catchfire.finance All digests

Daily network digest

Sunday, September 1, 2024

Directed information-flow networks for this session, covering Cryptocurrency. Each class is estimated independently over a trailing one-year window; the tables below rank the assets carrying the most influence within their own network.

Cryptocurrency

REPUSD was the most central name, carrying 4.74× the influence of an average node across 335 assets.

assets
335
Density
0.156
Clustering
0.365
Diameter
2

Most central

Ranked by PageRank within the session's directed network. “vs average” shows influence relative to a node with no advantage — 1.00× would mean perfectly ordinary.

Symbol#PageRank vs averageReceives fromSends toStrongest outgoing
REPUSD 1 0.01416 4.74× 51.2% 21.9% WAXLUSD, RBCUSD, ASMUSD, SEIEUR, BATUSD
KRLUSD 2 0.01332 4.46× 50.6% 11.4% EWTUSD, SHPINGUSD, RADUSD, PAXUSD, TRXUSD
XDCUSD 3 0.01000 3.35× 44.6% 14.4% GNOUSD, ZECUSD, IMXUSD, TUSD, SUPERUSD
SHPINGUSD 4 0.00968 3.24× 45.2% 15.6% NYMUSD, SOLUSD, XLMUSD, DOTUSD, MUSEUSD
BLZUSD 5 0.00895 3.00× 42.8% 29.3% DIMOUSD, PUNDIXUSD, FILUSD, JSTUSD, XRDUSD
TRBUSD 6 0.00863 2.89× 49.7% 12.3% BTRSTUSD, BONDUSD, AMPUSD, PHAUSD, REPUSD
TSDUSD 7 0.00821 2.75× 45.2% 36.5% POLISUSD, CGLDUSD, SDNUSD, MINAUSD, EGLDUSD
MULTIUSD 8 0.00813 2.72× 38.0% 29.3% CLVUSD, BANDUSD, OSMOUSD, DOGEUSD, TLOSUSD
XRPBTC 9 0.00780 2.61× 36.8% 23.4% CVCUSD, RADUSD, AURORAUSD, ACAUSD, KEYUSD
PNKUSD 10 0.00780 2.61× 38.3% 37.7% SUKUUSD, PERPUSD, XCNUSD, ENSUSD, SCRTUSD

Rising influence vs 2024-08-31

SymbolPageRank PreviousChange
KRLUSD 0.013320.00517 +0.00816
TSDUSD 0.008210.00203 +0.00618
XDCUSD 0.010000.00394 +0.00606
REPUSD 0.014160.00831 +0.00585
TRBUSD 0.008630.00340 +0.00523

Falling influence

SymbolPageRank PreviousChange
TEERUSD 0.002700.01187 -0.00917
DESOUSD 0.002950.01060 -0.00765
NODLUSD 0.002770.00901 -0.00624
EULUSD 0.002650.00881 -0.00617
CTXUSD 0.003810.00975 -0.00594

How these numbers are produced

Transfer entropy measures how much knowing one asset's past reduces uncertainty about another's future, beyond what that asset's own past already explains. Unlike correlation it is directional, so an edge from A to B means A's moves carry information about B's next move — not merely that they move together.

Each night we estimate transfer entropy across three universes — roughly 1,500 US common stocks, 350 cryptocurrencies and 550 currency pairs — over a trailing one-year window, conditioning on a broad market series so that shared market-wide movement is discounted rather than counted as a link. Edges surviving the significance filter form that session's directed network, from which the centrality figures on this page are computed. Coverage differs by class: crypto reaches back to January 2024, equities to August 2025, so older digests may cover fewer classes.

PageRank ranks an asset by how much influence flows into it from other influential assets. In-degree and out-degree are the share of the network sending information to it and receiving information from it. Betweenness captures how often an asset sits on the shortest path between two others.

The volatility watch score is a model estimate of the probability that an asset sees elevated volatility in the coming sessions. It is a research output, not a forecast of price direction, and nothing on this page is investment advice.