Launched this week
ChainVol
Desktop-grade crypto order flow, on your phone
20 followers
Desktop-grade crypto order flow, on your phone
20 followers
Crypto order-flow analytics on mobile: footprint charts, CVD, liquidity heatmaps, four spoofing-detection algorithms, and price-vs-order-flow divergence tracked from forming to confirmed. Real-time from Coinbase and Kraken.
how does the spoofing detection actually perform when a venue is just naturally messy and not trying to spoof anything
@nehir240588 Messy isn't the same as manipulative, and the detectors care about the pattern, not the mess. What gets flagged is an order that shows up, leans on price, then pulls instead of filling. Genuine churn usually doesn't have that shape. Where it does get murky is thin books — that's what the confidence score is for, so you can weigh it rather than take a yes/no.
How does the spoofing detection actually hold up during low-liquidity hours on Coinbase, since the algorithms might flag genuine large orders as manipulation when the book is thin?
@selahattinkucx Size alone doesn't trigger it — the detectors look at behaviour. An order appears, price approaches, it cancels or repositions instead of getting filled. A big resting bid that actually gets hit isn't flagged.
But you're right that thin books are noisier. Market makers quoting and pulling look structurally similar to spoofing when there's less depth. That's partly why every anomaly comes with a confidence score and a timeline — so you can discount the weak ones rather than trust a binary label.
If you catch it being clearly wrong, send it my way. That's the useful feedback.
How does the spoofing detection actually work in real time without false flagging legit large limit orders, and do you plan to add Binance or Bybit feeds beyond Coinbase and Kraken?
@aleyna3zum Behaviour, not size. A legit large order rests and gets filled — not flagged. What triggers it is an order that appears, leans on price, then cancels instead of trading. Each flag carries a confidence score and a timeline, so borderline cases are yours to weigh. Coinbase and Kraken for now, no plans beyond that. Each venue needs its own calibration, so I'd rather keep two solid than four shaky. If demand shows up, I'll revisit.
How does the spoofing detection handle legitimate large orders being pulled quickly, since that seems like a fine line to walk without flagging too many false positives?
@eyda9c3c That's the actual hard case, and I won't pretend it's solved. A market maker pulling size fast and a spoofer pulling size fast look similar — I can't read intent.
What the detectors look at is the shape around it: did it lean on price first, did it reposition and lean again, was any of it filled. That catches the deliberate patterns and lets most ordinary quoting through — but some legitimate pulls will get flagged, honestly.
Which is why nothing is binary. Every flag has a confidence score and a timeline, so you weigh it rather than trust it.
Footprint charts on mobile actually feel usable here, which I didn’t expect. The spoofing detection flagged something real on a thin order book I was watching earlier.
@ferhatozon2647 Thanks — footprint on mobile took the most reworking, getting per-level numbers readable without mush wasn't obvious.
Thin books are the hard case, so good that it held up. If you spot it flagging something that looks wrong, send it my way.
real-time footprint charts on mobile finally make sense, the divergence tool caught a fakeout on SOL this morning that saved me a bad entry. would love dark mode for the heatmap
@cihan8qim Thanks — footprint was the piece I spent the most time on, glad it reads right on a phone.
On the divergence though, I'd be careful: it describes what order flow is doing, it doesn't predict the move. Sometimes it lines up with a fakeout, sometimes it doesn't — I don't want to sell it as an entry filter, because I've tested that and it isn't one.
On dark mode — the app is already dark throughout, so I want to make sure I understand you. Do you mean the heatmap colours themselves are too bright against the background, or something else?