Hunter Killer
SEC/01 · ACCURACY SCORECARD ▸

The scorecard we couldn't
find anywhere else.

Coinglass and Hyblock publish no walk-forward accuracy scorecard for their liquidation levels or heatmaps, and we now track eleven head-to-head comparisons. Ours is public: every number carries its sample size and Wilson 95% confidence interval, including the symbols we switched off. This page states the gap factually, and shows the actual numbers.

All competitor characterizations on this page are sourced from their own public product pages and documentation, as accessed on the research date cited beside each quote (most recently July 2026). Their products change; if a cited characterization is no longer accurate, tell us and we will correct it. Not affiliated with, endorsed by, or sponsored by any platform named. No claim beyond what those materials show.

Our live scorecard, as of now

Loaded from the live accuracy ledger. Every figure is walk-forward, out-of-sample, with n and Wilson 95% CI. Auto-disabled symbols are counted and shown.

N/A global precision@N
all tracked symbols · weighted
N/A majors precision@N
BTC / ETH / SOL / XRP · weighted
N/A global Wilson 95% CI floor
conservative lower bound
N/A effective independent samples · de-overlapped ·
walk-forward measurement · the number nobody else prints
N/A symbols tracked
incl. auto-disabled shown honestly
N/A walk-forward span
out-of-sample measurement days

Loading live data from /proof/stats

What each platform publishes: factually

The table below reflects what each platform's own public product pages and documentation state. We make no claims beyond what those materials show.

Hunter Killer (us)

Walk-forward accuracy scorecard Published: /proof
Sample size (n) on every figure Yes, per symbol
Wilson / CI on accuracy figures 95% CI lower bound
Auto-disabled symbols shown Yes, by design
Per-symbol accuracy pages Yes: /coins/*
Feed reconciliation record Public on /proof
Accuracy model methodology Published

Coinglass

Walk-forward accuracy scorecard Not published
Sample size (n) on accuracy n/a, no accuracy data
Wilson / CI on accuracy n/a, no accuracy data
Self-description of the model "simulating liquidation events" (from their own FAQ)
Per-symbol accuracy pages Not published
Feed reconciliation record Not published
Accuracy model methodology Not published

Hyblock

Walk-forward accuracy scorecard Not published
Sample size (n) on accuracy n/a, no accuracy data
Wilson / CI on accuracy n/a, no accuracy data
Self-description of the model "estimate derived from OI… not actual order book" (from their disclaimer)
Per-symbol accuracy pages Not published
Feed reconciliation record Not published
Accuracy model methodology Not published

Competitor rows sourced from publicly accessible product pages and FAQs as researched 2026-07-11. If any product has since published accuracy data, please let us know and we will update this page promptly.

What the platforms say about their own data

These are verbatim or closely paraphrased from each platform's own published product pages and FAQs. We quote only what they say about themselves, not inferences.

"The data is generated by simulating liquidation events based on various leverage multiples and open interest data."
Coinglass FAQ / product documentation (publicly accessible, 2026-07-11 research)
"The data shown is an estimate derived from open interest data and is not an actual leaked order book."
Hyblock disclaimer / product documentation (publicly accessible, 2026-07-11 research)

Both platforms are honest that their liquidation maps are derived models, not direct market observations. The gap we note is not that they use models, so do we. The gap is that neither platform publishes a walk-forward record of how well their model performed: no precision@N, no n, no confidence interval, no per-symbol outcome tracking.

Our position on this is structural, not rhetorical: a track record is the one feature money cannot buy retroactively: only time makes it. Every month we operate, the ledger grows. Every month it grows, the gap widens. We began publishing outcomes before we had a sales page.

Every head-to-head, in depth

We publish a full, sourced comparison page for each platform traders ask about. The accuracy line on each card states only what the linked page supports from that platform's own public materials — nothing more.

Coinglass

Broad crypto-derivatives data portal: liquidation heatmaps, open interest, funding rates, long/short ratios and a wide market dashboard across many exchanges.

Walk-forward liquidation accuracy scorecard: not published in their public materials (research date on the linked page).

Full head-to-head →

Hyblock

Broad crypto analytics terminal: liquidation levels and heatmaps, order-book and liquidity metrics, long/short and retail-sentiment indicators.

