WOW,...I must comment here. I will give you a set of rules you MUST follow. These are rules that you are OBVIOUSLY NOT following. You will blow up your acct again probably before the end of the year if not before Halloween if you do not follow these rules.
1) Never risk more than 0.5% of your capital on any one trade UNTIL you have proven through statistics that your "strategy" will in fact make you money. Realize that your backtesting has inherent errors in it. It assumes you will trade your system 100% perfectly which obviously you are not doing. Because of human errors, emotional errors as well as market dynamic your system will NEVER perform as well in real life trading.
2) Keep detailed records of your trades. I recommend using a good spreadsheet designed for this. I have no affiliation with this site but I use their product and it is GREAT for keeping track of trades. BUY this...
http://tradingspreadsheets.com/default.aspx
3) Read Dr. Van Tharp's book "Supertrader". I was like you in many ways because of a "good strategy based on backtesting". I started out thinking I knew how to trade and kept blowing up my acct and blaming it on the market or not following rules and ended up with highly volatile swings but net loss. This book changed all that for me. This was the basis for understanding risk management and position sizing. You SAY you understand it but it is obvious you do not. This book will help you realize your weaknesses and improve on them. Once you read that I also recommend reading his book "The Definitive Guide to Position Sizing". In that book you will learn to evaluate your system in depth and determine more appropriate position sizing. Then you can optimize your % of capital to risk on each trade. The difference is you are doing it based on stats. IF your system truly shows a positive expectancy in real world trading NOT backtesting then you should be able to risk more than 0.5% of capital per trade. You will need at least 100 real world trades to do this. This is where your spreadsheet data will help you fine tune.
If you do not follow these rules I guarantee you will blow up your acct again. I can not explain in detail why on here as it would take too long. Suffice it to say the reasons are in the sources I have given above. Basically your position sizing is WAAAY to high and you are going to suffer huge drawdowns.
A big part of your problem appears to also be psychological. If you TRULY had enough confidence in your system you would not be taking these other types of trades that do not follow your rules. You are doing them because you still are not sure and you have not proven it to yourself that is does work. Statistics on your data will show you what works and what doesn't....again the spreadsheet is vital here. You will also notice that keeping stats and examining them periodically will force you to do more of that which works and less of that which doesnt. Also, you do not seem to have appropriate stop losses. There should be NO way you could lose 40% on a couple of trades if you have 2% supposedly at risk. I suspect you do not enter stops or you average down on losing trades. Both are bad!!
I also suspect you are quite young (< 30 years old) and do not have enough capital to trade well. You really need at least $25K in my opinion to trade effectively. Someone with frequent trades has high round trip commisions that you must overcome in order to be profitable. Add up all the round trip commissions over the course of 1 month of trading and figure what it would be after 1 year. I suspect with your acct size it is AT LEAST 35% in commisions alone and possibly more like 100% depending on how many trades you do in a month!!. An acct size of >25K will at least give you a fighting chance. In the mean time you need to raise capital, READ, and preferably paper trade and keep stats on that. If you must use your own money follow my rule of 0.5% until you have stats. You quite possibly will still blow up your acct with 0.5% because your commisions will eat away at your acct but this trading time will only be for collecting data. You should analyze the data WITHOUT commisions to get an idea of what could happen with an acct big enough to make them less of an issue.
IF you have read this far into my post then at least thats a step in the right direction. I encourage you to print this out and keep it. You may or may not listen to me now. If you dont listen to me now then SOME DAY after blowing up another 10 accounts you will possibly be at a point in your trading life where you CAN accept some of what I am saying. What I am suggesting is looking at yourself and treating this more like a business than a slot machine that you think you can rig in your favor.
Good luck.