Go-live — the 12-step protocol from demo-ready EA to live account deployment
Going live with an EA is not a single decision — it is a structured sequence of 12 verifiable steps. Each step either proves the EA is ready or surfaces a reason to wait. This lesson is the final checklist before your first live tick.
अंतिम समीक्षा:
The 90-second version
Every pilot runs a pre-flight checklist before every flight — not because they distrust the aircraft after 10,000 hours, but because a missed item on the ground is cheaper than a missed item at 35,000 feet. Your go-live checklist is the same logic applied to automated trading.
- Most EA failures on live accounts happen in the first 30 days — nearly always due to a skipped infrastructure or broker step, not a strategy failure
- The broker steps are about counterparty risk: you need to verify the broker is regulated and that you can get your money out before you put money in
- The VPS step is about uptime reliability: an EA that misses entries or can't close positions because your home internet went down is not a deployed EA — it is a risk
- The EA verification steps are about baseline: you need a documented record of what the EA does on demo so you can distinguish 'normal variance' from 'something is wrong' on live
- The capital steps are about survival: starting small and reviewing at 20 trades gives you real live data at negligible cost before you scale to your intended position size
FundedNext's 2024 monitoring dashboard revealed that traders who documented their go-live setup — inputs, VPS region, broker, initial lot size — resolved account breach disputes 40% faster than those who had not, because they could prove what the EA was configured to do.
FundedNext ने 2024 में एक live monitoring dashboard लॉन्च किया जो funded traders को उनके EA का equity curve, daily drawdown, trade count और win rate real time में दिखाता है — 'मेरा account breach क्यों हुआ' वाले support-ticket की मात्रा 40% कम करते हुए।
स्रोतLaunch Protocol Tracker
Check off each step as you complete it. Any step can be checked or unchecked independently — this is not a linear wizard. Complete all 12 before going live.
Why each step exists
Why a go-live protocol exists — the cost of skipping steps
Commercial aircraft have pre-flight checklists that haven't changed fundamentally since the 1930s. They are not there because pilots are incompetent or because aircraft are unreliable — they are there because human memory under pressure is unreliable, and a single skipped item in a standard sequence can cascade into something unrecoverable. The checklist does not express distrust. It expresses the opposite: a professional's respect for what can go wrong when something simple is assumed rather than verified.
An EA go-live protocol is the same structure applied to a different domain. Your EA has been running on demo for weeks. You know the strategy. You have reviewed the backtest. You have watched the demo trades execute. None of that makes the protocol redundant — it makes it necessary. Because the moment you add real capital, every assumption you did not verify becomes a variable that could produce a loss you didn't expect.
Consider the four categories of the 12-step protocol. Broker steps address counterparty risk: if the broker is unregulated, undercapitalised, or operationally dysfunctional, the risk is not in the market — it is in the institution holding your funds. Infrastructure steps address operational reliability: an EA that runs perfectly on your home PC for three hours but misses two hours of trading because your internet dropped is not a deployed system. EA steps address behavioural verification: the difference between 'I believe my risk rules are configured correctly' and 'I have confirmed my risk rules fire correctly on demand.' Capital steps address sequencing logic: why the first 20 live trades should be sized at minimum, not at your intended production lot.
Each skipped step is not just a missing box on a checklist. It is a risk that has been consciously moved from the 'known' column to the 'uncontrolled' column. Professional traders keep the 'uncontrolled' column as short as possible. The protocol is how you do that.
The 12 steps by category — what each one verifies
Broker category (Steps 1–3). Step 1 — Verify broker regulation and fund segregation: check the regulator's public register to confirm the broker holds a valid licence and that client funds are segregated from company funds. This step takes 15 minutes and rules out the most serious category of counterparty risk. Step 2 — Open a live account matching your demo environment: the live account must use the same account type (ECN/Standard), the same base currency, the same leverage, and the same server as your demo. Spread, commission, and swap structures should be identical. Any difference means your demo results are not predictive of live behaviour. Step 3 — Test a withdrawal process on demo: before depositing real funds, verify that the broker's withdrawal process works. Make a small deposit, initiate a withdrawal, confirm it processes within the stated timeframe. If the broker cannot demonstrate a clean withdrawal before you deposit, this is a red flag.
