A trading robot candidate is ready for the next stage only when the available evidence justifies a stronger form of testing.
A positive result is not enough.
A large backtest is not enough.
One successful paper week is not enough.
Even a technically clean demo period does not prove that a robot is ready for live-money trading.
The correct question is not:
Did the robot perform well?
The better question is:
Do we have enough evidence to justify the next specific research step?
What Is a Trading Robot Candidate?
A trading robot candidate is a system that has progressed far enough to deserve continued evaluation.
It may already have:
- historical test results;
- paper observations;
- demo observations;
- technical logs;
- weekly reports;
- version history;
- documented research decisions.
The word candidate is important.
It means the system is still being evaluated.
It is not a guarantee that the robot will become a finished product or that it will be suitable for live trading.
Readiness Depends on the Next Stage
There is no single universal definition of “ready”.
A robot may be ready for:
- historical screening;
- paper observation;
- controlled demo testing;
- another week of unchanged observation;
- a technical review;
- a new version;
- discontinuation.
Each decision requires different evidence.
For example, a robot does not need broker-execution evidence before entering paper simulation.
But a move from paper testing to demo execution may require confidence that the logic, safety controls and technical state are sufficiently stable.
Readiness must therefore be evaluated relative to the next research question.
A Positive Result Does Not Automatically Mean Ready
A robot can produce a positive result for many reasons.
It may have benefited from:
- a favourable market regime;
- a short volatility pattern;
- one unusually strong trade;
- a small sample;
- random variation.
A useful readiness review should ask:
- How many trades produced the result?
- How was profit and loss distributed?
- Did the robot behave as expected?
- Were there technical problems?
- Did the system remain within its intended risk structure?
- What evidence is still missing?
A positive number may justify more observation.
It does not automatically justify promotion to a stronger stage.
Sample Size Matters
A system with five observed trades provides much less evidence than one with a larger and more varied sample.
Small samples can be misleading.
A candidate may look unusually strong or weak simply because too little has happened.
Questions can include:
- How many trades occurred?
- How many different days or weeks were observed?
- Were different volatility conditions included?
- Did the robot experience losing periods?
- Was performance dependent on a small number of events?
More trades do not guarantee future success.
They simply provide more information.
Historical Evidence Is Only the Beginning
Historical testing can show whether the strategy had useful structure on past data.
It can help identify:
- trade frequency;
- drawdown;
- weak conditions;
- directional imbalance;
- parameter sensitivity;
- periods where the logic failed.
But a historical result is still based on known past data.
A strong backtest may justify forward observation.
It does not prove that the candidate is ready for stronger execution testing.
For more background:
Backtesting vs Forward Testing for Forex Robots
Forward Evidence Adds Another Layer
Forward observation checks how the system behaves on new market data.
This may involve:
- paper simulation;
- controlled demo execution;
- another non-live research environment.
Forward testing can reveal problems that were not obvious historically.
For example:
- trade frequency may be too low;
- signals may appear at unexpected times;
- filters may behave differently;
- drawdown may develop differently;
- technical continuity may fail;
- the historical hypothesis may not survive.
These findings affect readiness.
When Paper Observation Is Enough for the Next Step
Paper simulation can be useful when the main research question concerns strategy behaviour rather than broker execution.
A paper candidate may be considered for demo testing when:
- the logic has behaved consistently enough to justify broker interaction;
- state continuity is reliable;
- open positions are tracked correctly;
- safety rules are defined;
- the candidate has no unresolved technical problems;
- the next unanswered question concerns actual demo execution.
This does not mean the strategy is proven profitable.
It means the research question has changed.
When More Paper Observation Is Better
A candidate may remain in paper simulation when:
- trade frequency is very low;
- the sample is too small;
- results are mixed;
- technical state recovery is uncertain;
- filters are still being evaluated;
- the strategy has not experienced enough varied conditions.
Continuing paper testing is not a failure.
Sometimes the evidence is simply incomplete.
Demo Observation Tests Different Questions
A demo environment allows the robot to interact with broker infrastructure.
This can provide evidence about:
- order submission;
- broker responses;
- position management;
- lot validation;
- safety restrictions;
- execution flow;
- recovery after interruption.
These are operational questions that paper simulation cannot fully answer.
