How Much Does Program Trading Cost? A Full Breakdown
Program trading costs fall into five groups: software, data, computer, trading capital and per-trade costs. The first three can start with free or existing resources; the real barriers are trading capital (margin plus backtest max drawdown plus a buffer) and the commission and slippage paid on every trade.
By Soren Huang · Published · updated · Editorial policy
Contents
- How much does program trading cost? Start by splitting the money into five groups
- Program trading software costs: is free backtesting software enough?
- How much do historical data and a computer (or VPS) cost?
- How much minimum capital does program trading need? How is margin worked out for Mini-TAIEX and micro futures?
- Commission and slippage: ongoing costs beyond fixed monthly fees
- Frequently asked questions
- References
How much does program trading cost? Start by splitting the money into five groups
How much does program trading cost? It is clear once you split the spending into five groups: software and platform, historical data, computer or VPS, futures margin and trading capital, and the commission and slippage on each trade. The first three can start with free or existing resources in Taiwan, close to zero. What really sets the bar is trading capital, plus the costs you pay on every trade, which grow the more you trade.
It is like moving house: boxes and tape are cheap, while rent and the deposit are the big items. Many people focus on whether the software costs money, before working out how much to put in the account and how much each trade takes away. Slippage is the gap between the price you expected to be filled at and the price you actually got, and like commission it is paid on every trade. The figure below separates the five groups by how they are paid, then shows which can start at zero.
If you have not started yet, Program trading for TAIEX futures: zero to first strategy explains the steps without coding; this article now prices each group in turn.
Program trading software costs: is free backtesting software enough?
A free backtesting tool is enough to test your first idea. Paid software does not buy "more accuracy"; it buys a few extra checks. Most free tools can apply a rule to historical prices and draw an equity curve. The differences are whether commission and slippage can be set per trade, whether trades fill on the next bar, whether you can test many variants at once with the same settings, and whether you can run walk-forward analysis and export to a trading platform. A next-bar fill means the signal is generated when a bar closes and the trade fills at the next bar's price, so you never fill at a price you could not yet see.
The suggested order: use a free tool to test one rule down to a net result after commission and slippage, make sure you can read the report, then decide whether to pay for large-scale testing and validation. Free backtesting software for futures: what you can test covers what free tools can do, and First algo trading strategy: a 30-day route for beginners gives the weekly tasks for that stretch.
How much do historical data and a computer (or VPS) cost?
Daily bars and longer-timeframe minute bars can usually be had cheaply or free, you can start on the computer you already have, and a VPS becomes a fixed monthly fee only when the strategy needs orders running 24 hours. A VPS is a remote computer in a data center that stays on all day; you put your order program on it so your own computer need not stay on all night.
What you pay for in data is completeness, not quantity. TAIEX futures have a night session, and US index futures trade almost around the clock, so data covering only the day session makes a backtest miss half of the bars and gaps. How nearly 24-hour futures trading affects backtests explains how session settings change results, and After-hours trading explained: hours, liquidity, backtest pitfalls explains why night-session data needs a separate volume check. Before you get data, ask three questions:
- Does the data include the night session, and in which time zone is the bar close time?
- How does the continuous contract data handle contract month changes, and does it mark the roll dates?
- How many years back does data other than daily go? This determines price and cost.
On the computer side, backtest speed depends on the processor or graphics card, but a daily-bar strategy runs on an ordinary laptop. Hardware is worth upgrading only when you test a large number of variants at once.
How much minimum capital does program trading need? How is margin worked out for Mini-TAIEX and micro futures?
Minimum capital is not the single margin figure the exchange announces. It is margin plus the backtest max drawdown, plus a buffer. Margin is the money the exchange requires you to keep in the account to back your position; max drawdown is the largest fall of the equity curve from a high to a low during the backtest. First use the free TAIEX futures margin calculator to look up the current margin per contract for TAIEX futures (TX), Mini-TAIEX (MXF) and micro futures, then add the rest on top.
The formula is just one line: minimum account capital = exchange margin + backtest max drawdown + buffer. The buffer is extra money set aside because real drawdowns are usually deeper than backtested ones; a common practice is half of the max drawdown, or an equal amount. Mini-TAIEX (MXF) is NT$50 per point, and micro futures shrink the contract size further, so margin and drawdown amounts fall in proportion and all three blocks get smaller. Margin changes with exchange announcements; for actual numbers, rely on the TAIFEX contract specification page.
