Bounce Candidates After Today's Pullback — Reading the Tape the Rogue Alpha Way

I am not going to give you a ticker list and tell you to buy in the morning. That is not how I trade and it is not how this works.

What I can do is walk you through the exact screen I am running tonight on the names that got flushed today, the five filters that have to clear before something even shows up on my watch list, and the rules I use to decide whether a bounce is real or a bull trap on the open.

Run this on your own list. If a name clears all five filters and confirms in the first 30 minutes Wednesday, you have an actual setup. If it does not, you have a stock that went down. There is a difference.

Days that look like Tuesday — KOSPI down 10%, memory complex liquidated, momentum names hit with no single-name news — are positioning unwinds, not fundamental repricings. The selling is forced, not informed. Forced sellers do not check what they sell on the way out; they sell what they own. That is why the names that fall hardest in a cascade are often the names with the most crowded long book, not the worst businesses.

When the forced selling exhausts — usually within one to three sessions — the highest-quality names in the wreckage are the first to bid. The garbage stays down. The framework below is about separating the two before the open, not after.

Filter 1 — Earnings, not story. The name has to have a real, growing earnings base. Not a narrative, not a roadmap, not a forward-multiple argument. If the company would not be profitable without 'adjustments,' it does not make the screen. Forced selling clears positioning; it does not bail out unprofitable companies.

Filter 2 — Heavy-volume distribution, not a drift down. The today's-session move has to be on a volume multiple of the 30-day average. A 4% drop on light volume is a slow leak. A 4% drop on three-times average volume is a flush — and flushes are what bounce. No volume, no setup.

Filter 3 — Sympathy tag, not company-specific news. If the name went down because its sector got liquidated, that is a bounce candidate. If the name went down because the CEO resigned, guidance was cut, or a short report dropped, that is a falling knife. Read the headlines on every name before it makes the screen. No exceptions.

Filter 4 — A defined intraday low to risk against. The name has to have a clean intraday low you can put a hard stop under. If the chart printed a single capitulation candle and stabilized, you have a level. If it closed on the lows with no shelf, the seller is not done and you do not have a level yet. Wait.

Filter 5 — The sector or index it lives in stabilized into the close. Bounces inside a still-falling group are short-cover spikes, not real reversals. The cleanest setups are names that flushed early, then went sideways in the back half while the rest of the group caught down to them. That dispersion is the tell.

Within the names that clear all five filters, the highest-probability bounces are sympathy trades — names that got hit because of what they are tagged with, not because of what they are. After a memory-complex liquidation, that means equipment names that were dragged down with the pure-plays (LRCX, AMAT, KLAC profile). After an AI-momentum unwind, that means second-derivative AI names that have actual cash flows, not the narrative leaders.

The reason sympathy trades bounce harder: the forced seller did not care that the name was tagged incorrectly. They sold the basket. When the basket stabilizes, the mis-tagged names are the first to be re-bid because they were the first to be wrongly sold.

The rule: a sympathy bounce candidate must have a fundamental reason it should not have been sold as hard as it was. Write that reason down in one sentence before the open. If you cannot write it in one sentence, you do not understand the trade well enough to take it.

Every bounce setup has the same failure mode: a strong open, a fade through the first 30 minutes, and a break of the prior-day low by 10:30 a.m. ET. When that happens, the bounce thesis is dead and the next move is usually a fast retest of the next major support below.

The kill-switch rule: if the name trades through the prior-day low after the first 15 minutes of trade, the position is closed at market. No averaging down. No 'one more hour.' The setup was wrong, the seller was not done, and the only job is to be flat before the next leg.

Half of why bounce trades work over a sample size is that the losers are small. The other half is that the winners are sized to be meaningful when the setup actually triggers. Both halves matter equally.

Size each bounce position so that the distance from your entry to your hard stop equals a fixed percentage of your account — for me, somewhere between 25 and 75 basis points per name depending on conviction and how clean the level is. This is the only number that matters at entry. Not the upside target, not the percentage gain you are hoping for. The risk size.

If a name has a 2% stop distance and you are willing to risk 50 bps on it, your position is 25% of the account. If a name has a 5% stop distance and you are willing to risk 50 bps on it, your position is 10% of the account. The stop distance dictates the size. Period.

On a day after a forced-seller flush, I will typically run three to five bounce candidates rather than concentrate in one, because the cross-section gives me information. If four out of five trigger and one fails, the failure was company-specific. If one triggers and four fail, the bounce thesis was wrong and the cash is the right position.

Names from the memory and memory-equipment complex that closed in the middle or upper half of the intraday range, on heavy volume, with no company-specific bad news. Those are the cleanest sympathy candidates if the Korean session stabilizes overnight.

Energy and gold-miner names that bid quietly into the close while the broader tape was offered. Those are not bounce trades — those are rotation trades, and the framework for them is different (trend continuation, not mean reversion). I covered the rotation map in the prior piece; that is where most of the new capital actually goes after a day like today.

Software and second-derivative AI names that held the morning lows, traded sideways, and closed green from the midday low. A held morning low after a sector flush is one of the highest-probability single-name setups in the entire framework.

I am specifically not watching names that closed on the absolute lows of the day. A close on the lows tells you the marginal seller was not done. The bounce, if it comes, has to wait until that seller is actually out.

A pullback day does not create bounce candidates. It reveals them. The names that bounce on Wednesday were already the highest-quality names in their groups on Monday — the pullback just gave you a price you would not have gotten otherwise.

If you cannot articulate why a name was the best business in its group before the pullback, do not pretend the pullback made it one. Buy quality at a better price. Do not buy garbage because it is cheaper.

This article is a personal opinion piece by Guy Gentile and a description of a trading framework. It is not investment advice, a research report, an offer to buy or sell securities, or a recommendation of any specific stock. The author may at any time be long, short, or flat any of the names or sectors discussed, including through options or hedges, and may change those positions without notice. Trading involves substantial risk of loss. Do your own work and consult a licensed financial advisor before making investment decisions.