A crypto grade moved and the score did not. Twice, in opposite directions, on the same pair of our records.

Between our October 13 and our October 14, 2026 records, Kamino's score did not move. It read 65.6 on both. Its grade went from B to F, and it left our ranking from position 11. On the same pair of records, LayerZero's score also did not move, 54.4 on both, and its grade went the other way, from F to C, taking position 25.

Two assets, two letters, two numbers that stood completely still.

Our record attaches no cause to either move. That is not a dodge, it is the state of the record, and it is the honest place to start a piece about what a letter can tell you.

Our Friday piece applies the framework to a live incident. The four news domains this routine is permitted to read were each tried once on October 16, 2026, and each refused: coindesk.com, reuters.com and cftc.gov failed at the connection, and apnews.com answered HTTP 403. So there is no outside incident we read first hand.

We will not summarize somebody else's summary and present it as reporting. Our own first rule is that every claim carries its receipt. What we can show you in full, with dates, is our own record, so that is the incident: four dated records, three assets, and two letters that moved on numbers that did not.

Our framework scores six weighted dimensions: tokenomics and value accrual at 28 percent, on-chain usage and economics at 22 percent, security, decentralization and durability at 18 percent, ecosystem and moat at 12 percent, relative valuation at 12 percent, and market structure at 8 percent. The published bands are 80 and above for an A, 60 to just under 80 for a B, 40 to just under 60 for a C, and below 40 for an F.

On top of the weighted score sit nine pass or fail screens. A tripped screen places an asset in our F grade at any score, including a score high enough for a much better band. That is the published design, and it is why we always publish the grade our record carries and never work a letter out from a number.

Now put Kamino's October 14, 2026 letter against that. Its F sat at 65.6, which is inside our published B band. LayerZero's F on the October 13 record sat at 54.4, inside our published C band. Neither F is explained by its number.

Here is the part most ratings writing skips. A low score is not the only road into an F, and our record does not say which road a given asset took. We do not name the screen, because our record does not publish which screen applied to a given asset. We are not going to guess, and you should be suspicious of anyone who does.

What we can give you is the number, the letter, the position and the date on each. That is four facts per asset per record, and you can watch them move.

Kamino read F at 61.6 on our October 15, 2026 record, unranked. On our October 16, 2026 record it reads B at 62.8, at position 12. So its grade left the B band on a number that had not moved, spent two of our records out of the ranking, and came back to a B on a number 2.8 points below where it started.

LayerZero read C at 54.4 at position 25 on our October 15, 2026 record, and B at 60.4 at position 15 on our October 16 record, its score up 6.0 points.

The score-change record we read for October 15 to October 16, 2026 carries 68 changed assets, and its attribution status is "not published" for every one of them. That means no cause is attached to a single move in it, ours included. We report what moved, from what to what, and between which two dates. If you want to learn why an asset's score moved, log in to your free CryptoGrade account to learn more.

Gram, previously Toncoin, read C at 40.0 on our October 15, 2026 record and F at 37.4 on our October 16 record. Here both the number and the letter moved, which is the ordinary case and the one readers expect.

Its number now sits below 40, which is where our published bands place an F. Note what we are still not saying: our record does not tell us which road Gram took into that letter either, so we are not going to name one. A number below the band floor and a tripped screen are not distinguishable from the outside, and we are on the outside of our own record here in exactly the way you are.

Gram is excluded on our October 16, 2026 record, and we do not print a position for an excluded asset. An excluded asset is unranked, not last. Writing it in at the bottom of a list would be inventing a fact.

This is the quietest finding in the four records and the most useful one if you hold more than a handful of assets.

LayerZero's record names the assets it passed between October 15 and October 16, 2026: Decred from 15 to 16, BNB from 16 to 17, Polkadot from 13 to 18, yearn.finance from 17 to 20, Litecoin from 18 to 21, Jupiter from 23 to 25 and Filecoin from 24 to 26. Kamino, re-entering the ranking at 12, is named in both LayerZero's record and Polkadot's as having moved past them.

