Thursday US Catch-Up Screen: FTNT Broke Support, LHX Is Disqualified, and There’s No Clean Buy Yet

Research cutoff: Thursday, August 20, 2026 at roughly 12:30 BST / 07:30 ET.
U.S. price cutoff: Wednesday, August 19 regular-session close.
Horizon: roughly the next 1–15 U.S. trading sessions.
Premarket moves do not count as confirmation in this screen.

Today is one of those days when the best catch-up trade is no trade.

Fortinet still has exactly the kind of medium-term peer gap I want to find: it is materially behind Palo Alto Networks and CrowdStrike over 10 and 20 sessions, while its own business momentum remains strong. But Wednesday changed the setup in a way I cannot ignore: FTNT closed at $152.85, below the $154–$155 invalidation zone that this screen had already defined before the move happened.

At the same time, L3Harris has a fresh company-specific governance shock after its CEO left following a board investigation. Even with guidance reaffirmed, that makes its relative weakness much harder to treat as a clean mean-reversion opportunity.

So I am not going to fill the list just to produce three names.

Today's verdict

Bottom line: there is no clean catch-up buy today.

Relative-performance table

For FTNT and FFIV, the benchmark is the equal-weight average return of Palo Alto Networks and CrowdStrike. They are direct cybersecurity competitors, so the relationship is economically meaningful rather than a random chart comparison.

For XOM, I use Chevron as the simplest integrated-oil comparison. Both are exposed to the same crude-price, refining-margin and geopolitical drivers.

Negative numbers mean the candidate lagged the benchmark over that window.

FTNT is still the most interesting shape in the table. The 1-, 3- and 5-session windows are positive while the 10- and 20-session windows remain meaningfully negative. In isolation, that is the classic transition pattern I want.

But relative strength cannot override a broken absolute-price level.

Fortinet: the peer gap is real, but the setup failed its own rule

Wednesday was a useful stress test for this strategy.

FTNT fell 3.27% to $152.85. Palo Alto Networks fell about 3.8% and CrowdStrike about 5.3%, so Fortinet still technically outperformed the peer basket on the day. That is why the short-window relative numbers in the table remain positive.

But calling that bullish would be misleading.

The catch-up screen had already said that a sustained close below roughly $154–$155 would invalidate the setup. Wednesday delivered exactly that. If I lower the line now simply because FTNT still looks relatively cheap, the rule becomes meaningless.

The business itself is not the problem. Fortinet's latest quarter produced $2.05 billion of revenue, up 26% year over year; billings rose 33%; and adjusted EPS increased 41%. Management also raised parts of its 2026 outlook, and the company continues expanding its AI-security stack, including the recently announced Virtue AI acquisition.

That is why FTNT stays on the research list. It just does not stay on the active buy list.

What would make FTNT interesting again?

I am tightening the rebuild sequence rather than immediately recycling the old setup:

• No entry below $154–$155. That zone now acts as the first repair level, not an entry area.
• First rebuild: reclaim roughly $154–$155 and hold above it.
• Better repair: reclaim roughly $158–$160.
• Old catch-up trigger: roughly $165.75, but only after the lower structure has already been rebuilt.
• Confirmation: at least two regular sessions above the rebuilt level while FTNT is no worse than neutral versus the PANW/CRWD basket over five sessions.

That sounds stricter than yesterday because it is. The stock broke the line. It has to earn its way back into the screen.

F5: the old gap is huge, but the price action is still wrong

FFIV is another good example of why the 20-session number cannot dominate the decision.

F5 remains more than 11 percentage points behind the PANW/CRWD basket over 20 sessions. But Wednesday closed at $382.63, well below the roughly $400 invalidation level from the earlier setup.

Its one- and three-session relative numbers are positive largely because the whole cybersecurity group sold off hard. The five-session relative number is still negative, and the absolute chart has not stabilized.

F5's underlying business is not obviously broken. Its latest reported quarter showed 11% total revenue growth and 19% product revenue growth, and management raised its fiscal-year outlook. That keeps the company fundamentally interesting.

But the trade is not interesting yet.

For FFIV to return to the active list, I want to see:

1. a reclaim of roughly $400;
2. at least two sessions holding above that level;
3. five-session relative performance versus the cyber basket back to neutral or positive;
4. only then does the old $420–$424 catch-up trigger become relevant again.

L3Harris: this is exactly the kind of laggard I should reject

LHX would normally be tempting. The company reported a record backlog, solid revenue growth and raised its 2026 outlook earlier this earnings season.

But on August 17, L3Harris announced that CEO Christopher Kubasik had stepped down following a board investigation into conduct that violated company policy. The company said the matter was unrelated to financial reporting, operations or customer relationships and reaffirmed its 2026 outlook.

That distinction matters fundamentally, but it does not remove the stock-specific catalyst.

A catch-up trade works best when a company is lagging despite broadly shared drivers. A fresh CEO-governance event gives investors a specific reason to discount LHX relative to defense peers. That means the divergence is contaminated.

