Open Scoring Framework · v1.0

AI Trading Bot Review Methodology

How we score AI crypto trading bots — a 6-criterion weighted scorecard, 5 auto-fail rules, and an honest conflict-of-interest disclosure you rarely see elsewhere.

Framework Version
v1.0 · 2026-05-26
Auto-Fail Rules
5 rules · any hit = elimination
Live Evidence
View /live-proof →

6-Criterion Weighted Scorecard

Out of 100 points. Capital safety carries 30% — we consider losing principal the most expensive failure mode in this category.

Auto-Fail Rules

Hit any rule → no rating

These are structural problems, not matters of degree. No matter how well the other dimensions score, a single hit is disqualifying.

1

Withdrawal permission required on API keys

Once leaked or coerced into an update, funds can be transferred away directly. No legitimate AI trading bot needs withdrawal access — any system that requires it is out under this framework.

2

Custodial setup / funds not held in the user’s own exchange account

A custodial model stacks operator-bankruptcy risk on top of exchange risk. We only evaluate non-custodial systems where the user retains funds.

3

Performance cannot be independently verified on the exchange side

Products that rely solely on vendor screenshots and proprietary dashboards have an extremely low cost of fabricating numbers. A third-party–verifiable path on the exchange is mandatory.

4

Internal broker / OTC matching only — no major public exchanges

Systems that do not connect to Binance / OKX / Bybit / Bitget cannot have their prices or fills externally validated; they are effectively an internal ledger. Out under this framework.

5

No third-party–queryable live evidence whatsoever

Even a single public leaderboard or open stats endpoint is better than zero. Fully black-box products are not evaluated under this framework.

Criteria in Detail

What each dimension evaluates, and what we count as evidence.

Capital Safety & Permission Isolation

Funds never leave the exchange; minimal API permissions; sub-account isolation; Asset Guard–class circuit breakers

30%
Evaluation checklist
  • API permission scope (does it require withdrawal access?)
  • Sub-account or isolated capital channel availability
  • Asset Guard / Equity Guard–class capital protection mechanisms
  • Systemic circuit-breaker or position-reduction logic under extreme market conditions
  • Private key and credential management practices
What counts as evidence

API permission whitelist must be independently verifiable on the exchange side; protection mechanisms must have public documentation or a triggerable demo

Live Performance & Verifiability

Live trading ≠ backtest; floating P&L and drawdown are disclosed; any third party can independently verify

20%
Evaluation checklist
  • Whether live and backtest data are clearly separated
  • Whether floating P&L is published alongside closed P&L
  • Whether maximum drawdown is publicly visible and independently verifiable
  • Whether a non-user can independently query performance (e.g., a public leaderboard page)
  • Whether the test window covers multiple market regimes (bull, bear, range-bound, black-swan)
What counts as evidence

Required: exchange-side dashboards, official public pages, or time-stamped API data. Vendor marketing screenshots do not count.

Cost Transparency

Pricing is clearly comparable; no hidden fees; profit-share terms are unambiguous

15%
Evaluation checklist
  • Whether the pricing model (subscription / profit-share / point card) is public
  • Presence of hidden costs (FX spread, forced upgrades, slippage cuts, etc.)
  • Whether profit-share triggers are clearly defined (closed-profit only / floating profit / gross volume)
  • Whether free-tier vs paid-tier feature differences are documented
  • Whether cancellation and refund rules are user-friendly
What counts as evidence

Public pricing page + user-facing billing receipts that can be reconciled. Point-card products must publish a clear "1 USDT card processes X USDT of profit" conversion.