Infrastructure category (Steps 4–5). Step 4 — Provision a VPS in the same region as your broker's server: most brokers publish their execution server location (LD4, NY4, TY3 are the major forex co-location hubs). Choose a VPS provider that operates in the same facility or the nearest region. The goal is latency under 5ms between the VPS and the broker's execution server. Step 5 — Verify MT5 connects and EA runs 24h on VPS demo: log into the VPS, install MT5, connect to your broker's demo server, attach the EA, and leave it running for 24 hours. Check the Expert Journal at the end. Any connection drop, delayed execution, or missing entry in the log during that window is a fault to resolve before going live.
EA category (Steps 6–9). Step 6 — Confirm demo results match backtest expectations: compare the last 30 demo trades against the backtest's average win rate, average trade duration, and average profit factor. You are not looking for a match — variance is expected. You are looking for a plausible range. If the demo win rate is 22% against a backtest win rate of 61%, something structural is wrong (overfitting, wrong symbol, wrong timeframe). Step 7 — Verify all risk rules fire correctly: run the deliberate-trigger test from Lesson 9 — trip the daily loss limit, trip the drawdown cutoff, confirm the news filter logs its message, verify magic number isolation. Document the results. Step 8 — Document all input settings in a permanent record: write down every EA parameter value, the date it was set, the version of the EA, and the account it was attached to. This is your baseline. If behaviour changes, you need the original configuration to diagnose the cause. Step 9 — Sign off against your personal go-live criteria: define in advance what 'ready' looks like. Minimum demo track record length, minimum win rate threshold, maximum acceptable drawdown. If the EA does not meet those criteria, it is not ready, regardless of how confident you feel about it.
Capital category (Steps 10–12). Step 10 — Start at minimum lot size: on most retail accounts this is 0.01 lots. The purpose is not to be conservative — it is to collect live data at negligible financial cost. Live execution, live spreads, live slippage, and live swap conditions may differ from demo in ways that only appear once real money is on the line. Step 11 — Review after 20 completed trades: 20 trades is a statistical minimum for any meaningful pattern assessment. At 20 trades, calculate win rate, average R multiple, and whether actual fills matched demo fills closely. If there are systematic fill differences, resolve them before scaling. Step 12 — Scale to planned lot after consistent performance: 'consistent' means the first 20 trades performed within your pre-defined acceptable range. Double the lot size only when you have written evidence that the live account behaves as expected at the current size. Scale in increments, not in a single jump.
After go-live — the first 30 days review
Going live is not the end of the protocol — it is the beginning of a monitoring cycle. In the first 30 days, three things require daily attention: account equity (is the drawdown within the range seen on demo?), fill quality (are entries and exits executing at prices comparable to demo?), and Expert Journal (is the EA logging normally, or are there unexpected errors or skipped entries?). Any of these three diverging significantly from the demo baseline triggers a pause: stop trading, do not close positions hastily, review the logs, compare input settings against your documented baseline.
A pause is not a permanent stop. A pause is a decision to collect more information before the EA continues trading. The distinction matters because many traders either over-react (close everything and restart without diagnosis) or under-react (ignore the signal and let the divergence continue). The protocol is: observe the divergence, compare to the baseline, identify the cause, fix the cause, restart with documented changes.
What triggers a permanent stop rather than a pause? Three conditions: (1) the EA breaches its drawdown cutoff and you have no confidence in the diagnosis — the EA should not be restarted until the cause is understood; (2) a broker execution issue surfaces that you cannot resolve — if fills are systematically 3+ pips off the expected price and the broker will not engage, move to a different broker; (3) the strategy's core assumption is invalidated by a structural market change — this is rare in the first 30 days but possible if a major regulatory or liquidity change affects the instrument. In all other cases, a divergence is an investigation, not a failure.
Scaling capital after a successful first month follows the same step-by-step logic as the initial deployment. You have 30 days of live data. Calculate the metrics again at the end of the month: win rate, profit factor, maximum drawdown observed. If all three are within the acceptable range defined in Step 9, increase the lot size by a factor of two. Wait another 30 days. Repeat. Never jump to production size in a single step — the scaling ladder is itself a verification protocol, not a display of caution.
