Follow-Through Day

Key takeaways · 5

1. 01A Follow-Through Day is a 1.7%+ close on a major US index, on volume above the prior session, between day 4 and roughly day 12 of a rally attempt off a market low.
2. 02William O'Neil designed it to detect institutional commitment, the only kind of buying powerful enough to start a new uptrend. His claim: no major bull market since the 1900s has begun without one.
3. 03The signal is invalidated if the index later closes below the FTD day's intraday low. That close is the framework's stop-out level.
4. 04About half of historical FTDs lead to sustained advances (March 2009 and April 2020 are textbook winners). The other half fail (March 2008 is the canonical loss). The signal tilts the odds, it does not guarantee the outcome.
5. 05TickerStance encodes the FTD as a binary signal in the Breadth subscore. It flips to 1 when an FTD fires and back to 0 when the rally invalidates.

Any candidate session that misses one of these three is not an FTD.

First, the index must close at least 1.7% above the prior day's close. The threshold has changed over the years. O'Neil used 1% in the 1960s and 1970s. IBD raised it to 2% briefly around 2000, then settled on 1.7% in the mid-2000s as average daily volatility increased and 1% sessions became too common to mean anything. TickerStance uses the current 1.7% threshold.

Second, total volume on the FTD session must be greater than the prior session's volume. This is the institutional-commitment test. Mutual funds, pensions, and other large pools of money cannot enter quietly; their footprint shows up as a volume spike. A 1.7% up day on lighter volume is short-covering or a mechanical rebound, not the regime change the framework is looking for.

Third, the session must fall on day four or later of the current rally attempt. Days one through three are not eligible, even if a 2% session prints. O'Neil considered the first three days too noisy to count as confirmation. The window extends to roughly day twelve before the signal weakens; quantitative work by Quantifiable Edges in 2008 found that FTDs after day ten have lower forward base rates than those between days four and seven.

The major US indexes that count are the S&P 500, the Nasdaq Composite, and the Dow Jones Industrial Average. An FTD on any one of them confirms the rally attempt; the others do not have to confirm. In practice the Nasdaq tends to fire FTDs first because it carries higher-beta names that lead off bottoms.

Three rules in one picture. The green session qualifies as an FTD because it gains 1.7% or more, prints volume above the prior bar, and lands inside the day 4 to day 7 sweet spot of the rally attempt.

The piece beginners get wrong is the day count. It does not start when the news improves or when the chart looks pretty. It starts on a specific session that satisfies a specific rule.

The trigger: the index has been declining. On some session, the index trades below its prior intraday low (an "undercut") and then closes higher than where it opened. That session is day one of a rally attempt. The next trading day is day two, and so on. The undercut is not strictly required by every reading of the rule; some practitioners count day one from the first up close after a clean bottoming pattern. IBD's strict version uses the undercut-and-reverse session.

A worked example. In March 2009, the S&P 500 made a fresh closing low on March 6 at 683.38. On March 9 the index closed at 676.53, but its intraday low of 666.79 was a fresh undercut and the close finished near the high of the day. March 9 was day one. March 10 was day two (a +6.4% snapback). March 11 was day three (consolidation). March 12 was day four, and the +4.07% close that day on heavier volume was the FTD.

The count resets if the index undercuts the rally-attempt low at any point. A new low cancels the prior count and a fresh attempt begins on the next undercut-and-reverse session. This sounds finicky, and it is. Bottoms produce a lot of false attempts before the real one prints, which is why a single FTD is more credible than the third or fourth attempt during the same decline.

The cleanest FTD in modern memory printed on March 12, 2009, four days into the rally attempt off the Global Financial Crisis low.

The setup. The S&P 500 had fallen 57.7% from its October 2007 high of 1,565 to the March 9, 2009 intraday low of 666.79. Lehman, Bear Stearns, AIG, the auto bailout, the bank stress tests: every major shoe had dropped. Bear-market rallies through the prior twelve months had printed and failed repeatedly. Sentiment was at one of the most negative readings in postwar US history.

What happened. March 9 produced an undercut-and-reverse close that started a new rally attempt. March 10 ripped 6.4% on volume in what felt like a relief bounce. March 11 consolidated. March 12, on day four, the S&P 500 closed up 4.07% at 750.74 on volume above the prior session. That was the textbook FTD.

What followed. The S&P 500 ran from 750 to 1,200 by year-end 2009, a 60% rally in nine months. The bull market continued essentially uninterrupted until February 2020, eleven years and roughly a 400% total return later. Anyone who recognized the FTD on the day it printed had a green light to start building long exposure with a clearly defined invalidation level: March 12's intraday low. That is exactly the entry-with-stop the framework was designed to produce.

March 12, 2009: day four of the rally attempt, the S&P 500 closes up 4.07% on heavier volume than the day before. The bull market that started here ran for the next eleven years.

The pandemic crash delivered a different kind of FTD: faster, with a less crisp setup, and almost daring traders not to take it.

The setup. The S&P 500 fell from its February 19 high of 3,393 to the March 23 close of 2,237, a 34% decline in 23 trading days. The speed of the drop had no real precedent outside 1929 and 1987. The Federal Reserve had already cut rates to zero, restarted quantitative easing, and announced unlimited Treasury and mortgage-bond purchases. Treasury yields collapsed, credit spreads blew out, the VIX printed all-time intraday highs above 80. Most desks were trading from kitchen tables.

What happened. March 23 was the close low. March 24 ripped 9.4% on heavy volume, almost certainly day one of the rally attempt. March 25 and 26 added more. The decisive test was April 2: a +2.28% close on volume above April 1, on day eight of the rally attempt. That was the FTD by IBD's strict criteria. (Some retrospective analysts cite April 6, the +7.0% session on day ten, as a stronger confirmation; either way the rally was confirmed in the first half of April.)

