Backtesting and Demo

Like most of you, when I was a complete noob, I did a lot of demo and backtesting. I'd sit there and see all those imaginary green and red dollars, and the systems that had 90% winrates. Yet, I never made money from either. I started to stray from "playing the cards" to "playing the players". This ended up making me realize that players don't bet their entire bet in one entry, but they will more often scale in/out due to their position sizes. When they enter the market, they can move price.

This is very important to note, as demos and backtesting do not account for actual trades taking place. Buying a gazillion lots of any pair on a demo account will not show what really happens in the real world. While price may move for big accounts, we should also note that even retail traders contribute to inventory balances of their brokers. The brokers will try to match orders in house, but they will always be exposed as a broker. While a minilot may not move price, it will ultimately alter the course of price simply due to the simple fact that the fact exists. Things may not change drastically, but how can you test something which doesn't take into account your actions?

I feel like I've overlooked something, but I don't know what that could be. Anyone have any thoughts?

Your thinking is definately on the right track. Large traders must have liquidity to enter positions. If a large company needs to convert a large number for something mundane like payroll they still have to have a strategy to make that trade because the slip that would result from fifty or sixty million hitting with a market order is unacceptable. Often when stops get triggered it is just a skilled bank trader getting a good fill for his customer.

I recommend that people only test and learn to trade with a very small account. I think it helps to learn very early what really happens when you try to enter when the market is moving fast in your direction. I have seen even a single minilot slip a few pips with a market order. It makes you start to be a little strategic with your entries. I generally only enter short in a rising market and long in a falling one because I want the slip to be in my favor.

You are also correct that your fifty thousand sell could just be the last shovel full that fills the broker's bucket and sends him to the interbank market to offload his risk.

http://fxevolve.com/understanding-li...-of-execution/

I know there are supposed to be "trading simulators" or at least I've read about them. I wonder if they are any different. However, I wonder how people are trained as traders. I have very recently started looking into entry level jobs for firms and brokers. (I doubt I'll get a job, but that would be wonderful.)

I also agree with the very small account. If one joins a firm or bank to be a trader, in the beginning of their career, their managers usually don't give them full buying power. The saying is the same in poker and trading, there's tuition you gotta pay. In poker, mostly online, they have a lot of play money rooms alongside the real money rooms. The only thing this is good for is getting used to how the game is played, where the button goes each hand, how each card is revealed. You can't develop strategies because no one can take that game serious, unless you are playing with a group that has it in mind to take the game serious. People constantly just go all in every hand, or they are just playing bingo. Even in the micro stakes, it's hard to actually have any strategy, yet there is just a little. Once you start getting up into the bigger buyins where you are putting down a few hundred to a few thousand, then you can really start getting into strategy.

While the subject of poker is in the air, for online poker, there's poker trackers which track the hands you play. You can find out how often you put in money voluntarily and how aggressive you are and on which streets and how often a person will fold to a 4bet on the river after opening up preflop and etc... It would be interesting if there was a trader tracker, lol.

Thanks for the link, btw. You provide a lot of good information.