As a quantitative trader who has been losing and losing for seven years, I've been meaning to write a post summarizing some of the losses I've suffered and some of the experience I've accumulated over the years, in the hope that it will give you a little inspiration and that you will be able to walk a little more smoothly on the path of quantitative trading.
# The starting point of everything
Back to the starting point of quantitative trading, it is certainly alternative to manual trading, that is, you have to have a logical closed-loop trading strategy before you can further research, at first you can go on trying some EA publicly available online, but eventually you will return to your own implementation of the strategy The road of course does not mean you have to write the code yourself, but you have to be able to quantitative description of your trading strategy, write the code or something, you can find a fighter, such as the industry's gold medal fighter, group master 😊. (Group masters are also genuinely expensive in terms of fees)
# Theoretical foundations
To make money in this market, you first need to figure out what you did right when you were lucky and why you made money.
Many people's trading style is to pursue short and quick, do intraday ultra-short, but in fact, such transactions cost very high, we must first be able to beat the transaction costs , survive down, before it is possible to profit, especially on the transaction costs must be understood clearly to understand the cost of the market. In particular, transaction costs must be well understood.
- Spread cost (the difference between the bid price Ask and the ask price Bid is called spread)
- Slippage cost (the difference between the price of your order and the actual price of the transaction is called slippage, the vast majority of platforms will be slippage, and it is not in your favor slippage)
- Swap interest (different varieties of overnight interest) (the overnight interest rate varies from one instrument to another and can change)
- Price cost (topping and bottoming is about getting a lower price cost)
- Margin cost (the margin tied up in a position is also a cost)
- Time cost (the time spent holding a position is a mild horizontal line cost)
- Risk cost (when the market price moves in an unfavorable direction, the risk cost increases)
Once we understand these transaction costs, our primary goal is to beat these transaction costs in our trading strategy, estimate them consistently, and come up with out/in and holding conditions that are favorable to our profitability .
# EA (Mt4/Mt5 Trading Expert)
First of all, to be clear, you must first learn to use EA, including loading EA and EA backtesting, the initial stage can download some free EA from the Internet to learn, or better yet, if you can read the code, download some open source EA to understand how it determines the entry and exit conditions and place orders, especially the quantification of indicators How the results are used.
For EA backtesting, Mt5 is recommended, after all, Mt5 naturally supports tick backtesting and simultaneous backtesting of multiple currency pairs (each quote sent by the platform server is called a tick), it should be noted that do not believe in EA backtesting, especially the backtesting results of EA's that do not see the source code, for several reasons:
- Because of the existence of cheating methods such as time functions, these EA's can backtest within a specified time period. These EA's can get exponential backtest net worth growth in a specified time period
- Even tick-level EA backtesting is massively different from the actual trading environment because backtesting doesn't slip and doesn't Network disconnects 🤣
- Markets change, the longer you test, the greater the range of change, an EA that makes money this month may not work next month, EA backtesting can only ever backtest past data, it can't test the future market
Once you understand this, you can open a demo account and load EA for simulation testing, you need to observe whether the EA can open and close positions properly, set stop-loss and take-profit normally, etc. It is important to note here that there are also major differences between demo and real! The company's main goal is to provide a strong and reliable service to its customers.
# What does a mature quantitative trading system look like
Before actually building your own quantitative trading system, it is necessary to understand what a mature quantitative trading system looks like, through the group owner's observation and practice over the years, it mainly has the following characteristics:
- Clear exit/entry signals, if necessary you need to work out a set of custom indicators yourself, it would be better if you could derive exit/entry signals without indicators, after all, there is a lag in indicators
- Clear hard and soft stops, hard stops are used to resist huge risks in case of black swans, soft stops are used to clarify the stop loss conditions of the trading strategy
- Define the maxDD (net maximum float loss) that is acceptable for the trading strategy and build the overall risk control system for the account based on the maxDD
- A robust risk control system where the benefit risk control and stop loss risk control can be unequal, but a certain percentage must exist
- No logic holes, which are so terrible that they can open too many orders at the same time, or close out medium and long term positions all closed out in one fell swoop (the group owner has suffered losses on this, the whole thing was confused when it happened)
- Be able to outperform the last few years of historical market data (at least one year) and get a growing net worth curve
- Be clear on the varieties used by the trading strategy, a strategy that works for all varieties does not exist, behind each currency variety there is a There is a country behind each currency pair, and each country follows a different monetary policy, so their actual performance is very different, and there is no one-trick pony
- Define the number of lots and the total number of orders opened under margin, this is very important, some EA's can run exponential growth results in backtesting because it is backtesting to hold the trend, to make an unlimited number of intensive open orders, the end result chart It is nice to see, but the real world can kill people, so it is important to specify the number of lots and the total number of orders, for example, 10,000 $ margin, 0.01 lots per single, cumulative position of 0.1 lots, that is, the total number of orders 10
- Not using strategies that are mathematically known to be bound to fail, such as Martin and grid trading, both of which will inevitably lose due to risk and transaction costs
- Being able to survive in the live market for more than a year with strategies that initially preserve at least capital
# The Dawn of Victory
Well, by this point maybe you have your own understanding of quantitative trading strategies and are dumb enough to move, the group still has a few words of advice that need to be said to conclude this post:
- Don't believe in the Holy Grail (the pursuit of a one-and-done strategy), the market is changing, and quantitative trading systems need to follow the market to make corrections, usually on a half-year or yearly basis
- In the process of practicing quantitative trading, paying tuition to the market is inevitable. strong>is inevitable, do not be frustrated, when you encounter difficulties actively review, timely replenishment of ammunition
- The most difficult part is to preserve capital, and then is to win, so to adhere to the path of quantitative trading, we must firmly believe in their own capital preservation and reflected in the trading strategy, first to live, to have survival, before there is a future.