Key terms
Go-live monitoring in the real world
In 2024, FundedNext launched a real-time EA monitoring dashboard that showed funded traders their equity curve, daily drawdown consumption, trade count, and win rate. The product emerged directly from the support-ticket data: the firm found that 60–70% of account breach complaints could not be diagnosed because the trader had no record of what their EA was configured to do at the time of the breach.
FundedNext ने 2024 में एक live monitoring dashboard लॉन्च किया जो funded traders को उनके EA का equity curve, daily drawdown, trade count और win rate real time में दिखाता है — 'मेरा account breach क्यों हुआ' वाले support-ticket की मात्रा 40% कम करते हुए।
FundedNext's monitoring dashboard reduced 'why was my account breached' support tickets by 40% in its first six months. The reduction came not from the dashboard preventing breaches — it came from the documentation the dashboard enforced. Traders who used the dashboard had their EA settings, equity curve, and trade log in one place. When a breach occurred, the diagnosis was a 10-minute conversation rather than a two-week investigation. The lesson extends beyond prop firms: every EA deployment benefits from the same documentation discipline, whether the account is funded or personal. What the FundedNext case demonstrates is that the 12-step go-live protocol — and especially Step 8, the permanent input record — is not bureaucracy. It is the minimum evidence base required to diagnose and fix a live EA that is behaving unexpectedly. Without that record, you are not running an automated trading system. You are running an automated guessing game.
स्रोतPractice
Run the full 12-step checklist for your current or planned EA
Work through each step for an EA you currently use, plan to use, or are evaluating. If you do not have a live account yet, complete all steps you can complete on demo — and identify explicitly which steps are blocked until you open a live account.
- 1
Broker verification: look up your broker on the regulator's public register (FCA, ASIC, CySEC, or the relevant authority for your region). Confirm the licence is current and active. Check that the account agreement states client funds are held in segregated accounts. Write down the broker name, regulator, and licence number. If you cannot find a current, valid licence, do not deposit.
- 2
Account parity check: open your demo account settings and your live account settings (or the live account application form) side by side. Confirm: same account type, same base currency, same leverage, same execution server. If your broker offers multiple servers, identify which server your demo is connected to, and ensure the live account uses the same server. Differences in server mean differences in spreads, latency, and available instruments.
- 3
VPS region check: find your broker's published execution server location (most reputable brokers publish this in their FAQ or infrastructure page — look for references to LD4 London, NY4 New York, or TY3 Tokyo). Select a VPS provider with a presence in the same region. Install MT5 on the VPS and measure the ping from the VPS to the broker's server (MT5 shows server ping in the bottom-right corner of the platform). Target under 5ms. Document the result.
- 4
24-hour VPS demo run: attach your EA to a demo account on the VPS. Leave it running for exactly 24 hours without logging in remotely. After 24 hours, open the Expert Journal and read every entry. Look for: connection drop messages, order placement errors, missed entries during active session hours, and any message that includes 'error' or 'failed'. If any of these appear, diagnose and fix before going live.
- 5
Risk rule verification: run the deliberate-trigger test for every risk rule configured in the EA (see Lesson 9). For each rule, write down: the parameter name, the value set, the condition that triggered the rule, and the Expert Journal message that confirmed it fired. If a rule has no corresponding Expert Journal message, it may not be implemented correctly — investigate before deployment.
Mastery check
Four questions. Pass at 75% (3/4). Focus on why each step category exists — not on memorising step numbers.
Mastery check — Lesson 12
4 प्रश्नों के साथ अपनी समझ परखें। पास होने के लिए 75/4 सही चाहिए।
Reflect
चिंतन
अपने ईमानदार उत्तर लिखें — केवल इसी डिवाइस पर सहेजे जाते हैं। अगले सप्ताह इनका उपयोग अपनी ट्रेडिंग सोच में पैटर्न पहचानने के लिए करें।
Pro deep dive
The 12-step protocol covers individual EA deployment. Professional operations add two layers: multi-EA portfolio deployment coordination, and post-launch quantitative review frameworks that use statistical criteria rather than gut feeling to make scaling decisions.
