There may be nothing better than the small luxury of a Friday evening in September, when it’s still warm but no longer hot, and the sky lights up with multiple pink hues. The work week is over, with a full reset waiting as I exit the office. If I can top it off by going to bed in freshly washed sheets, it's literal heaven.

This year's biggest stock losers are chasing that same reset feeling. A fifth of the 1,000 largest US stocks are down more than 15%, and the hope, for a lot of them, is that the bad news is over, waiting for a fresh start. But only some of them have actually earned that. 

The obvious part first. For the US stocks down more than 15% this year, you can probably guess the 2 clusters: enterprise software (names like Adobe, Autodesk, and HubSpot, down 15-50%) and, maybe more surprisingly, defense (names like Kratos, AeroVironment, and Northrop). The "SaaS-pocalypse" and the defense-trade-cooling-off stories have both been written about by many already.

But, of the 207 names that are down big, which ones does Wall Street think will keep growing earnings anyway, and does that make them safer or more exposed?

Well in 2002, accounting researchers Doug Skinner and Richard Sloan published a paper entitled: "Don't Let an Earnings Torpedo Sink Your Portfolio." Their finding: while growth stocks priced for a growth story that the market has bought into don't miss earnings any more often than value stocks do, when they do miss, the stock gets hit disproportionately harder (aka a torpedo). The miss isn't just a bad quarter, it's potential evidence that growth assumptions are off. Value stocks, priced for little, don't have that far to fall when they disappoint. This is why you might have heard me say: “expectations are everything”.

And that asymmetry, Skinner and Sloan argued, is part of why "growth" as a factor can underperform "value" at times. Like this year: the Russell 1000 Value Index has returned 19% vs the Russell 1000 Growth up 6% this year, through September 24th.

So I split this year's losers into 2 buckets:

Group 1, the torpedo bucket, in green in the chart below, is down double digits, but the Street still models positive earnings growth into 2027. Almost 70 names fit in this, including mostly names in software along with some from the defense sector. The stock prices have already voted against the story; the analyst models, at least in some cases, haven't caught up yet.

Group 2, the reset bucket, in black in the chart below, is down just as much or more. But the earnings estimates in each have already been cut. This includes over 120 names, mostly within the consumer sector. Here, the bad news is already in both the price and the models.

Looking through these names, one could naturally assume the reset bucket, the ones that already saw their earnings growth estimates go negative, would have the worst behind it and trade at a lower, more attractive multiple. But instead both buckets sit at almost the same median PEG ratio, around 1.25x. Taking the earnings estimate hits hasn't earned “reset” stocks a valuation break relative to the ones potentially still waiting on theirs. 

What I take from this: 

Some of the ‘torpedos” are perfectly good businesses caught in a real re-rating of what "software" is worth in an agentic-AI world, and estimates holding up might just mean the earnings power is genuinely there. But if a confirmed miss doesn't earn a cheaper multiple, I wouldn't assume the torpedo bucket's decline so far is the end of the story. Its estimate hasn't been tested yet. Per Skinner and Sloan, when it is, such as the next earnings report, estimates tend to move down to meet the price, not the other way around. 

What will settle this is the next earnings season. I’ll be watching guidance on names in both buckets. That's when a growth cut shows up, if one's coming, and where an "already-priced-in" case is confirmed.

Until then, I’ll be washing my sheets Thursday nights for my own Friday resets, even when September has passed.

We will not publish next week as we’ll be at HoodSummit 2026. Join us for live sessions on September 29 and 30.