We Checked 46 Stocks Jim Cramer Told Viewers to Buy in August. 14 Have Beaten the S&P 500

Jim Cramer tells his Mad Money audience what to buy almost every weeknight, in monologues, interviews and the rapid-fire Lightning Round. We went back to August, wrote down a rule for what counts as a buy call before looking at a single stock price, and scored every one he made on the show that month. He closed the Aug. 5 episode with this:

From the close the next day through the Oct. 27 close, AMDAMD returned +14.5%, NvidiaNVDA −0.4% and IntelINTC −6.1%, against +0.4% for the S&P 500, including dividends. A few seconds earlier he had said, “you can buy Okta, you can buy Cloudflare too.” OktaOKTA returned +57.5%, the best result of the month, and CloudflareNET +15.2%. He made a related case about cybersecurity in October, on what AI can’t take from CrowdStrike.

How We Scored the Calls

We pulled all 21 episodes that aired from Aug. 3 to Aug. 31 from CNBC’s podcast feed, transcribed them by machine and checked the calls we counted against the audio. A call counted only if Cramer’s own words told viewers to buy or own one named stock right then, such as “I think it’s a buy,” “I would own it” or “I’m a buyer.” We left out holds, praise without a buy, buys that hinged on a pullback or a price, hypothetical examples in his teaching segments, and anything a guest or caller said. Calls he made on other CNBC shows or on X are not in the count.

That left 46 stocks. When he backed a stock more than once, we used the first call. Each pick is measured from the closing price on the next trading day after the show aired to the Oct. 27 close, including dividends, and compared with the SPDR S&P 500 ETF over the same days. A pick beat the market if it returned more.

The Scorecard

Of the 46 picks, 14 have beaten the S&P 500 so far. The median pick returned −5.2%, while the index returned a median +0.1% over the same windows, and 31 of the 46 lost money outright.

It was a flat stretch for the market. The S&P 500 closed at 7,709.96 on Aug. 6 and at 7,722.60 on Oct. 27, with a record close of 7,818.93 on Oct. 6 in between. That made the bar low, and most picks still fell short of it. The biggest misses were Carpenter TechnologyCRS, which he called “a buy” on Aug. 13 and which has lost 47.0%, CoreWeaveCRWV (−34.3%) and Sterling InfrastructureSTRL (−31.4%). In all, 14 picks trailed the index by 10 percentage points or more.

The winners included PalantirPLTR (+16.2%), AppleAAPL (+8.6%) and DellDELL (+7.6%), which he called an opportunity again in mid-October.

Cramer’s August buy calls, through the Oct. 27 close

When the Calls Came

The early calls held up best: 9 of the 18 stocks he backed in the first week of August beat the index, led by the chip and software names above. Of the 28 he backed after that, 5 did. Those first-week picks have had the most time to work, and the groups are small, so the split may reflect how chip and software stocks traded more than his timing.

Some of his loudest calls came late. On Aug. 25 he called RedditRDDT “such a buy,” and it has returned −12.2%. On Aug. 31 he told viewers to buy Viking HoldingsVIK, “buy, buy, buy the dip,” and the stock has returned +0.5%, behind the index’s +1.3%. NetflixNFLX, which he told a caller on Aug. 11 to buy a little more of, has returned −15.8% as the stock slid to a two-year low.

What This Test Can and Can’t Show

A few months is a short window, and Cramer often tells viewers to own a stock for years. Some of his best-known calls took that long: on Feb. 20, 2024, he said Nvidia was nowhere near its peak, and the stock had roughly tripled by Oct. 27, 2026, split-adjusted. A machine transcript can also miss a call, so the count could be off by a few names, and a looser reading of “buy” would add more.

The scorecard stays open. The rule can be rerun at any later close, and the stocks to watch are the 14 that trail the index by 10 points or more, since a few big recoveries would change the median faster than anything else.