Hypothetical example
Suppose a Mini-TAIEX (MXF) strategy has margin of NT$50,000 per contract, a backtest max drawdown of NT$80,000, and a buffer of half the drawdown, NT$40,000. Minimum account capital is then about NT$170,000. Switch to a micro contract one-fifth the size and all three blocks shrink to one-fifth, so minimum account capital is about NT$34,000. The numbers are only illustrative; plug in your own backtest drawdown and the margin the exchange announces that day.
Commission and slippage: ongoing costs beyond fixed monthly fees
Commission and slippage are paid on every trade, and the more you trade, the more you pay. The way to judge them is to compare each trade's cost with the average trade. Below is an actual backtest from AlphaPrime (Valynox): ES daily bars, January 2, 2019 to June 30, 2025, long and short, 1 contract each time, with a US$20 commission and US$25 slippage charged on every trade, US$45 in total per trade. This search used an ATR channel rule as the only entry and exit condition, with no stop loss, profit target or filter. It is a stand-in rule for the topic of costs, not the subject itself. The report is the one the search kept from 1,024 candidates, ranked by return relative to drawdown. Picking the best one after the fact makes the numbers look better by itself, so these numbers do not mean the rule earned this money, and they are not a result you should expect.
Source: AlphaPrime product engine (CPU), run 2026-10-31. ES daily bars, 2019-01-02 to 2025-06-30, long and short both traded, 1 contract, $20 commission and $25 slippage per trade.
| Number | Result |
|---|---|
| Net profit (closed trades) | $78,397.5 |
| Max drawdown | $56,087.5 |
| Trades | 32 |
| Win rate | 50% |
| Profit factor | 1.881 |
| Average trade | $2,449.92 |
| Time in market | 95.68% |
What matters in this backtest is the ratio. A cost of US$45 per trade against an average trade of US$2,449.92 is under 2%, and over more than six years there were only 32 trades, so total costs came to a little over US$1400. Put the same US$45 on a strategy that trades once a day and it is charged more than 200 times a year; if the average trade is only a few tens of dollars, costs eat most of it. That is why a backtest should include costs from the very first run, and Trading costs in backtests: commission, slippage and the average trade has the full calculation. For actual commission, futures transaction tax and slippage figures, rely on your futures broker's fee schedule and the exchange's announcements. When you read your own results, the trade list and average trade in the AlphaPrime performance report are the two numbers you need to work out this ratio.
Next step: open the report of your latest backtest and divide the average trade by commission plus slippage per trade. Then check today's margin on the TAIFEX contract specification page and use the formula in the previous section to work out your minimum account capital.
Frequently asked questions
How much minimum capital does program trading need?
Not just the exchange margin. Margin plus the backtest max drawdown plus a buffer is the minimum account capital that lets the strategy ride through a deep drawdown. Use your own backtest figures for the drawdown and the TAIFEX announcement of the day for margin; both change over time.
How much does Mini-TAIEX program trading cost?
Mini-TAIEX (MXF) is NT$50 per point, so the same strategy's drawdown amount on Mini-TAIEX is one quarter of that on TAIEX futures (TX), and the margin is much lower too. Mini-TAIEX margin plus the Mini-TAIEX backtest max drawdown plus the buffer is your starting amount. If that is still too high, you can use the smaller micro futures to shrink all three blocks further.
What fixed monthly costs does program trading have?
Only three are common: the monthly fee for paid software or a platform, the subscription for real-time or historical data, and the monthly rent for a VPS that holds the order program. All three can be zero at the start: use free tools, free or one-time-purchase daily data, and your own computer. Commission and slippage are not monthly fees; they are charged on every trade.
Do I have to rent a VPS for program trading?
Only if the strategy needs orders working overnight and you do not want your own computer on all day. A daily-bar strategy, or one that trades only the day session, can run backtests and place orders on an ordinary laptop. First confirm which sessions the strategy trades in, then decide whether a monthly fee is worth adding.
References
- Taiwan Futures Exchange. Mini-TAIEX Futures. https://www.taifex.com.tw/enl/eng2/mTX
- Bailey, D. H., Borwein, J., López de Prado, M., & Zhu, Q. J. (2014). Pseudo-Mathematics and Financial Charlatanism: The Effects of Backtest Overfitting on Out-of-Sample Performance. Notices of the American Mathematical Society, 61(5), 458–471.