The alert fired. I didn't buy.

An entry signal fired on a name I've been watching for weeks. I didn't buy it.

That's the whole story of the last two weeks. No trades, no drama, a portfolio that finished almost exactly where it started — down -0.59% on the adjusted curve. The only real decision I had to make was whether to act on the signal. I decided to wait. This post is me showing my work on that call, because waiting is a decision too — and this one has already cost me a little.

The name that triggered

On July 21, two of my TradingView strategies fired on ACN — Accenture. In my system that's the actual buy signal. The screener is a watchlist; it tells me a business is worth owning. The TradingView strategies tell me the chart is set up to enter. When both line up, it's time to make a call.

So I made one. I pulled the full Stock Rover report, ran it through my whole framework, and got a second opinion on top. Here's what I found.

Accenture is cheap. Not a-little-cheap — it's down roughly half from its high, trading around 11 times earnings against a decade where it normally fetched close to 28.¹ It passes eleven of my twelve fundamental gates. It has net cash, not net debt. Its free cash flow has topped its net income every year for five years, which is about as clean as earnings quality gets. On the numbers alone, this is a great business on sale.

So why is it down 50%? One question the market can't answer yet: does AI eat Accenture's business? A huge chunk of what Accenture sells is people writing code, testing software, and running processes by the hour. If AI does that work instead, the billable-hours model shrinks. On top of that, U.S. government contracts — a real slice of revenue — are getting cut. Bookings slipped. Guidance came down.²

Why I didn't buy

Here's the part that kept me out. My momentum score — NMS, on a 0-to-6 scale — reads 4 on Accenture. Not 5, not 6. A 4 means "warming up," not "ready." And the reason it's only a 4 is exactly the thing I'd want resolved before I buy: bookings are shrinking and analysts are trimming their estimates. The business isn't falling apart, but it isn't turning up yet either.

That's the discipline. Cheap is not the signal. Cheap plus a business that's stopped getting worse is the signal. Right now Accenture is cheap and still drifting the wrong way. My rule exists to keep me out of exactly this trap — the stock that looks like a bargain on every value screen while its earnings quietly keep sliding.

What waiting has already cost

I'll be straight about it. Since the strategies fired, Accenture has drifted up — about 5% higher per share. The discount I was looking at has narrowed with it: the margin of safety compressed from around 41% to 27% in two weeks. If it keeps running, the clean entry I wanted may never show up, and I'll have watched a bargain walk away.

I've done this before. I've left gains on the table by waiting for confirmation that never came. It stings every time. But the alternative — buying a falling knife because it's cheap — is the more expensive mistake over a career. The next real read comes on September 24, when Accenture reports full-year results and gives its first look at next year. If bookings turn back up, I buy with the momentum confirming the value. If they slip again, I'm glad I waited. Either way, I'd rather be a few percent late on a right decision than on time for a wrong one.

The tech names gave it all back

Last month I wrote about META running 16% while I wasn't watching. This fortnight it handed most of that back — down -11.1%, the worst name in the book. GOOG was right behind at -10.1%, and AMZN gave up -5.4%. The high-flyers that carried the portfolio in early July were the drag in late July. Same names, opposite direction, two weeks apart. That's the case for not chasing them when the RSI is hot.

The surprise on the other side was PYPL — up +21.2%, the best name of the period by a wide margin. PayPal has been on my "review this thesis" list longer than anything else and I still haven't done the work. A 21% move is the market telling me to stop putting it off. INTU (+7.8%) and CF (+7.0%) also pulled their weight — CF doing exactly the zig-when-tech-zags job I bought it for.

Gold caught its breath

After three brutal weeks, the gold cluster steadied. AEM was down just -1.2%, NEM off -2.2%, and Barrick actually finished green. No repeat of the double-digit bleed from last month. The fundamental case never changed; the prices just stopped falling. Nothing to do here but hold.

How the portfolio did

Flat. Down -0.59% on the adjusted curve — price-only, with the background auto-invest drips stripped out so they don't read as real gains.

The shape is a shrug: a small rise into mid-July, a roll-over on July 23, a partial bounce to close. The tech drawdown and the PayPal-plus-energy-plus-Intuit gains roughly cancelled. When the winners and losers offset like that, the return isn't the story — the decisions are. And the only decision that mattered was the one I didn't make.

The screener kept running

While I sat on my hands, the screen churned as usual. Six new names showed up this cycle — UBER the highest-ranked of them. Two names I'd flagged as buy-ready last month, AU and KGC, quietly fell off the shortlist as the gold cluster reshuffled. And ADBE is now in its ninth-plus week near the top of Bedrock with a full thesis I still haven't written. It's the longest-standing gap in my own reasoning record, and it's moving up the list.

What's next

Three things. Write the Accenture thesis into a proper dossier so the September 24 call is a check against a plan, not a fresh scramble. Do the PayPal review I've deferred for two months — a 21% week earns it. And close the Adobe gap before another entry window opens and shuts while I'm still not ready.

The alert fired. I didn't buy. Ask me again after September 24 whether that was patience or a miss.

Footnotes

1. Valuation history from the Stock Rover research report (July 21, 2026): current P/E near 11 versus a 9-year range topping 28, and EV/EBITDA near 7 versus roughly 18 over the prior four years.
2. Accenture plc — Third-Quarter Fiscal 2026 Results (Accenture Investor Relations, 2026). New bookings down 2-3% year over year, full-year revenue growth guidance trimmed, with a stated impact from the U.S. federal business.

— Mark