Promote a strategy from replay to live in Botfolio

When this how-to matters

• Replay looks “pretty good” and impatience is rising
• You changed risk or session rules and need a fresh promote decision
• You are entering a new account stage (combine vs funded)
• You historically went live on courage instead of evidence

Promotion criteria (process checklist)

Promote only if:

• Rules are fully specified (including stand-down and session bounds)
• Replay covered multiple regimes, not one lucky week
• Risk per trade fits the stage budget
• You have practiced stop / flatten / disconnect
• Credentials are stored safely
• You will supervise early live sessions
• Current Topstep rules were re-checked
• Demotion criteria are written

If any box is theatrical rather than true, do not promote.

What “promote” means in the product story

1. Replay-ready historical sessions informed the rules
2. Rules approved — entry, exit, sizing, stand-down explicit
3. Controls armed — stop, disconnect, flatten, diagnostics visible
4. Live path connected under supervision

That sequence is the homepage narrative translated into ops.

Size promotion separately from logic promotion

Changing MES→MNQ→ES/NQ style sizing is its own promotion. See micros vs minis. Promoting both logic changes and size changes simultaneously makes diagnosis painful.

After promotion: watch the right things

• Connection state
• Working orders / protective exits
• Limit proximity (day loss / stand-down)
• Whether you are tempted to override emotionally
• Whether live friction matches replay assumptions enough to stay honest

Use the first session checklist.

Hypothetical promote board (not a result)

Promote: Rule sheet complete, replay packet includes ugly days, stand-downs validated, flatten drilled, micros size, supervised calendar blocked.

Demote: Replay only includes friendly trend days, stand-downs undefined, operator plans to “watch on phone between meetings.”

The demote decision can be the highest-EV click in the whole week — and it is not a trading signal claim; it is process hygiene.

Common failure modes

1. Promoting to end boredom
2. Promoting after one green live day as if research is over
3. Changing three parameters during promote week
4. No demotion criteria
5. Treating promotion as permanent graduation

Operator checklist (promote day)

1. Re-open rule sheet
2. Re-skim official Topstep pages
3. Verify risk controls
4. Verify connection while flat
5. Enable only the promoted strategy
6. Supervise
7. Journal divergences immediately

What Botfolio does / does not do

Does

• Market a deliberate promote step between replay and live
• Require explicit account/sizing/risk settings in the story of promotion
• Expect supervision afterward

Does not

• Guarantee live will match replay
• Guarantee profits
• Remove platform risk
• Make promotion irreversible wisdom

Promotion grades (use whatever labels you like)

Grade A — promote under supervision: evidence pack complete, controls drilled, size conservative, calendar blocked.

Grade B — promote to paper/controlled conditions only if your process includes that concept: still not a profit claim; still supervised learning. (If your stack lacks a clean intermediate mode, stay on replay longer rather than inventing one.)

Grade C — demote: incomplete rules, highlight-reel replay, unpracticed flatten, unsupervised plan.

Be honest about grade inflation. The market does not grade on a curve.

Communication template if you trade around family/work constraints

If someone else’s schedule interacts with your supervision plan, say:

• “I am live only during these windows.”
• “If I am interrupted, the bot must already have stand-downs and I must be able to flatten quickly.”
• “I will not enable live mode during meetings.”

Supervised automation includes social reality. Botfolio’s command-center model assumes attention is part of the system.

Post-promote first five sessions

Track a simple scorecard:

• Did connection behavior match expectations?
• Did stand-downs behave?
• Did exits/protective orders behave?
• Did I override?
• Do I still understand the rules cold?

If overrides pile up, you did not promote a strategy — you promoted a debate.

Promote packet contents (copy/paste outline)

```text

Strategy name:

Date frozen:

Rule sheet link/notes:

Replay dates reviewed:

Stand-down observations:

Session window:

Fixed risk + contract choice:

Emergency drill date:

Credentials storage confirmed (y/n):

Official Topstep pages reviewed (date):

Supervision plan (sessions/times):

Demote triggers:

Signer (you):

```

If you cannot fill it, you are not ready.