Walk-forward liquidation accuracy scorecard: not published in their public materials (research date on the linked page).

Full head-to-head →

CoinAnk

Broad crypto-derivatives data and analytics portal: liquidation heatmaps and map, open interest, funding, order-book and market dashboards across many exchanges.

Walk-forward liquidation accuracy scorecard: not published in their public materials (research date on the linked page).

Full head-to-head →

Laevitas

Broad crypto-derivatives analytics platform: option chains, implied volatility and Greeks, funding, futures basis, order-book data, aimed at quantitative analysts.

Walk-forward liquidation accuracy scorecard: not published in their public materials (research date on the linked page).

Full head-to-head →

Decadent

Same category as us: a liquidation heatmap / liquidation map tool showing where leveraged positions would be force-closed.

Walk-forward liquidation accuracy scorecard: not published in the public materials we could review (research date on the linked page).

Full head-to-head →

TradingView liquidation indicators

Community Pine Script overlays on the TradingView charting platform that estimate liquidation levels from price and assumed leverage tiers.

Walk-forward accuracy scorecard for those estimates: not published in their public materials (research date on the linked page).

Full head-to-head →

The Kingfisher

Purpose-built liquidation heatmaps visualizing where clustered leverage sits, delivered via web interface and Telegram.

Same category, different proof posture: the comparison is about what is measured and published — our walk-forward scorecard is public.

Full head-to-head →

Aggr

Real-time multi-exchange trade-tape aggregator: what the tape is doing right now, in one stream.

Different primary category (live tape vs liquidation map) — the comparison is about focus; the walk-forward scorecard is ours.

Full head-to-head →

Coinalyze

Broad futures-analytics dashboard: liquidation data, open interest, funding rates, multi-exchange aggregation, with a free tier.

Different primary category (wide dashboard vs focused map) — the comparison is about focus; the walk-forward scorecard is ours.

Full head-to-head →

Velo

Institutional-grade market-data terminal: clean, exportable series for funding rates, basis, open interest, and options data.

Different primary category (data terminal vs liquidation map) — the comparison is about focus; the walk-forward scorecard is ours.

Full head-to-head →

TensorCharts

Order-book and market-microstructure charting: order-book heatmaps, footprint charts, volume bubbles, liquidation overlays on price.

Different primary category (order-book charting vs liquidation positioning) — the comparison is about focus; the walk-forward scorecard is ours.

Full head-to-head →

Category descriptions are from the linked comparison pages, each sourced from the platform's own public materials as of the research date shown on that page. If a platform has since published a walk-forward accuracy scorecard, tell us and we will update the card promptly.

Why accuracy grading matters in practice

When a liquidation level is drawn on a chart with no accuracy provenance, a trader has no basis to weight it against other signals. It looks like a prediction. It might be a tool landing 2-in-10 or 7-in-10. Without the outcome record, the number on the chart carries the same visual weight either way.

Our approach is different by design: every magnet band on the Hunter Killer map carries its own measured reach rate (n, %, horizon). Below our minimum threshold, the band is suppressed from the chart automatically. A map with fewer, graded bands is more useful than a map with many unverified ones.

Symbols that drop below threshold are auto-disabled, and that state is shown honestly on the per-coin pages and the full proof ledger. We gain no advantage from hiding failures: the moat is the ledger, not the appearance of one.

Per-symbol accuracy pages

Each tracked coin has its own graded accuracy page: walk-forward precision@N, n, Wilson CI, and a delayed liquidation snapshot. Graded symbols are marked in green; auto-disabled symbols are shown with their honest state.

See the full scorecard

The complete walk-forward accuracy ledger, every symbol, every horizon, every auto-disabled case shown, is public on /proof.

All figures on this page are walk-forward, out-of-sample, not backtested on historical data with look-ahead · n and Wilson 95% CI printed on every accuracy figure · auto-disabled symbols counted and disclosed · no comparison-to-chance framing anywhere · no predicted-profit claims · competitor characterizations from their own public materials only · this page last checked for accuracy 2026-08-09 · full methodology · technical specification · verifiable evidence locker · engine track record