.03695
Waiting for 6 confirms for the coin to show up in gox.            so far the effective premium is 8.53%

An hour later I was able to sell the 8.8289
btc on Mt gox  for an avg price of 129.755                                    Cost  $6.87       Total 92.19

So I currently have  $1,138.83  cash in gox.  less the 30 wire transfer and the 2.99 ok pay fee =  $1,105.84

Time 2 weeks
profit 105.84.  or  10.58%
cost:  92.19    or   09.22%

(fyi this is not scientific and numbers are only back of the envelope calculations)  To me is seem then unless you can hold large amounts of cash and/or coin on the various exchanges it it a bit more trouble then it is work. - Then what do you risk if one of these exchanges gets shut down/hacked or disappears?

Be curious to hear anyone else's experience.