Token profiles
A five-day-old token and a two-year-old token are not the same kind of object, and analysing them the same way produces confident nonsense on at least one of them.
So the analysis picks a profile, and the profile decides what matters.
🆕 Fresh
under 14 days old
Volume and buy/sell pressure
🌱 Early
14–60 days old
Trend plus momentum
🏛 Established
60 days and up
The full indicator suite
Unknown age is treated as Established — the conservative default, since it avoids applying memecoin assumptions to something that might be a blue chip.
The first three are age bands. The fourth is not: it is chosen by what the token is, and it wins over age. That distinction is the whole point of it — see below.
Why the emphasis differs
On a Fresh token, a mean-reversion indicator is measuring the mean of about four days of trading, most of which was a launch. That mean has no authority. What does have authority is whether volume is still growing and whether buys outnumber sells — is anyone actually still here.
On an Established token, the reverse. There's real history, so the oscillators mean something, the moving-average stack means something, and the trend has been tested. Volume trend becomes noise on a mature chart and is weighted accordingly.
Early sits between them, because a three-week-old token is neither.
A tokenized equity is a different question again. Robinhood minted its stock tokens months ago, but the companies behind them have traded for decades, so scoring them by pool age put Apple in Early — where the engine throws away the first day of candles as a launch and treats DEX swap counts as sentiment. No age threshold fixes that, because age was never the variable that mattered.
The equity profile keeps every price-derived indicator at its Established weight and nearly mutes the two that come from the DEX. A few hundred swaps a day is a rounding error against the billions the same share turns over on its home exchange, and on a thin listing it would otherwise let a handful of trades speak for Apple. Trend and the moving-average stack take that weight instead: they are what equity traders read, and unlike a swap count they are computed from the price arbitrage keeps honest.
The weighting is also asymmetric: on a young token, growing volume is genuine information while a quiet hour is mostly noise, so the same signal is allowed to help the score more than it's allowed to hurt it. That asymmetry runs through the whole table, in both directions depending on the signal.
Stripping launch candles
The other thing the profile controls is how much of the launch gets thrown away before anything is measured — a longer window for the youngest tokens, a shorter one for Early, nothing at all for Established or a tokenized equity.
Why: launch price action is a vertical line into a peak nobody will see again, followed by a cliff. Left in, it becomes the "resistance" every level is measured against, it drags every moving average, and it makes "−96% from ATH" look like a discount when it's just the shape of a launch.
A tokenized equity is exempt for a different reason than Established: it has no launch to strip. Its first on-chain candles are an ordinary trading day for a company an exchange was already pricing, so trimming them would delete real history and manufacture a spike that never happened. For the same reason the launch-spike flag below never fires on one.
Safety valve: if trimming would leave too little chart to analyse, the untrimmed frame is used instead. Better a noisy chart than no chart.
The launch-spike flag
Trimming fixes the analysis, but the all-time high is still an interesting fact — and a misleading one when it was set in the first hour.
So the trimmed chart's high is compared against the untrimmed one. If the real peak lives entirely inside the launch window, the ATH was an artifact, and the card flags it.
Practical read: when that flag is up, ignore every "% below ATH" number on the card. It's measuring against a candle that existed for ten minutes and never recurred.
Last updated