Inferring aggregate flow signal from privacy-mode metadata — without decrypting any payload.
IT'S ALPHA
Pure research & development. This page exists so we can test whether the on-chain data patterns are coherent enough to derive insight at the boundaries of what's observable. Methodology is disclosed, error bounds are surfaced where possible, but accuracy is not guaranteed and we may rewrite, replace, or remove anything you see here at any time. Do not trade or make decisions on this output without independent verification.
What is alpha: read the full alpha protocol →
Private txs in sample
0
Total txs observed
0
Privacy share
—
Session age
1 min
Privacy Heatmap (7d × 24h)
1 · Privacy Long/Short Bias
inferred from action-type ratio in encrypted setSample: 0 private txs. Need ≥50 for meaningful inference.
▸Methodology
We can't see the direction of an encrypted order, but we can see its action type — PlaceOrder vs CancelOrder. A surge in private place-orders during a price breakout, vs balanced ratios during chop, gives a noisy but real signal of intent. v1 reports the raw place/cancel ratio + a 95% bootstrap CI over the in-session sample. v2 will weight by adjacent-block price drift.
2 · Price Couplingv2
privacy bursts × ASTER price drift (5-min buckets)Reading the signal. Each row is a 5-min bucket. Left bar = privacy tx count in that window (violet, saturates at 30). Right bar = ASTER price drift over that same window (green = up, red = down). When privacy bursts consistently co-occur with price drops, correlation goes negative → bearish-leaning. Positive corr ≥ 0.4 → bullish. Within ±0.15 → no real coupling. v2.1 will widen the sample to 24h server-side; right now we only see what the page session has accumulated.
▸Methodology
v1 inferred bias from the action-type ratio alone — a coarse signal. v2 adds a price-coupling layer: we bucket privacy txs into 5-min windows, compute the ASTER price drift over each window, and report the Pearson correlation between privacy tx count and signed drift. Strong negative correlation (privacy bursts coincide with price drops) is the textbook signature of insider/MM short pressure leaking through encryption metadata. Price source is the AsterDEX public ticker, polled every 10 s and buffered 30 min in localStorage.
3 · Whale Temporal Profilesv2
per-wallet bias signature · click to expand▸Methodology
v1 ranked privacy whales by frequency only. v2 builds a full bias signature per wallet: action mix (place/cancel ratio), size class (from cipher byte-length), first-/last-seen timestamps, and a combined bias score that blends ratio bias (60% weight) with price-coupling bias (40% weight, when sample allows). Confidence is reported per wallet: low < 10 txs, medium ≥ 10 txs, high ≥ 20 txs and ≥ 5 price-overlap points. Click any whale to see the detailed signature.
4 · Correlated Clusters
co-occurrence in same block▸Methodology
Pairs of addresses appearing in the same block multiple times in a row are statistically unlikely to be unrelated traders. Could be: same entity behind two wallets, MM cluster, copy-trading. v1 uses block-level co-occurrence; v2 will use millisecond timing for tighter signal.
Where the data comes from: bias, clusters, top wallets, and action mix are pulled from a rolling server-side sample (the /api/collect cron records cipher-tx metadata — hash, sender, action, block, timestamp; no decrypted payload, privacy is preserved). Widgets 2 and 3 (price coupling, whale temporal) still use the in-tab WebSocket buffer because they need the live price feed alongside, so give them ≥ 10 min of session age.