This post covers the Coinbase Exchange daily candle session that closed on 2026-09-05, across the 20 USD-quoted crypto pairs tracked by Skia Paper. The session that follows is complete; the figures below are final for that date.

The regime label for the session is broad-advance: every one of the 20 pairs closed higher, the median session change was +2.90%, and the spread between the strongest and weakest mover was 9.55 percentage points. The table rendered alongside this post carries the full per-pair breakdown.

What the Session Did

Breadth on 2026-09-05 was absolute: all 20 of 20 tracked pairs closed higher. That is the widest possible reading on the breadth measure this desk publishes, and it is a relatively rare condition in the dataset. The median session change of +2.90% indicates the typical pair moved by a meaningful margin rather than edging fractionally into positive territory.

Dispersion — the gap between the strongest and weakest session performers — was 9.55 percentage points. OP led at +9.74% and BTC trailed at +0.20%. The gap is wide enough to matter: pairs at the top of the distribution moved roughly 49 times as far as the pair at the bottom within the same session. The table shows the full ranking; the narrative here focuses on what the shape of that distribution reveals about the session's character.

Range position adds a second dimension. BTC, ETH, and NEAR all closed with range positions at or above 86% of their respective 30-day ranges, meaning their closes sat near the top of recent experience even though their session gains were modest. OP, by contrast, gained +9.74% on the session yet sat only 63% up its 30-day range, reflecting that its prior 30-day context was wide enough to absorb a large single-day move without reaching a new ceiling. ICP closed at 92% of its 30-day range after a +8.09% session gain. STX carried the most striking 30-day figure in the table at +107.53% over that horizon, closing at 81% of its range.

The combination of universal positive breadth, a median above +2.50%, and a dispersion figure near 10 percentage points describes a session where participation was broad but the magnitude of gains was unevenly distributed — concentrated in a subset of smaller-cap and layer-1 tokens rather than spread evenly across the field.

Regime and What Distinguishes It

The broad-advance label is applied when the breadth count is high and the median session change is positive by a margin that exceeds noise. A session with 20 of 20 pairs higher and a +2.90% median sits unambiguously in that category. The neighbouring regimes — mixed and compressed — would require either a split breadth count or a median close to zero; neither applies here.

A dispersed session shares the wide spread seen here (9.55 pp) but does not require uniform directionality; what makes this session broad-advance rather than dispersed is that the dispersion occurred entirely on the positive side of zero. Every pair contributed to the advance, even if the contributions ranged from negligible (+0.20% for BTC) to substantial (+9.74% for OP).

The 7-day and 30-day columns in the table provide context for where each pair arrived at the close, but they describe prior price history, not the session's cause. NEAR's 7-day figure of +17.34% and LINK's 30-day figure of +47.18% indicate those pairs had already moved significantly before the 2026-09-05 session opened; the session close extended or held those moves. That context is descriptive of the state of each pair's recent range, nothing more.

Cost of Involvement in Conditions Like These

Coinbase Exchange publishes a tiered maker-taker fee schedule. At the lowest volume tier, taker fees run up to 0.60% per side, producing a round-trip cost of up to 1.20% before spread and slippage are considered. Maker fees at the same tier are lower, but a maker order is not guaranteed to fill.

Against the session moves recorded on 2026-09-05, the cost picture varies sharply by pair. For OP at +9.74% or ICP at +8.09%, a 1.20% round-trip represents a fraction of the session's gross move. For BTC at +0.20% or XRP at +0.96%, a round-trip at standard taker rates would consume the entire session move and more, leaving a negative net figure before any spread or slippage is added. The median move of +2.90% sits above a standard round-trip cost, but only by a margin that spread and slippage can close quickly in thinner conditions.

Spread widens during periods of rapid price movement, and slippage scales with order size relative to available liquidity. A broad-advance session with large moves in smaller tokens — OP, FET, APT — typically corresponds to elevated spread in those pairs precisely when the moves are occurring. The cost of a round-trip is therefore not fixed; it tends to be highest in the pairs where the session move was largest.

Strategies that clear round-trip costs on sessions like this generally do so through one of three routes: holding period long enough that the fee is a small fraction of cumulative move, position size large enough that maker rebates become meaningful, or execution frequency high enough that per-trade costs are offset by volume-tier reductions. A single session is not a holding period in the first sense; it is a description of what one day's candle contained.

What One Session Does Not Establish

A single session with 20 of 20 pairs closing higher is a complete description of one day. It is not evidence of a trend, a regime shift, or a repeating pattern. The broad-advance label describes what the 2026-09-05 candles showed; it carries no implication about the session that follows.

Breadth at its maximum is, structurally, a condition that cannot be exceeded. A session reading of 20 of 20 has nowhere to go but sideways or lower on the breadth measure alone. That is a mathematical property of the scale, not a signal about direction.

Range position figures above 80% — present for BTC, ETH, LINK, NEAR, ICP, and IMX — indicate those pairs closed near the top of their recent 30-day windows. That is a description of where price sat relative to recent history on this one date. It does not indicate whether the range will expand, contract, or hold.

The dispersion figure of 9.55 percentage points describes the spread within this session. Past sessions with similar dispersion have been followed by sessions of every character in the dataset. The classification is a label for what happened, and the failure mode of treating it as predictive is well-documented in the literature on regime-based analysis: the label changes when the session changes, and the change is only visible after the fact.

The 2026-09-05 session closed with its broadest possible breadth reading and a median gain of +2.90%, while a 9.55 percentage-point gap between OP and BTC illustrated that uniform directionality and uniform magnitude are different things.

Note: This explains how a mechanism works. It is educational and is not financial, investment or trading advice, not a recommendation, and not a signal service. Any figure describes a session that has already closed and says nothing about what happens next. Trading cryptoassets risks losing money, including all of it.