Extremely wide initial stop-loss
The source configures stoploss = -0.345. That is a configured threshold, not a guaranteed maximum loss or account drawdown. Overrides, leverage, gaps, fills and slippage require separate validation.
Public audit case study
Bandtastic: what a 53/100 source-code audit found.
Bandtastic is a public community strategy for Freqtrade. Vetta reviewed the source as text only and retained seven findings with exact line evidence—without importing the strategy, running it, or claiming it can make money.
Public community strategyView the public source
Check the evidence, not just the score.
- Fixed source commit (UTC)
- Source evidence rechecked (UTC)
- Original audit date: unverified
- No audit rerun in this review
The links below pin a historical source revision, not the latest strategy. This review checked the quoted code and line references only. The published 53/100 score, severity labels and Reject verdict are Vetta's audit judgments, not independently reproduced performance results. No backtest market data or live trading evidence is used here.
Sanitized report excerpt
The report, without private account data.
This public rendering reproduces the production score, classification, primary risk, and two evidence excerpts. Account email, audit ID, source hash, internal records, and the private report URL are excluded.
07 / 07
Seven findings survived the evidence gate.
Every retained finding includes a source line and exact snippet. Three unsupported model findings were removed before the final report was stored.
Entry confirmation filters disabled
RSI, MFI, and EMA entry guards default to disabled, leaving the lower-Bollinger-band trigger with limited confirmation.
Parameter default outside its declared range
buy_slowema declares a range of 1–126 but a default of 364. The target Freqtrade release should be checked before optimization or deployment.
Large hyperparameter search space
Sixteen tunable parameters increase the need for out-of-sample, walk-forward, and regime-separated validation.
ROI and sell-signal interaction needs testing
The ROI schedule and upper-band sell condition can produce different exit timing; only empirical testing can establish which path dominates.
EMA calculation errors can be silent
A broad exception handler can suppress EMA calculation failures, reducing observability when an indicator path breaks.
No Freqtrade protections block
The reviewed source declares no protections block. This whole-file observation does not establish that runtime configuration, inherited behavior or another deployment layer has no protections.
Critical evidence
A configured stop is not a loss guarantee.
stoploss = -0.345
trailing_stop = True
trailing_stop_positive = 0.01
trailing_stop_positive_offset = 0.058The report separates the observed literal values from the inference. The source proves the stop and trailing settings; loss distribution, slippage, and live behavior still require testing.
Fixed source lines L36–L42Medium evidence
The default exceeds its own declared range.
buy_slowema = IntParameter(low=1, high=126, default=364, space='buy', optimize=True, load=True)This is a concrete configuration inconsistency, not a profitability claim. It should be validated against the exact Freqtrade parameter implementation in use.
Fixed source lines L46Method boundary
Static source review, not a backtest.
Vetta parsed the source as text, extracted strategy facts, ran deterministic checks, and retained only model findings whose evidence bound to exact source lines. It did not execute code, verify returns, model fees or liquidity, or prove deployment safety.
What should be tested next
- 01Validate the IntParameter default behavior in the target Freqtrade version.
- 02Stress-test tighter stop-loss settings across trend, range, and high-volatility periods.
- 03Enable and compare at least one entry confirmation or regime filter.
- 04Run official lookahead and recursive-analysis checks.
- 05Use out-of-sample, walk-forward, fee, slippage, and dry-run validation before risking capital.
Private strategy review
Audit your own strategy before risking capital.
Upload one Freqtrade strategy for a private static audit. Your code is parsed as text and is never imported or executed.
Login and 1 audit credit required. This public case is free to read; a private audit is not a free diagnostic.
Bandtastic is a public community strategy and is shown for educational audit demonstration only. Vetta does not claim ownership of the strategy. This static audit does not establish profitability, safety, or suitability and is not investment advice or a recommendation to trade.