--- name: "um-analysis" description: "The UM framework (@0xuberM): non-derivative test, kingmaker question, best-proxy method, casino-vs-tables, liquidity-as-unlock. Run on every DD, token screen, contract-address paste or \"is this early\" question, alongside okay-lets-ride and the normal on-chain screen." --- # UM analysis Ben, 19 Sep 2026: "When I say we do due diligence, you will then apply the 'let's run it' approach as well as the UM approach, and this is how you evaluate coins going forward." ## THE STANDING DD SHAPE Every DD is now three layers, in this order: 1. **FACTS.** The normal on-chain screen. Deployer, funder, bundle and sniper check, holder reconstruction, LP and lock state, taxes, impostor-contract check, seller ledger. Nothing here changes. Facts must be right before any framework is applied. 2. **OKAY LET'S RIDE** (Coulou, @okay_lets_ride). Moneyness, pairing count, penetration, prior-cycle deflation, holder tiers, drawdown quality, meta lifecycle, business taxonomy, stack layer. 3. **UM** (@0xuberM). Everything below. Layers 2 and 3 answer "does this have potential, where is it in its life, is it aligned with the narrative". Layer 1 answers "are the facts real". Never let a framework overwrite a fact. When the two frameworks disagree, SAY SO and give both, with the one measurement that would settle it. They already disagree on ORBIO and on reflections. Source: the Flow State ep.6 X broadcast, 19 Sep 2026, transcribed in full. Detail and quotes live in memory at /areas/uberm-framework.md. Read that file before running this. --- ## 1. THE NON-DERIVATIVE TEST. Run this first, it gates everything else. > "Everything I see is just a derivative of something that's already been done before... eventually something is going to appear that has elements to it that you never really saw before. Whenever that happens, you have to put all your attention in it." Ask: **what about this has never been done before?** - Name the closest three comparables and say exactly what this does that they do not. - **If it comps cleanly, it is a derivative.** Say so. A derivative can still be a trade, but it is a normal-sized, invalidation-bearing trade, never a conviction bet. - If it genuinely does not comp, say that plainly, because that is the rare signal: "there's nothing to compare it to." - Be honest about frequency. These appear a few times a year. UM, asked what he liked that day after looking every day: "there's really nothing right now." A DD that finds a non-derivative every week is miscalibrated. **Consequence for sizing and exits.** On a true non-derivative UM runs NO invalidation and is willing to dust the position, because the upside is uncapped. On a derivative he keeps normal invalidations. State which one this is, because it decides whether a stop makes sense. ## 2. THE KINGMAKER QUESTION. The single most load-bearing variable. Not "is this good" but **"do the people who own this venue want this trade to work?"** Evidence to go and find, not assume: - What has the chain, launchpad or foundation's leadership said ON THE RECORD about this category? Quote it with a date. - Do their team accounts post about it? Do they engage the coin or its founder? - Has the venue king-made a comparable coin before? - Does this coin's mechanism DIRECTLY advance the venue owner's stated goal? (long.xyz deepening tokenized-stock pools is the model case: Robinhood needs that liquidity before it can lend against tokenized stocks.) **Then compress the whole bear case into ONE sentence, and it should almost always be about the venue owner's attention.** UM's entire STONK bear case was: "the only bear case is if Solana doesn't care about it." A DD that cannot name the kingmaker, or that names one with no dated evidence, is not finished. ## 3. BEST PROXY, NOT BEST COIN. The eight-step AI replay. The question is "what is the best PROXY for this narrative on chain", never "which coin is cheapest or has the nicest chart". Run these in order: 1. **Stated intent.** What has the venue owner said this venue is FOR? 2. **Proven appetite.** Which coins on this venue already ran under this theme? Name them with numbers. 3. **The primitive and its first mover.** What is the new mechanic, and who shipped it first? 4. **Rule out the rival variation.** Which competing design is structurally worse, and why? (UM ruled out reflections: CEX-listing friction and regulatory ambiguity.) 5. **First and highest.