Soft-launch vs hard-launch mindset

Soft-launch: promoted logic, conservative size, max supervision, quick demote willingness.

Hard-launch: same logic after soft-launch behavior matches expectations, still supervised, still not a profit claim.

Many traders skip soft-launch because it feels unserious. In prop contexts, unserious is skipping brakes.

Promote anti-patterns unique to automation

1. Promoting because the backtest equity curve is pretty in one rendering
2. Promoting to justify the software purchase emotionally
3. Promoting five correlated ideas at once
4. Promoting and immediately leaving the desk
5. Promoting without a demote path, as if admission of error were forbidden

Botfolio’s marketed loop includes iterate/refine. Demotion is part of iteration, not a moral failure.

Promote decision tree (text form)

1. Are rules frozen and complete? If no → demote to rule sheet.
2. Does replay evidence include ugly days? If no → demote to replay lab.
3. Are risk/stand-down/window settings explicit? If no → demote to risk config.
4. Are emergency controls drilled? If no → demote to drill day.
5. Is supervision actually available for early sessions? If no → wait.
6. Were official Topstep pages reviewed recently? If no → read, then revisit.
7. Only then promote one strategy at conservative size.

If you skip the tree because “I know,” you are using confidence as a control surface. Confidence is not a control surface.

Metrics that matter after promote (process metrics)

Track for two weeks:

• Override count
• Stand-down fires and whether they felt early/late
• Unclear-state events
• Times you changed rules mid-session (should be ~zero)
• Times you demoted (can be healthy)

Do not use short-window P&L as the sole promote validation. Short-window P&L lies cheerfully.

Freeze discipline: what “rules frozen” actually means

Frozen means you will not casually edit entry logic mid-session because a candle annoyed you. It does not mean you will never improve the system. Improvements go through replay and a new promote cycle.

Create a freeze tag in your notes: FROZEN-YYYY-MM-DD. If live notes require a logic change, the tag becomes THAW-CANDIDATE and the strategy returns to replay. This lightweight ceremony prevents silent mutation.

Promote-day environment setup

• Quiet enough to supervise
• Runbook open
• Official Topstep tabs available if conduct questions appear
• No ambitious “also launch two other bots” agenda
• Food/water handled so you do not abandon the desk for the wrong reasons during the first hour

Environment is part of supervised automation. Botfolio cannot supply your attention.

Handling early green results without mythology

A green first promote session proves you had a green session. It does not prove the evidence pack was unnecessary. Keep process grading primary for at least several sessions before any size conversation.

Extended operator notes for promote strategy from replay to live

This section exists to make the page operationally complete enough for evergreen SEO depth without inventing product UI fiction. Use it as a working appendix.

Pre-commit questions

Before you claim this part of the Botfolio workflow is “handled,” answer:

1. What artifacts prove it (notes, drill logs, evidence packs)?
2. What would a skeptical co-pilot say is still fuzzy?
3. What official Topstep constraint could invalidate today’s plan?
4. What Botfolio control will you use if live state disagrees with the plan?
5. What demote/abort condition is written down?

Failure rehearsal

Write a short paragraph describing how this area fails in the wild (not how it succeeds). Then write the containment step that uses supervised controls rather than hope. If you cannot describe failure, you have not finished design.

Weekly maintenance

Revisit this topic weekly for the first month of live use, then monthly. Tools drift, emotions drift, and stage rules can change. Evergreen process only stays evergreen if someone maintains it.

Boundary with claims

Nothing in this appendix promises profits, payouts, Combine passes, or risk-free trading. It is process depth around verified Botfolio themes: replay, promote, supervise, local credentials, risk limits, and emergency actions. Confirm in-app details for your version whenever this appendix meets the real UI.

Quick takeaways

• You can install quickly; promotion should still wait on evidence and controls practice.
• No. Promotion means the process gate passed — not that markets owe you anything.
• Expect differences. Pause, diagnose, re-replay with better assumptions, then re-promote deliberately.
• Usually no. Parallel promotes multiply uncertainty.