Where the market and the fundamentals disagree

Across the whole universe, bigger tokens do score better. Median composite falls from 157 in the mega-cap band to 95 in the micro-cap band. Ranking the whole set at once therefore measures size, not mispricing. We tried it: the resulting list of bargains came out entirely below $100M with a median of $14.7M, while the list of expensive names had a median of $211.7M. That is arithmetic restating what everyone already knows.

Inside a single band it stops. Composite score and market capitalisation correlate about 0.03 among small caps and minus 0.11 among mid caps. Once the size bracket is fixed, being larger says almost nothing about scoring better. So each token is ranked only against others its own size, and the gap is converted to percentile points of its band, because a 109-member band allows a gap of 108 and a 28-member band allows 27.

A rank is only worth reading if every row is a live, distinct, researched asset. Three filters run before the ranking, and each was added because the unfiltered version published something wrong.

Protocols their own page reports as offline or wound down are removed, because a score written before the event does not describe what is there now. Retired tickers and wrapped duplicates are removed, because POL and MATIC resolve to one asset and were counted twice. And 35 tokens are held out for low confidence: their 25 variables sit mostly on 4 and 5, which is the pattern an analyst produces with no information rather than a measurement. Ranking those against researched names turns missing research into a buy signal.

Sorted by the widest gap. A token here ranks well on the 25 variables and is priced as though it does not. That is a disagreement to investigate, not a verdict: the market may be pricing something the framework does not measure, and often is.

The same arithmetic the other way. These rank far higher by market capitalisation than the framework ranks them, which usually means the market is paying for something the 25 variables do not capture. Attention is the most common answer.

Which crypto tokens are undervalued?

On this measure CoW Protocol (COW), Scroll (SCROLL), Flux (FLUX), Morpho (MORPHO), PoolTogether (POOL), Alpaca Finance (ALPACA) sit furthest above their market-cap rank on fundamentals. That is a statement about the gap between two rankings, not a price target.

Which crypto tokens are overvalued?

Ethereum Classic (ETC), Zilliqa (ZIL), Zcash (ZEC), SushiSwap (SUSHI), Popcat (POPCAT), Ravencoin (RVN) carry the widest gap the other way, ranking far higher by market capitalisation than by fundamentals.

How is the gap calculated?

Each token is ranked twice against the others in its own size band, once by composite score across the 25 variables and once by market capitalisation. The gap is the market-cap rank minus the fundamental rank, converted to percentile points of that band so the bands compare. A token ranking 2nd on fundamentals and 50th by size in a 56-member band carries +86.

Why does this gap mean anything?

Because inside a size band score and capitalisation do not move together. Across the whole universe larger tokens score higher, which is why ranking everything at once produces a list of small tokens and nothing else. Inside a band the correlation collapses to roughly zero, so a wide gap there is a real disagreement rather than an artefact of size.