Liquidity is the first threshold in bottom research: slippage, fake volume, and inability to sell

Why Liquidity Is the First Threshold

For small-cap coins at price bottoms, the biggest fear is not a wrong directional call, but being unable to buy when you want to buy and unable to sell when you want to sell. When order book depth is insufficient, even a not-so-large buy order can push the price up, causing slippage; conversely, if there are not enough counterparties when you want to exit, you may be forced to accept a lower price or even be unable to execute for a long time. Wash trading also creates false prosperity: the surface turnover is high, but the actually executable liquidity is far lower than the book numbers. Therefore, the hard filters treat insufficient or unstable liquidity and almost no trading as direct filter items.

For ordinary investors, the risks of slippage and being unable to sell are amplified in small-cap bottom targets. We only study Binance USDT spot precisely because these trading pairs are relatively standardized, but even so, liquidity must be gated separately. This filter is not because we are bearish on a particular coin, but to avoid mixing targets that cannot actually be traded into the research pool.

Distribution of Today's Failures to Pass Hard Filters

Figure 1 shows the path from the latest market-wide scan to the review council: 414 Binance USDT spot trading pairs, 267 within the market cap research scope, 68 passing all hard filters; then entering the review council in the past 24 hours, 19 sent to review, 16 approved. The scan numbers are a snapshot, not a cumulative total over the past 24 hours.

Among the reasons for not passing the hard filters, 168 have risk tags, 107 are outside the market cap research scope, 33 are not yet in a bottom area, 21 have insufficient or unstable liquidity, and 1 has almost no trading. Risk tags and out-of-scope market cap are the two categories with the most filtered counts, but liquidity-related filtering remains critical: it excludes targets that might have acceptable scores but whose real trading depth cannot support the strategy.

Potential Board and Review Council: Scores Do Not Replace Review

In the top 10 of the potential board scored by the rule engine, YGG has a potential score of 69 and is 91% below its peak, GALA has a potential score of 60 and is 90% below its peak, and POLYX has a potential score of 55 and is 74% below its peak (Figure 2). These scores come from the rule engine, not from AI screening. The below-its-peak decline is relative to the stage high on Binance daily candles and does not represent the all-time high.

In the past 24 hours in the review council, the stance distribution of the four reviewers is shown in Figure 3: Fundamentals analyst buy 2, hold 1, avoid 0; Skeptic buy 2, hold 1, avoid 0; Technicals buy 3, hold 0, avoid 0; Risk officer buy 16, hold 1, avoid 2. The review council conclusions are cross-given by AI, and the Risk officer avoided 2, indicating that even after passing hard filters and rule scoring, a small number of targets are still rejected from a risk perspective.

Public Signals and Sample Limitations

Among the publicly disclosed review-approved signals, POLYX received 4 approve votes with 4 reviewers participating; YGG received 3 approve votes with 4 reviewers participating; KAITO, ONE, SOXLB, BARD, TUT, and GALA each received 1 approve vote with 1 reviewer. These numbers only reflect the reviewers' approve votes and do not constitute any buy or sell advice.

The public track record of live trading has only 1 closed signal, a very small sample, so no conclusion can be drawn yet; in the last 7 days, closed 0 and profitable 0. Even though the single trade averaged a holding period of 4.4 days, average 9.89%, maximum drawdown 0.0%, these statistics come from only 1 sample and cannot represent future performance. At present, what deserves more attention is whether the filtering chain is robust, rather than the closed trade track record.

What This Means for Users

What users see is a very vertical funnel: from trading pairs to market cap scope, then to hard filters, and finally to the review council and AI stances. Liquidity-related filtering treats 'can buy in and can sell out' as a research premise, rather than considering it only after a signal appears. This makes the research pool smaller but improves the operability of the research.

The limitations are also clear: strict filtering may miss some targets whose liquidity is temporarily weak but improves later; the review council sample size is small and public signals are still insufficient; with only 1 closed signal, strategy stability cannot be verified. Therefore, we only output research observations, do not provide return promises, and do not advise readers to buy any coin based on this.