'm testing my automatic trading system in stock market (data mining system). I'm modeling day by day for 30-days and calculate profit in every step. Suppose that my system predicts tomorrow close price is higher and today close price. When I check the real direction of tomorrow price it is higher than today. So we have same direction in real and predicted close price. Now I want calculate profit. Can I use this formula?
profit_for_this_trade = ( next_day_close_price_sell - today_close_price_buy) - commissions and other buy and sell related costs What do you think?
(We used today and previous days closing prices of today to train a model and predict next day closing price) What is your proposed formula for calculating this profit for a close price predicting system? Currently I'm using only accuracy of system for performance calculation.
Dear Rahimikia
Your formula faces some problem
1. the prices in your formula are the closing prices which will not be your actual prices if you actually trade a stock
2. In a trading system, you may not be able to sell a stock immediately after a day (your stocks may be available for after 2 or 3 days)
Best
Canh
@Nguyen Phuc Canh. Thank you for your answer. We're currently in testing stage not practical stage so we want calculate approximate profitability.
1. That's true but I think it's the best price that we can use in this situation.
2. This trading system is trading in an online system so currently we don't have this lag between buy and sell orders.
Dear Rahimikia
I see, so i suggest that you use the average price for measuring the returns, that will be more reality for evaluating and then we can implement into practice
Best
Canh
@Nguyen Phuc Canh. Currently I have daily open, high, low and close price (and last price). I should average between open, high, low and close? As you know there is difference between calculating close price comparing to open, high and low price. Is is appropriate to add close price to calculate average of prices?
Dear Rahimikia,
In your set-up period, I would say to not care about what kind of price you are using, but to concentrate on what has potential big effects. At least I can see to major points:
- You have to consider dividends and corporate actions such as spits or mergers (and all others having an effect on the price).
- You also have to make your mind about using a P&L method. What stocks are you selling first : the last you have bought or the first ones ? (FIFO, LIFO, Average price…).
The mechanisms to manage properly both aspects might be tricky
Finally, using different kind of prices refers to different usages. For example, Portfolio accountants are using closing prices, whereas for an immediate P&L you should use last prices. In any cases, you should use bid prices when you want to evaluate the price you will get for a security you own, and the ask one for a security you want to buy. This is not the commission you referred to.
Not to mentionne that I am very dubious about predicting a closing price (or any other) ;-))
For a realistic assessment, please take into account trading costs. These consist essentially of two components:
(i) on the one side, the known fees you have to pay to your stock broker to execute the trade for you on the stock exchange (if you want to get direct market access to the stock exchange computer you need to become a member institution with lots more costs involved).
(ii) and on the other side the "hidden trading costs" resulting from the fact, that your own order to be executed at the closing auction will have an impact on the closing price. This market impact of yours is much more difficult to estimate because you only know from historical data what the closing prices are without you being active.
(iii) don't underestimate smart market players who identify you and your activities (=trading strategies) and react a bit quicker. Some people might call this "front running"...
Good luck!
- Badges
- Science topic
- Similar topics
- Computer Science
- Data Mining