Regional Localization & Compliance

UI / support language coverage; regional regulatory clarity; broad exchange depth

10%
Evaluation checklist
  • English UI completeness and English customer support availability
  • US user accessibility (including state-level restrictions)
  • UK / EU regulatory standing (MiCA, FCA, etc.)
  • Major Asian-market coverage (Hong Kong, Taiwan, Singapore, SE Asia)
  • Depth of supported exchanges (Binance / OKX / Bybit / Bitget — full or partial)
What counts as evidence

Onboarding and funding flows that a reader from the relevant region can replicate; regulatory stance must be stated publicly or in compliance documentation

User Experience & Setup

5-minute onboarding; zero-tuning default strategy; clear error messages

15%
Evaluation checklist
  • Shortest path from sign-up to first trade (step count and total time)
  • Whether default strategies are beginner-friendly (run without parameter tuning)
  • Whether critical actions (API binding, transfers, position changes) have protective confirmations
  • Mobile experience
  • Whether error messages are readable and guide the user toward resolution
What counts as evidence

Reviewer walks the entire flow with a fresh account, recording time-per-step and any obstacles encountered

Support & Operating History

Response speed, knowledge depth, years operating + record of handling black-swan events

10%
Evaluation checklist
  • Customer support response time (business hours / 7×24)
  • Completeness of FAQ / knowledge-base documentation
  • Public disclosure of operating years and team information
  • How the platform handled major historical events (hacks, outages, extreme market moves)
  • Community atmosphere (Telegram / Discord / forums)
What counts as evidence

At least 3 live support tests; public incidents must include timestamps and resolution outcomes

Limitations

What we do not evaluate

Honesty matters more than completeness — below are the boundaries of this framework.

We do not audit backend code. Unless the vendor open-sources or publishes a third-party audit, we do not pretend to see inside the model.
We do not verify "total users / total profit" marketing claims — only the independently-verifiable subset (e.g., numbers on a public leaderboard).
We do not predict future performance. This framework evaluates currently visible evidence and is not investment advice. Markets change; so does AI.
We do not evaluate long-term stability beyond our testing window. Our live-trade window is 30+ days, which exposes a bot to multiple regime shifts but cannot replace years of operating record.
Conflict of Interest · Honest Disclosure

What's our relationship with CoinTech2u?

This site is an official CoinTech2u promotion channel and earns affiliate revenue when readers sign up. We do not pretend to be neutral.

Almost every seemingly "independent" AI-trading-bot review site collects affiliate commissions — the only difference is whether they admit it, and how prominently. The common playbook is to cultivate an "independent lab" persona while burying the affiliate disclosure in fine print at the bottom of the page.

We chose differently. We openly identify as a channel, and we want you to notice what that means —

  • If we publish data that gets disproven, CoinTech2u will immediately terminate the relationship and pursue consequences.
  • If we mislead readers into bad decisions, the channel status that funds this site is gone.
  • Our reverse incentive — sustaining long-term revenue — is perfectly aligned with being honest with you.

Put another way: the neutrality of a so-called "independent review site" rests on a promise. Our neutrality rests on aligned financial incentives — which is usually the more reliable mechanism. But you don't have to take our word for it: the live data and the official leaderboard page are both independently verifiable.

Products We've Applied This Framework To

The framework is the rule. Below are the results of applying it.

Reviewed

CoinTech2u

Live data pulled hourly directly from the official system — no editorial filtering.

See live evidence →
In review

3Commas / Pionex / Bitsgap

Scored against the 6 criteria + checked against auto-fail rules. Results will be published.

Coming soon
Candidate

Reader Suggestions

Which AI trading bot would you like to see scored under this framework?

Tell us →

Version Changelog

This framework will evolve as we apply it. Every meaningful revision is logged here, and prior versions are not deleted.

v1.0 2026-05-26 Current

Initial release: 6-criterion weighted scorecard (Capital Safety 30% · Live Performance 20% · Cost Transparency 15% · Regional Localization 10% · UX 15% · Support 10%) + 5 auto-fail rules + honest conflict-of-interest disclosure.

You've seen the rules — now see the evidence

Rules are what we say. Evidence is what we have. Both are here. Judge for yourself.

Start Free Trading →