Current vs Daily spread search

This guide explains the two price-spread search modes in the Arbitrage section — Current and Daily — and how to choose between them. You’ll learn what each one shows, how to read it, and when to use which: a one-shot trade right now versus a repeatable / bot strategy.

The Arbitrage section has five search modes, switched at the top of the page: Current, Daily, Funding, Order book and Ping-pong. The first two are about cross-exchange price spreads and are covered here; Funding is about earning the perpetual funding rate (own article); Order book is single-venue maker market-making (own article); Ping-pong is inventory-based oscillation trading without transfers (see the strategy overview in Strategies).

What does Current mode show?

Current mode shows the live cross-exchange spread for each coin at this moment: where it’s cheapest to buy, where it’s dearest to sell, and the profit of each trade variant (spot-spot, spot-perp, perp-perp) at your capital, after fees and order-book slippage. Prices and spreads refresh every ~30 seconds.

Use it when you want to act now: you see a spread that exists this second and can take a one-shot trade. The min/max spread filter narrows the list; the upper bound also trims suspiciously large spreads, which are more often ticker collisions than real arbitrage. A “Best options” (★) toggle in the sort bar narrows it to only the genuinely profitable cards — those with at least one variant in real net profit after fees, funding and order-book slippage (at your fee tier, on a 100-USDT reference size) — and ranks them by that net. It loads the order books for every card, so give it a moment. (Every mode has its own “Best options” toggle, each meaning “profitable” for that mode.) You can also type a specific coin into the “Check a coin” box to see all of its exchange pairs directly — even when the spread is below the filter.

• Buy / sell exchanges and the live spread — the nominal one (from last prices) plus the executable top-of-book spread.
• Profit for each variant at your capital, including the real spread after walking the order book.
• For the perp variants (spot-perp / perp-perp) — an extra “if no convergence” figure: what the same trade nets if the spread never converges, i.e. you unwind both legs on the opposite (unfavorable) book sides at unchanged prices, round-trip fees and funding included.
• Order-book depth, funding rates, and the networks available for the spot-spot transfer.

A card sometimes shows no perp variant at all. That is deliberate: a perp that is being delisted keeps quoting a stale price over an empty book, so any spread measured against it is fictitious — and the convergence check can’t see it, because convergence is measured on SPOT candles. Each perp leg is therefore health-checked (funding published? does the perp itself have 24h turnover? is its price still attached to spot?), and a leg that fails has its variants dropped rather than shown with an attractive number.

Spot legs get the same check, applied even earlier — before the exchange pair is picked, because a frozen quote looks like the cheapest venue by construction, which is exactly why it would end up on a card. A real example: XMR spot sat at 118.70 on one exchange (its price at delisting) while every other venue and all six XMR perps traded at 332 — and the exchange kept reporting 24h volume for it. A leg counts as dead when it has almost no turnover, or when its price is more than 25% away from the consensus of the other venues (perps get a vote in that consensus: they agree with each other far more reliably than thin spot books do). A 10–25% gap isn’t fatal, but it keeps the variant out of “Best options”. Note the scope: this asks “is this market real”, not “is it liquid enough” — a thin but live market stays, and its cost is shown to you by the order-book walk and the risk score.

A current spread is a snapshot — it can close before you fill, especially on a thin book. Check the depth and the executable spread, not just the headline number.

The perp variants’ displayed profit is priced on the spread converging to zero. When the last 24h never saw a ≤1% convergence and the no-convergence figure is negative, the variant gets a red “no convergence in 24h” chip — the green number relies on a convergence that isn’t happening. And when on top of that the coin can’t be transferred between the two exchanges (no common network), the whole card gets a “structural gap” banner: nothing can close such a spread, so treat it as untradeable rather than attractive.

What does Daily mode show?

Daily mode looks at the last 24 hours of a pair’s spread and counts convergences: how many times the spread diverged past a threshold and then came back (crossed zero). It is built from hourly snapshots for the liquid, curated set of coins; other pairs can be computed on demand.