So I am not setting a price trigger for LHX today. I want the governance story to settle before treating its peer gap as a mean-reversion signal again.

Energy: oil is moving, but the stocks are not diverging enough

Oil is the obvious macro driver this morning. Brent and WTI pushed to roughly three-week highs as supply concerns around the U.S.-Iran conflict intensified.

That made integrated oil names worth screening for a laggard.

The problem is that Exxon and Chevron are basically moving together. XOM trails CVX by only about 1.5–1.7 percentage points over the five- and ten-session windows, while their 20-session performance is essentially aligned.

That is not enough dispersion for this strategy. I would rather pass than manufacture a relationship that is not actually stretched.

The same logic applies to Occidental today. Its fundamentals have improved sharply — including strong free cash flow and debt reduction in Q2 — but the recent relative gap against the energy group is still too inconsistent to qualify as a clean multi-window catch-up setup.

Seven-session walk-forward check: one rule gets tougher

I reran the current logic over the prior seven U.S. trading sessions using only information that would have been available at each point in time.

The most useful lesson is that the absolute-price gate has been doing its job.

• FTNT's earlier one-day relative bounce did not complete the reclaim-and-hold sequence, and the stock reversed afterward.
• The short-window transition appeared again on August 18, but the stock still had not repaired the absolute structure.
• On August 19, FTNT then closed through the predefined invalidation.
• FFIV can look superficially better on one- and three-session relative performance when the peer basket falls faster, even while F5 itself keeps making lower closes.

I am therefore adding a post-invalidation quarantine rule:

Once a candidate closes through its predefined invalidation, positive 1- or 3-session relative performance is not enough to put it back on the active list. It must first reclaim the invalidation zone, hold above it for two regular sessions, and show at least neutral five-session relative performance.

This is not about making the model more conservative after every loss. It is about stopping a broken setup from repeatedly reappearing just because its peers happen to fall harder for a day or two.

Analyst, news and community read

Fortinet: the fundamental analyst backdrop remains constructive, with recent research pointing to accelerating billings, strong execution and AI-security demand. The counterargument is valuation: several valuation-focused analyses and investor discussions have questioned how much future growth is already priced into cybersecurity leaders. That tension is useful context, because it means the stock can have a strong business and still deserve a period of multiple compression.

Community: recent retail discussion around FTNT and the cybersecurity group is still broadly optimistic about long-run demand, but valuation and timing concerns are common. That is not a trading signal by itself; it simply argues against assuming every dip must mean-revert immediately.

L3Harris: current discussion is dominated by the leadership investigation rather than by defense-sector relative value. That is another reason to keep LHX outside a clean peer-catch-up framework for now.

Today's macro risk

The broader tape is not especially friendly to forcing a mean-reversion trade.

U.S. futures were muted Thursday morning while Treasury yields resumed climbing, with the 10-year yield around 4.67% and the 30-year around 5.22%. Walmart earnings, weekly jobless claims and Federal Reserve commentary are all near-term macro inputs.

Oil is also up sharply again on supply concerns tied to the U.S.-Iran conflict. That can help energy cash flows while simultaneously keeping inflation and bond-yield pressure alive.

For growth stocks, that combination makes premarket rallies less trustworthy. I continue to count regular-session closes, not overnight or premarket spikes, as confirmation.

Bottom line

Today the screen is doing something more valuable than producing a ticker: it is refusing to move the goalposts.

FTNT still has a real medium-term peer gap and strong fundamentals, but it broke the $154–$155 invalidation. FFIV has a huge old relative gap but an even weaker absolute trend. LHX has fresh company-specific governance risk. Energy names are moving too closely together to create a useful mismatch.

So there is no clean U.S. catch-up buy today.

FTNT remains the first name I will check again. But it now needs to rebuild the broken support zone before the old $165.75 catch-up trigger matters. Until that happens, the right signal is not “cheap versus peers.” It is wait.

Sources

Price history / relative-performance calculations

• Investing.com — Fortinet historical data
• Investing.com — Palo Alto Networks historical data
• Investing.com — CrowdStrike historical data
• Investing.com — F5 historical data
• Investing.com — Exxon Mobil historical data
• Investing.com — Chevron historical data

Fortinet / cybersecurity

• MarketWatch — Fortinet falls 3.27% on Aug. 19
• Fortinet Investor Relations — quarterly earnings
• Fortinet — Virtue AI acquisition
• Zacks — Fortinet outlook and valuation discussion
• Reddit — cybersecurity valuation discussion

L3Harris

• Reuters — L3Harris names new CEO after Kubasik steps down, Aug. 17, 2026
• L3Harris — Q2 2026 results

Energy / macro

• Reuters — Oil rises to three-week high on supply worries, Aug. 20, 2026
• Occidental — Q2 2026 results
• Reuters — U.S. futures muted as bond yields resume uptrend, Aug. 20, 2026

This article is for educational and informational purposes only. It is not financial advice, investment research, or a recommendation to buy or sell any security. Relative relationships can break permanently rather than mean-revert.