What this kit is
This is the page to read before you spend anything on algo trading. It pulls the whole beginner path into one place: the four layers your setup needs, one ready-made stack with honest pricing, five checks to run before any live trade, and two free calculators that do the maths for you. No signup, no purchase required for any of it.
Stackbot's house rule: if a tool's price isn't published, we say so instead of guessing — and the same goes for this kit. Every price below is taken from the directory listing it links to.
The four layers
Every working algo setup has four jobs, whether it costs $28 a month or $2,800:
1. Strategy — something that decides when to buy and sell. A ready-made EA, a hosted bot platform, or a strategy you build yourself. 2. Hosting — an always-on server (a VPS) where the terminal actually lives, so a sleeping laptop or a dropped connection can't strand a trade. 3. Testing — backtesting on historical data before real money is involved. Practise on the past, then risk money on the future. 4. Scorekeeping — a journal and tracker that records what your strategy actually did, as opposed to what you remember it doing.
Skip a layer and you will feel it eventually — usually at the worst moment. The full walkthrough of each layer is in Your first algo trading stack.
Your first stack: the MT4 Starter Stack
One complete, lean first setup from the Stacks page:
| Job | Tool | Cost |
|---|---|---|
| Run the strategy | Forex Fury | $249.99 one-time licence |
| Prototype ideas | fxDreema | Free basic tier |
| Keep MetaTrader online 24/7 | ForexVPS.net | $28.80/mo billed annually |
| Track and verify performance | Myfxbook | Free tier |
Known monthly total: $28.80/mo (the one-time licence is a separate, upfront cost — it is not folded into the monthly figure, because hiding it would be dishonest maths). If you want a different shape — gold EAs, TradingView suites, crypto bots, a backtest lab — the other four stacks on the Stacks page are built the same way, with their monthly totals shown up front.
The five pre-flight checks
Run all five before a strategy touches a live account. A strategy that fails a check is not ready; that is the check doing its job.
1. Backtest it on real data
Not on a screenshot, not on a vendor's demo account. Use historical tick data where you can — free Dukascopy data is the classic starting point — and watch for the four traps in Backtesting before you believe it: bad data, slippage and spread, look-ahead, and overfitting.
2. Size every position from risk, not from feeling
The formula is one line: position size = risk amount ÷ distance to your stop. Risk amount is your balance × the percentage you are willing to lose on one trade; distance is entry minus stop. Our position-size calculator does the arithmetic and shows each step — try it with a $1,000 account at 1% risk before you trust any number in your head. The worked examples are in Position sizing for algo traders.
3. Check the grid's worst day, not its best week
If your strategy adds to losers (a grid or martingale), run the grid-risk simulator first. Enter your spacing, levels and lot size and read the loss table: a flat 0.01-lot grid at 20-pip spacing loses $30 at 5 levels, $110 at 10, $240 at 15 and $420 at 20 — and a ×2 martingale reaches $114 by step 5. Grids win small and lose big; the simulator shows the big half before the market does.
4. Prove the VPS stays up
Your terminal lives on the VPS now. Confirm the server location against your broker, check that the terminal restarts after a reboot, and know how you will be told if it goes down. The specs that matter — latency, RAM, Windows support — are in How to choose a trading VPS.
5. Set the kill-switch before you start
Decide, in writing, the loss at which you stop the strategy and walk away — a daily drawdown limit, a total drawdown limit, or both. Write it in your journal on day one. A journal like TradeZella ($35/mo Essential) or TradesViz (Pro $19.99/mo) is where the rule lives and where you check whether you kept it; Trading journals: what to track covers what to log so the review actually teaches you something.
Your starter checklist
- [ ] One strategy seat filled — one EA or bot, not two fighting over the same account
- [ ] VPS running, terminal live on it, restart behaviour confirmed
- [ ] Backtest done on real data, traps checked, results recorded
- [ ] Position-size maths done for your account, risk % chosen in advance
- [ ] Grid/martingale worst case simulated (if applicable) and survivable
- [ ] Journal open, kill-switch rule written down before the first trade
- [ ] First week treated as a test: small size, everything logged, review on Sunday
Keep going
When the checklist is done, the next steps are the rest of the Learn section: crypto bots, honestly if you are heading toward exchanges, and the Tools Library whenever you need the maths done again. The directory itself — 305 tools across 13 categories — is there whenever a layer needs upgrading. Upgrade one layer at a time, and keep the journal open while you do it.