The key number is convergences: how many times over 24 hours the spread diverged past the threshold and came back (crossed zero). You set the threshold (minimum spread) and the minimum number of convergences in the search form. A pair whose spread regularly opens and closes is a candidate for the arbitrage bot / mean-reversion: you enter when it diverges and exit when it converges — repeatedly, not once. The chart shows the signed spread with the zero line in the centre and your chosen threshold dashed; it opens on the last 24h and you can scroll left for up to 72h of history (the convergence count stays over the last 24h).

Use it to evaluate a pair for recurring trading: a big spread that never converges is useless (and risky — see the risk score), while a smaller spread that converges many times a day can be traded again and again. Daily has no live profit number, so its “Best options” toggle keeps only the reliably-reverting pairs — those that converged at least 3× at your threshold, a repeatable pattern rather than a one-off divergence.

• The number of convergences over the last 24h, for the threshold you set in the form.
• The signed-spread chart — opens on the last 24h, scrolls back to 72h (zero line centred, chosen threshold dashed).
• The same trade variants plus an arbitrage-bot forecast for the pair.

How are trades executed? (both modes)

Watched cards of BOTH modes — Current and Daily — carry the same execution actions: «Execute» and «Set trigger» for the spot-spot variant, «Open trade» for the spot-perp / perp-perp variants (a two-exchange delta-neutral position, auto-closed by the real exit basis), and «Launch bot» for repeatable mean-reversion campaigns. The Daily card is where the bot fits naturally: its convergence statistics tell you the trigger and the cycle count worth setting.

On a watched card the spot-spot variant has two execution buttons. «Execute» runs the trade now: the platform buys the coin on the cheap exchange, you move it to the dear exchange yourself (the exact network, its fee and ETA are shown before you confirm — withdrawals stay manual on purpose, so your API keys never need withdraw permissions), and once the coin arrives the platform sells it automatically (or, if you untick auto-sell, notifies you and shows a «Sell» button). «Set trigger» arms the same trade to fire later: it executes automatically once the EXECUTABLE spread — walked through the order books for your capital, net of fees — reaches your threshold; an unfired trigger expires after 7 days.

Demo mode is honest about the transfer: the coin “travels” for the network’s real ETA and the sell is priced off the live order book at arrival — so a demo result includes the mode’s main risk, the price moving while the coin is in transit. Live execution additionally requires trade-enabled API keys on BOTH exchanges and the live-trading opt-in in Settings.

What does «Launch bot» do?

When the pair has a perpetual on at least one leg, the “Arbitrage bot” block offers «Launch bot» — a mean-reversion campaign (demo for now): it enters a delta-neutral position each time the EXECUTABLE spread (walked through the order books for your capital) diverges past your trigger, exits when it converges to your target, and repeats — up to the cycle count you set or 7 days. It holds nothing between cycles, stops after a losing cycle, and both entry and exit are measured on executable prices, never the headline ticker. The same honesty now applies to every open spread position: the auto-close fires on the real exit basis — the price of actually unwinding both legs — not the nominal spread.

Mind the “book width” line under the forecast: the chart is built from last-trade prints, which bounce between bid and ask inside each book — on a wide/dead book the chart’s “peaks” can be pure bounce with no executable spread behind them. Since the trigger fires on the executable spread, the nominal spread must beat your threshold by roughly that width; when no 24h peak could have cleared it, the forecast is struck through and a warning explains that the bot would simply stay silent (that silence is the protection — a real entry on such books would lose money).

Current vs Daily — which should I use?

• One-shot trade, act now → Current.
• Repeatable / bot trading, see how often a spread actually closes → Daily.
• Earn the funding rate, delta-neutral (no bet on price) → Funding.
• Make the bid-ask spread inside one exchange as a maker → Order-book spread.

The two are complementary: find a recurring opportunity in Daily, then switch to Current to time the entry.

Active-exchange filter: in Settings → Active exchanges you can pin all search modes — Current, Daily, Funding, Order book and Ping-pong — to a chosen set of exchanges, so every search only returns pairs on the venues you actually trade on. Leave it empty to search across all exchanges.