Multi-EA portfolio go-live sequencing
When deploying multiple EAs on the same account, the 12-step protocol must be run for each EA individually, but the capital steps (10–12) must be coordinated at the portfolio level. The core constraint: the combined drawdown of all live EAs must stay below the account's configured drawdown cutoff, regardless of individual EA performance. This requires calculating the expected worst-case simultaneous drawdown across all EAs before the first trade is placed on any of them. The calculation uses the EAs' historical drawdown profiles and their pairwise correlation: if two EAs trade the same instruments and are likely to draw down simultaneously during the same market conditions, their worst-case combined drawdown is approximately the sum of their individual worst cases. If they are genuinely uncorrelated, the combined worst case is approximately the larger of the two. A portfolio go-live sequence typically staggers deployments — EA-1 runs for 20 trades before EA-2 is activated, EA-2 runs for 20 trades before EA-3, and so on. This produces a verified baseline for each EA before the portfolio-level stress is introduced.
Statistical frameworks for scaling decisions
The 20-trade review in Step 11 is a practical minimum. Statistical significance for a trading system's win rate requires a much larger sample — typically 100–200 trades — before the confidence interval narrows to a useful width. A 20-trade sample with a 55% observed win rate has a 95% confidence interval of roughly 33–77%. That range is too wide to draw strong conclusions. What the 20-trade review does is different: it tests for gross failure (win rate below 30% is structural, not variance) and for operational correctness (are the fills, positions sizes, and risk rules behaving as designed?). For scaling decisions based on statistical significance, wait for 50+ trades and apply a t-test or Wilson score interval to the win rate. For scaling decisions based on operational verification, 20 trades is sufficient. These are different questions requiring different sample sizes.
Prop-firm specific go-live considerations
Prop-firm accounts add a layer of constraints on top of the standard 12-step protocol. Every prop firm publishes a set of trading rules — maximum daily drawdown, maximum overall drawdown, minimum trading days, prohibited strategies (grid, martingale, news trading, latency arbitrage) — and violation of any rule results in account termination regardless of profitability. Before Step 6 (EA demo performance review), add a prop-firm constraint mapping step: list every rule the prop firm imposes and verify that the EA's configured inputs cannot violate any of them. Daily loss limit in the EA must be set below the prop firm's daily drawdown limit (not at — below, to give a buffer). Maximum drawdown cutoff must be set at least 2% above the prop firm's drawdown limit. News filter must be enabled if the prop firm prohibits news trading. Magic numbers must be unique per challenge account. Violating a prop-firm rule is different from losing on a trade — the former terminates the account regardless of the account's profitability. The standard risk rules and go-live protocol map directly onto prop-firm requirements, but the calibration of each parameter requires the firm's specific rules as constraints.
Post-launch monitoring infrastructure
Professional EA operations use monitoring infrastructure that goes beyond the MT5 Expert Journal. The minimum professional setup includes: a trade journalling platform (Edgewonk, TraderVue, or a spreadsheet tracking every trade's entry/exit, spread cost, net P&L, and fill quality against expected price); a VPS uptime monitor that sends an alert if the VPS loses connection for more than 5 minutes; and a daily P&L summary that compares the current drawdown-from-peak against the go-live baseline. FundedNext's 2024 monitoring dashboard demonstrated that automated monitoring dramatically reduces the response time to anomalies — the data that FundedNext centralised for funded traders is the same data that a rigorous individual trader should be tracking locally. The tools exist for free (FXBlue for MT5 statistics export, UptimeRobot for VPS monitoring, Google Sheets for trade journal). The discipline is the variable.
Sources
उत्तर दिखाएं
Broker (Steps 1–3) addresses counterparty risk — verifying regulation, account parity, and withdrawal process. Infrastructure (Steps 4–5) addresses operational reliability — VPS region, latency, and uptime. EA (Steps 6–9) addresses behavioural verification — demo results, risk rule testing, input documentation, and personal go-live sign-off. Capital (Steps 10–12) addresses deployment sequencing — starting at minimum size, reviewing at 20 trades, scaling only after verified performance.
केवल शैक्षिक सामग्री — निवेश सलाह नहीं। ट्रेडिंग में पूंजी हानि का जोखिम होता है। हमेशा डेमो पर अभ्यास करें और स्टॉप-लॉस का उपयोग करें। ← Automated Trading पर वापस