A robot may therefore be ready for demo observation even when its future profitability remains uncertain.
Demo Success Does Not Mean Live-Ready
This distinction is critical.
A candidate can operate successfully in a demo environment and still be unsuitable for live-money trading.
Demo accounts can differ from live accounts in:
- execution quality;
- slippage;
- liquidity;
- spread behaviour;
- order handling;
- other broker-specific conditions.
A demo result can justify continued demo observation.
It does not create a guarantee of live performance.
Technical Stability Is Part of Readiness
A strategy can have promising trading logic and still be unready because the software is unstable.
Technical review may include:
- duplicate-order prevention;
- correct trade-state recovery;
- consistent timestamps;
- correct position tracking;
- safety restrictions;
- complete logs;
- clear version identification.
A robot that cannot be audited reliably should not move to a stronger stage simply because its trading result looks good.
Risk Must Match the Research Stage
Readiness also depends on risk control.
Important questions include:
- What is the maximum observed drawdown?
- How large are losing streaks?
- Does exposure increase after losses?
- Are position sizes controlled?
- Can live-account execution be blocked when required?
- Are safety limits appropriate for the testing stage?
A candidate should not progress by taking more risk simply to create stronger-looking results.
Version History Must Be Clear
Each version should have a reason for existing.
A useful history should explain:
- what weakness was observed;
- what changed;
- why it changed;
- what remained unchanged;
- what evidence the new version is expected to collect.
If parameters are changed repeatedly without a clear reason, it becomes difficult to know what is actually being tested.
For more background:
How We Track MT5 Robot Versions
Sometimes “No Change” Is the Correct Decision
A new stage does not always require a new version.
Sometimes the strongest decision is:
continue unchanged.
This can be appropriate when:
- the sample is still growing;
- the logic has not produced a clear failure;
- technical operation is stable;
- changing parameters would make evaluation harder.
Frequent changes can destroy the value of observation because the system being measured never remains stable long enough.
Weekly Reports Help Support Readiness Decisions
A weekly report creates a structured checkpoint.
It can summarize:
- trade activity;
- positive and negative results;
- technical findings;
- open positions;
- changes from the previous period;
- the next research decision.
This helps prevent one trade or one day from determining the status of a candidate.
For more detail:
How Weekly Reports Help Track Trading Robots
A Candidate Can Be Ready to Stop
Readiness does not always mean promotion.
Sometimes the evidence is strong enough to conclude that a research branch should end.
Reasons may include:
- persistent negative behaviour;
- insufficient activity;
- unstable logic;
- excessive drawdown;
- technical complexity that does not justify the result;
- a stronger research direction becoming available.
Stopping a weak branch is a valid research outcome.
Readiness Should Be a Decision, Not a Label
Labels such as:
Tested Logic
or:
Demo Candidate
can be useful for organization.
But the label itself should not become the evidence.
The important questions are:
- What has actually been tested?
- What was observed?
- What remains unknown?
- Why is the next stage justified?
A clear explanation is more valuable than a status name alone.
See Real Research Decisions
FX Trading Robot Lab publishes older completed research publicly when appropriate.
These materials show how robot candidates were continued, modified, moved between stages or discontinued.
You can browse them through:
Unmarked materials are publicly accessible.
Entries marked MEMBERS require an active membership, so the access requirement is visible before the material is opened.
When Is a Trading Robot Candidate Ready?
A trading robot candidate is ready for the next stage when the current evidence is strong enough to justify the next research question.
That may mean:
- move from historical testing to paper observation;
- continue paper observation;
- move from paper to demo;
- continue demo testing;
- create a new version;
- stop the branch.
The decision should be based on evidence, not optimism.
Risk Warning
Trading Forex, CFDs and other financial instruments involves significant risk and may result in the loss of capital.
Historical tests, paper observations, demo results, trading robot candidate reviews and past performance are not reliable indicators of future results.
Moving a candidate to a stronger testing stage does not guarantee profitability or suitability for live-money trading.
FX Trading Robot Lab does not provide investment advice, financial advice, managed account services, copy trading, trading signals or guaranteed trading results.
All material is provided for research and educational purposes. Users remain responsible for their own financial decisions, independent testing, broker choice, account configuration and risk management.