Thanks Low Salubrity Thug, How do I code and test? For my case, everyday I draw the support and resistance I want to trade against. Then I have a mechanical system I manually use to trade it. I am ok at coding and perform coding while in college. I use ThinkorSwim platform. Are the example coding I can build on? Thanks,
Java, I agree 100% with you. I have learned my lesson about not having a plan and the lesson cost too much. Therefore, I maintain a plan before live trading. This helps removes emotions.
Hold on. So you're watching the screen pretending to trade and writing down how much money that makes ? What if your strategy only works for 6 months ? The only way you can know if your strategy is good or not is to push a button and see what it did over the last 20 years. Try NinjaTrader, TradeStation, MultiCharts or even better some Python / C# reading a .csv of OHLC data.
Hmmmm.... I wouldn't say that. First you come up with the trading logic, e.g. when to buy and when to sell, following what behavior on what data. Once that your system is robust, not over-fitted, realistic and has a good correlation to the rest of your portfolio of systems (as in no correlation), it's only at this point that you know how much money it may start with. Then, depending on where you work, either (a) a completely different team will take your trade list and come up with execution algorithm; or (b) you will have to do it yourself or rely on external brokers systems. At our firm, it's a bit desk-based so each desk do whatever it wants. I personally have 3-4 good execution algorithms that I test and optimize on a list of simulated trades. Execution should always be segregated from trading logic
You need something which simulates your execution venue (e.g. exchange). Professional algo trading outfits spend a lot of time and money building simulators which model exchange behaviour down to the latencies involved in processing messages and the exact matching engine algos used for each product. It's the only way you can have any clear idea how your strategies are likely to perform in the real world. Designing and building these is not a trivial task. Maybe there's a commercial software package available which does a decent job. I'm afraid I don't know as I'm not familiar with the technology available to retail traders.
Sure. I suppose I was just nitpicking really. To mind whenever you work size in the market over a period of time you are scaling in/out. You at one of the old school Mayfair funds or somewhere more progressive? Agree. Execution is a discipline in and of itself. It can make a significant difference to P&L if done skillfully.
1. Fixed stop would be needed for this then. Pointless exercise in exits...This would require entry work, not exit work. 2, 3 and 4 are now rendered impotent, because you have lost sight of the goal. With a trailing stop, there should not be, in my mind, a pre-determined target. That's the point of a trailing stop. Your trailing stop is your target.
Bracket trading is futile. The sooner you get out of that box the better. What really separates the pros from the independents is the pros can't risk 100% and the independent must. Best way to go hungry is look for a free lunch. Identify what has to go wrong to wipe you out and start from there. Looking for a way to play this game without getting wiped out is waste of time, energy and especially money.
All in, scale in, stop, trail are all tools. When used to embrace risk each one can have it's own special purpose. It really doesn't matter how much you test a hammer if you don't know what you are trying to build. Testing a hammer to see if it will build a house is meaningless place to start. First understand what will destroy the house you want to build.