I can't do it here.
To be clear I am just saying it exists, not that it could be profitable or you will get good historical data, or whatever.
Even if it is not the historical data for that particular broker, using a data from a different broker, will allow you to back test your EA and see if it is profitable. So, just use the tick data from a very famous source and test your EA with that.
As I said, the tick data differs only by small deltas between brokers, so even for M1 data the differences are subtle and should not impact your EA too much and is the equivalent of the impact of small changes in spread.
If your EA however is so sensitive to those difference that it suffers greatly from that impact, then rest assured that in a real live forward trading it will do just as badly if not worse on the real data feed.
From what I know, it' just not possible as they have opposite interest to yours, you win they lost. How can someone really hope to earn something in the long run ?
Maybe there are some true brokers as mentioned, then it could be an other thing, to verify.
I don't know about the exact number, however, I can place a Link to a site giving the calculations as to why the broker has a high edge in its favour?
Am I allowed to place the link here?
PS! Actually, I don't have to place the link here. A quick Google search and you will find quite a few sites explaining why there is such a high edge in favour of the "house" for Binary Options.