Demo Vs. Historical Backtesting

Hi guys, I was pondering earlier today about how I am going to find time to demo trade my methods for the 3 Months James16 recommends.
I have a computer without internet access that I use to do some manual backtesting on historical data. Usually I plug in a blank EA, let it run and stop it whenever I see potential setups. Next I draw my lines and enter trades into a spreadsheet, let the EA run and see how it plays out recording wins/losses ect.
I have a theory that you guys can probably debunk rather quickly. Because the markets are more or less random, testing price action on historical data like this should equate to testing on real-time data, right?

My plan is to trade this way for the 3 (real) months and if I'm profitable at the end then take what I have learned Live with a 1/2 size acct.

What do you guys think, is "demo" testing on past data useless compared to real-time?

I think that is a perfect way to start. That is the method I used to become profitable. I would drag a chart one candle at a time and use a spreadsheet to log trades and monitor my performance.

This is a very good way to start. Demo trading comes after you have done this for a good, long time, and you have mastered your setup.

From my opinion the Back Testing and Demo is something not similar.
Demo allows you feel the reality and understand how everything works.

Back testing - allows you test your ideas and see how much you would have gained if you traded there.

I suggest you first use back testing, than demo and only after live trading.