** Inside the ecosystem that spawned the primitive, which coin ran first and ran highest? That is usually the proxy. 6. **Mechanism check.** Does the contract actually do what the pitch says? Verify on chain, never from the site. 7. **Biggest possible pair.** Is it attached to the largest underlying available? AI is paired to the biggest stock in the world. 8. **State the conclusion in his form:** "X is the best proxy for on chain", or say plainly that it is not, and name what is. ## 4. CASINO OR TABLES Launchpads are "Ponzis disguised as token factories": fees recycle into the pad's own token. - On a high-volatility, low-liquidity pad ecosystem, **the pad token is the position**, not the coins it launches, because a very small group wins those lotteries. - **Exception: venues whose coins are built for long lifespans** (thick locks, anti-PVP, anti-sniping). long.xyz is the named example. There the coins themselves are ownable. - So every DD on a launched coin must state which regime it is in, and whether the better expression of the same view is simply the pad token. ## 5. THE TRIFECTA > "Every chain sort of has this group of coins that intertwine together and they collectively go up." Precedents: SOL/LUNA/AVAX; Hyperliquid/Lighter/Aster; now PONS/STONK/PAID. If the candidate sits in a cluster, name the other legs, say whether the venue's operators benefit from all of them working, and check whether a different leg is the cleaner bet. ## 6. LIQUIDITY IS AN UNLOCK, NOT A WARNING This inverts the usual instinct and must be applied deliberately. - Deep liquidity lets a whole class of non-trenchers, who do not care about being early, buy the narrative LATE and WITH SIZE. That capital cannot participate in illiquid coins at all. - Test it: can the pool absorb a large market sell without breaking structure? AI took a single $2mm sell and just chopped above prior resistance. - Ask whether the coin's liquidity growth serves the venue owner's own roadmap. If yes, that is a durable reason for the venue to keep it alive. - Never write "too liquid, the easy money has gone" without testing this. ## 7. THE PAIRING QUESTION For any coin paired to an underlying asset: - Does that asset already have a community OUTSIDE crypto? - How strong is it? - How much money does it have? The whole stocker thesis is outside capital arriving, so a pair with no external community is a weak pair however good the mechanics. ## 8. HIS SECTOR MAP, as of 19 Sep 2026. Re-check before quoting. - Bull market. Stock-paired memes are the theme for at least the first half of the cycle, with "a lot of juice left". - Three places to be: the token factories (PONS, STONK, PAID); the long.xyz ecosystem coins; infrastructure that benefits from tokenized-equity activity, plus AI and privacy. - His strongest claim: "whatever is paired to Nvidia on long will be the Nvidia coin on chain." Every stock's definitive on-chain counterpart ends up on long.xyz. - End state: long.xyz becomes one of the largest holders of tokenized equities on Robinhood, which opens the door to institutional capital. - Rotation: every cycle eventually pivots back to AI in some form. Expect that AFTER the stocker mania. - Scale: Robinhood Chain does roughly $1bn to $1.4bn a day; Dogecoin peaked at $60bn to $70bn a day. Very early. - Vlad's incentive: chain revenue at about 90% margins, millions a day within two months, could dwarf the rest of Robinhood. ## HANDLING HIM AS A SOURCE - He is long AI, the long.xyz ecosystem, PONS, STONK and PAID. He talks his book. Use the method, verify every conclusion on chain. - He is honest about his misses (sold STONK at $2mm, sold AI on the way up), which makes his self-reported entries more usable than most. - By his own account he is early and executes badly. **Copy his entry logic. Do not copy his exits.** ## PRE-DELIVERY GATE A DD is not finished until it states, or says explicitly that it could not measure: 1. Derivative or non-derivative, with the three closest comps named 2. Whether an invalidation applies, which follows from 1 3. The kingmaker, with dated on-the-record evidence 4. The one-sentence bear case 5. Whether this is the best proxy for the narrative, or what is 6. Casino or tables, and whether the pad token is the better expression 7. The trifecta legs, if it sits in one 8. The liquidity absorption test 9. The pairing question, if it is paired to something 10. Where okay-lets-ride and UM disagree, and the measurement that would settle it