What followed. The S&P 500 ran from 2,527 to 3,756 by year-end 2020, recovering the entire crash and printing fresh all-time highs by August. The 2020 case is a useful study because the FTD printed during what felt like the worst possible regime, with pandemic deaths still climbing, the economy shut down, and earnings forecasts withdrawn. It worked anyway. The whole point of a mechanical signal is to override the emotional read.

April 2, 2020: the S&P 500 closes up 2.28% on heavier volume than April 1, on day eight of the rally attempt. The COVID bottom was three weeks behind, the all-time high four months ahead.

FTDs fail. The honest base rate, drawn from quantitative work on the modern era, is roughly half: about one in two FTDs goes on to lead a sustained advance, and about half are invalidated within weeks or months. The classic failure case in living memory is March 2008.

The setup. The S&P 500 had been declining since its October 2007 high. Bear Stearns nearly collapsed over the weekend of March 14, 2008, and was force-married to JPMorgan in a deal brokered by the Federal Reserve. The Fed cut rates 75 basis points on March 18 in an emergency response. The market rallied hard on the news.

What happened. March 17 produced an undercut-and-reverse close near the Bear Stearns weekend low at 1,276.60, starting day one of a new rally attempt. March 18 closed up 4.24% on heavy volume, looking like a textbook FTD. Most IBD readers at the time treated it as a confirmed signal.

What went wrong. The rally lasted six weeks. The S&P 500 rose to 1,426 by mid-May, then began rolling over. By July 7, the index closed at 1,252 — below the March 17 rally low of 1,276.60 — invalidating the FTD signal under the framework's own rule. The bear continued for another nine months and bottomed at 666.79 in March 2009. Anyone who treated the March 2008 FTD as a license to load up long without a stop lost a great deal of money over the following twelve months. Anyone who used the framework correctly cut exposure when the signal invalidated and saved most of their capital.

This is the most important lesson the FTD teaches. The signal is informational, not deterministic. It tilts the odds, it does not guarantee the outcome. The discipline that pays is the stop-out rule, not the entry.

March 18, 2008: a textbook FTD on the day after the Bear Stearns rescue. The rally held for six weeks, then rolled over. By July the index had closed below the rally low and the signal was invalidated.

Counting from the wrong day. Day one of a rally attempt is the first up close after an undercut-and-reverse session, not the day the news improved, not the day the chart looked nice. Get the count wrong and your "day four" becomes IBD's day six or eight, which changes whether the candidate session qualifies at all.

Treating an FTD as a buy signal for everything. An FTD is a regime confirmation. It does not tell you which stocks to buy; only that the regime now favors leadership over defense. The next step is hunting actual leaders inside leading sectors with real chart structure, not pressing every breakout in sight.

Ignoring the volume rule. A 1.7% up day on volume below the prior session is not an FTD. It is a relief bounce. The volume increase is the institutional-commitment test, and skipping it is what produces most of the false positives beginners report.

Forgetting the invalidation level. The single most useful piece of information the framework gives you, after the entry, is the FTD day's intraday low. If the index closes below that level, the signal is dead and the regime view should reset to neutral or defensive. Without the stop, the entry has no defined risk and the math falls apart.

Trading every FTD with the same size. A first FTD off a months-long decline (2009, 2020) has stronger forward base rates than the third or fourth FTD during a slow grinding bear (most of 2008, parts of 2022). Practitioners size up on first attempts after extended declines and size down on later attempts that follow a recent failed FTD.

For a retail swing trader, the practical workflow is short.

Watch for the rally attempt. After a sustained decline of 5% or more, look for an undercut-and-reverse session. That session begins the count.

Wait through days one to three. Even if a 2% session prints early, it does not qualify under the rule. Use those days to update your watchlist of names that survived the decline best, since those are tomorrow's leaders if the rally confirms.

Verify the criteria on day four through twelve. A close 1.7% or more above the prior day, on volume above the prior day, on the S&P 500, Nasdaq Composite, or Dow. Any one index counts.

Mark the invalidation level. Note the intraday low of the FTD session. That is your regime-level stop-out. If the index closes below it, the regime read flips back to defensive.

Start building, do not finish. The FTD is the green light to start scaling into leaders that have already broken out or are forming the right side of a base. The scaling is gradual because false signals exist. Full position size only comes after the rally has held for two or three weeks past the FTD without invalidating.

Cut exposure on invalidation. If the index closes below the FTD low, exit the trades that were initiated on the signal and step back to cash or hedges until the next rally attempt prints a fresh FTD.

TickerStance encodes the FTD as a binary signal called follow_through_day inside the Breadth subscore. It flips to 1 on the day an FTD fires on either the S&P 500 or the Nasdaq Composite, and back to 0 when the rally attempt invalidates (a close below the FTD day's intraday low) or when the next rally attempt begins.

The signal carries 15% weight inside Breadth, paired with distribution_days at 15% as the entry-and-exit pair for the breadth side of the score. Breadth itself is 30% of the headline Stance score, so an FTD lifts Breadth and the headline read together. The full math is on the methodology page.

In practice, on an FTD day the dashboard's regime narrative will note the signal explicitly. On invalidation, the signal flips back, Breadth eases, and the narrative will note the failure. The point is not to make you trade the signal in isolation. The point is to make sure that when you check the regime read on a morning that feels confusing, the most informative event of a market bottom is already in the number.

Frequently asked questions

Keep reading

• Market RegimeWhy market direction is the single biggest risk-management lever a swing trader has, and where the FTD fits in.
• MethodologyWhere follow_through_day sits inside the Breadth subscore, and how the rest of the Stance score composes.
• Stock LeadersOnce an FTD has fired, this is how you find the names that benefit most from the new regime.