Six of those seven carried the identical score on both records: Decred 60.2, BNB 59.9, yearn.finance 58.0, Litecoin 56.9, Jupiter 55.2 and Filecoin 54.5. Each of them sits lower because two other assets rose. Nothing about them changed at all.

Polkadot is the exception, and it is the one to watch if you own it: its own number moved, 61.3 to 58.7, and its letter went from B to C across the same two records.

So a ranking position is a relative statement. It can move while your asset stands still, and it moved that way for six named assets inside one pair of our records. A letter is the stronger signal of the two, and even a letter needs its date attached.

This is our model's standing assessment of each asset, published in our record on October 16, 2026. It is an assessment, not an explanation of the moves above, and our record does not connect the two.

On Kamino, our model's thesis reads that its supplied fundamentals "point to a materially used DeFi protocol, with $1.39B TVL and $4.92M in 30-day fees supporting a strong usage profile," and that "the main offset is token design uncertainty: supply can still expand toward 10B and no direct token-holder cash flow is provided, limiting tokenomics quality." Its own conclusion is that the asset "reads as operationally strong but not fully proven on value accrual." Our usage note for it is just as careful: large TVL and real fees indicate substantial activity, but "no token-holder revenue linkage is provided, so this is strong protocol usage rather than maximum token-economics credit."

On LayerZero, our model's thesis reads that "TVL context and live protocol fees point to meaningful adoption, while the supply profile still leaves substantial FDV dilution risk and the supplied record does not show direct token-holder cash flow," and that security and durability "are not strongly evidenced beyond the existence of the v2 design." Our tokenomics note puts circulating supply at about 400M against a 1B maximum.

Two assets our record describes very differently, one a decentralized finance protocol built on Solana that combines lending, liquidity provision and leverage, the other a project our record associates with interoperability across several chains while saying outright that its available sources do not establish what its token does, read on the same six dimensions and landing 2.4 points apart on the same record. Both carry the same flagged weakness, which is that protocol activity is documented and token-holder cash flow is not. That comparison is the entire reason a single framework is worth running.

The letter is not decoration on top of the score. In our model it carries a consequence, and ours is published at the grade level.

Read from our record on October 16, 2026 at our authenticated tier: our model's maximum single position for an A-graded asset reads "≤60% (BTC-class) / ≤30% other", for a B-graded asset "≤10%", and for an F-graded asset "0%".

Those are our model's rules at a grade, on that date. Everybody at a grade gets the same number. They are not addressed to you, they are not worked out from anything you own, and they are not advice: no one at CryptoGrade is a registered investment adviser or broker-dealer. What they do tell you is that when Kamino's letter went from B to F and back to B across four of our records, the model's own ceiling for it went with the letter, not with the 2.8 point move in its score.

That is the answer to the question in the headline. A letter tells you which of our model's rules an asset falls under. A score tells you where it sits inside our weighted dimensions. They are different facts, they move independently, and on October 14, 2026 two assets proved it in opposite directions.

A grade is a reading of an evidence record on a date. Every figure above is stated against the record it was read from, and that is deliberate: it is also the only way you can check us. We have not told you why any of these moves happened, because our record does not carry a cause for them, and a ratings company that fills that gap with a plausible story has stopped being a ratings company.

Not every asset in our ranking carries a letter either. Where our evidence record is incomplete, our model publishes no grade rather than guessing one, and we will not write a letter next to an asset that does not have one. Kamino, LayerZero, Gram and Polkadot all do, which is why they are the four we named.

Alerts, at https://cryptograde.ai/alerts You pick the crypto assets you care about and ask to be told when our grade on one of them changes. A letter changing while the score stands still is exactly the event that is invisible if you are watching a price, and it happened to two assets across one pair of our records.

The account is free, it costs an email address, and nobody we grade pays us anything. CryptoGrade does not work commercially with the assets or projects it covers, and it does not sell data.

Read the full framework, including the weights, the bands and the screens, at https://cryptograde.ai/framework