We should all thank our new poster and friend @wmitzlaff -- for starting what has turned out to be both an exceptionally interesting thread, as well as quite educational. Thanks @wmitzlaff !
I worked with a trading simulator before. Was doing so good and when I went live things completely turned on me. I think one of the issues about simulator is allowed me to fast foward or play at much higher speeds which gave me more patience and bravery then I actually possessed compared to trading on a real account.
My approach was not to backtest a "system" per se, but to forward test in real-time. This allowed me to most-closely approximate real-time market movement and the signals I would take at any given point while in sim. The method has been excruciating in the time needed, but at least, for me, it has not allowed me to make assumptions about my own methods if I tried to employ them through an artificially-speeded-up timeline, and previous market conditions which may cease in the near future when going live. I think "playing the data at high-speed" is great for testing an automated system with signal entries and ATM strategies...But it does nothing for the discretionary trader.
I disagree. If you cannot make money with your system on historical data, why would you ever trust it going forward? Moreover, backtesting allows you to discover the effect of stops. Too close, too wide, etc. One of the most valuable things you can learn is the max adverse excursion concept, ie when your losers move a certain amount against you, they are unlikely to come back. It's helpful to have an idea what that amount is.
Sometimes it's nigh on impossible to backtest a method. Well you possibly could backtest my system but it would require lots of inputs, different data suppliers, different software, very expensive and time consuming to set up. When you have a systematic system with discretionary elements looking at a number of other related instruments. For a one man band such as ourselves, forward testing with real money is simplest, just a bit risky, however risk can be contained. The method I use was never backtested because it was I thought with my budget impossible. I eyeballed it only and tried it, jumping in the deep end fixes all sorts of problems, but primarily analysis paralysis.
Counter-question...If you can make money with your system on historical data, why should you trust it going forward? It is like many folks say...Your system works until it doesn't. Conversely, your system doesn't work until it does. That may seem a pedantic comment, but there is truth to it. I guess my testing was to truly discover my own self, and not the system as it were.
Is it ok to say that I too trade without Stops... feels like coming out for the first time! I was originally taught to trade with the conventional thinking around use of stops. My win rate was well below where I thought it should be given my skills as an analyst and directional views on stocks. I played with wide and narrow stops and all sorts of rules around them. Eventually after analysing some 200 trades from my trading log I realised a big secret. That stops have a negative edge. I came to the conclusion that by not using stops you could blow up your account, but using them just means you lose more slowly. I see a traders journey to profitability as a series of traders dilemma's just like this. The answer for me has been not to use Stop Losses, as that has a negative edge, but to design my trading approach to manage risk in other ways. So for example I won't trade futures as too risky without stops, or similarly cfd's which is a widely used retail trading instrument here in Australia (banned in US I believe). Although not the only approach, I mostly take directional positions on equities using options, where in this case I have a defined risk, but the underlying stocks can spike up or down where ever it wants within the timeframe I'm trading without getting shaken out of my position. I'll also add that Stops can give a false sense of security. When you really need them, say the next market crash, a lot of traders are going to find they don't quite work as well as expected.
Stops can be mental, so trading without stops is like saying trading without loss-exits (aka "admitting to being wrong"). This doesn't have to be incorrect though, if you diversify enough and expectations can exceed +100% on other positions! I had an auto-system trading using fixed stops for a year, but needed to shut it down (guess why) and am trailing the rest of positions "manually" using fixed stops and discretion. Fixed stops or not, the system had expectations of following X00% moves in the market, so was so wide they didn't need to be fixed at all, thus vastly simplifying my next system. Turned out to be more complications and work to handle fixed stop-lists at broker-end than just handling orders directly. But if you at any point in time close a lossy position due to poor performance, I'd say you have some kind of Stop. Never understood targets. If the PA is good, I'd want to hold on to it.
It's really a good question because is so easy to say "yes stop loss, you are crazy without it". But it really depends on your trading style and strategy because if you have the luck to know how to test your strategies you will see that often in a trend follower strategy stop loss worsen your performance. And also have some targets (is better to have trailing profits for example). Personally I always put a daily stop loss in my discretionary trading, especially if I'm scalping (most of the time i'm contrarian) and the ultimate secret of a professional trader is: how to manage your risk with SIZE...