They're not the same because there's no single central exchange for FX - just lots of different banks all quoting their own price - different price feeds will each use one or more sources for their prices, so each feed will be different from the next one. In order to backtest effectively you need to be able to test using your brokers prices (because each broker will also quote slightly different prices from the next broker).
There are 2 ways that I know of to do this (although there may well be plenty more):
fxengines -
http://www.fxengines.com/ - i'm not sure about this one but I think they use FXCMs prices & you can then automate your strategy via FXCM
TradeStation -
http://www.tradestation.com/default_2.shtm - they use RJObrien as a price feed & as a broker
In each case, because you are backtesting with your brokers own prices, you should be able to rely (as far as is possible) on your backktest results. I haven't tested either of them for this purpose so cannot vouch for them.
If you aren't able to test with your brokers own prices then testing a short term forex strategy is entirely futile in my experience. (I thought I had a fantastic fx system a while ago, until live testing showed that the prices my strategy got were often nowhere near what my broker was quoting - when I tested it with TradeStations 'tradeable' prices, the backtest results fell a long way short of the results I got when testing with the eSignal prices that I was using previously)
Hope this makes sense/helps
Simon
