A seasonality rabbit hole · Aug 2026
A post from X caught my eye: the best Septembers tend to show up above the 200-day average. Fun idea, and it got me curious — so I chased it across all twelve months of the S&P back to 1928. Good news: the filter totally works. Funny news: it makes September look like the sleepiest month on the calendar, not the firecracker.
First off — I really liked this chart from @HostileCharts. Of the 20 best Septembers since 1928, 16 happened with the S&P above its 200-day moving average and 15 had a positive year-to-date. The takeaway — that September’s mood depends on the weather it starts in, and the same setup can cook up both the best months and the worst — is a genuinely good instinct. Trend state matters, and most people never even look at it. So: cool idea. I wanted to play with it.
The natural next question is just: does this only work for September, or does it work everywhere? So I ran the exact same above/below-the-200-day split on all twelve months and let the data talk. Two fun things popped out. The filter earns its keep across the whole calendar — nice. And once you look at September through its own filter, it turns out to be less firecracker, more warm glass of milk.
Above 200-day · all months
Average monthly return when the S&P enters above its 200-day line. 62.8% of those months finished green.
Below 200-day · all months
Same, when it enters below the line. 53.1% green. The filter is real — a 5× gap in average return.
September · above the 200-day
September’s average even when it starts above the line. 50% green — a coin flip, the weakest of any month.
Let’s start with the win, because it’s a good one. Take all 1,183 months on record and sort them by one little bit of info: was the S&P above or below its 200-day average heading into the month? Above-the-line months average +0.89% and finish green 63% of the time. Below-the-line months average +0.18% and are basically a coin flip at 53%. One trend switch, roughly a five-to-one difference in what you can expect. That’s a lot of signal from a single line on a chart — the post was onto something real.
S&P 500 · 1928–2026 · months entering above their 200-day average
Quick fun fact that trips everybody up (me included, at first). Because markets go up more than they go down, the S&P just lives above its 200-day line most of the time — 67% of all months start there. So any “best months” list is already going to be two-thirds above the line before you even look at returns. Sixteen of twenty (80%) beats that, but not by as much as it looks. The real magic is in the averages from section 1, not in the top-20 headcount.
And here’s where it gets genuinely weird — the best part of the whole rabbit hole. Rank every month in history by return and grab the 20 biggest up-months ever. Only 7 of them (35%) were above the 200-day line! The all-time monster months are bear-market rallies — the face-ripping bounces of 1932, 1938, 2020. If “explosive” means the biggest single moves, explosiveness actually lives below the trend. September just never throws those parties, which is exactly why its own top-20 leans so far above the line.
the baseline, the biggest up-months, and the best Septembers
Line the twelve months up by how they act after you apply the post’s own above-the-line rule, and September plops down at the bottom next to October. It’s the only month that’s a straight coin flip in an uptrend (50% green), and its average return is basically a rounding error next to December’s +1.76% or November’s +1.68%. The one place September does have a personality? When it starts below the line it gets genuinely spicy — −3.99% average, only 32% green. So September’s real signature isn’t explosive upside in a good tape; it’s a bit of a temper once the trend has already cracked.
| Month | Above: avg | Above: % green | Below: avg | Below: % green |
|---|---|---|---|---|
| Jan | +1.46% | 66% | +0.71% | 57% |
| Feb | +0.33% | 61% | −1.13% | 30% |
| Mar | +0.80% | 61% | −0.10% | 59% |
| Apr | +0.85% | 68% | +2.09% | 59% |
| May | +1.11% | 67% | −1.67% | 49% |
| Jun | +1.26% | 64% | −0.16% | 42% |
| Jul | +1.37% | 61% | +2.40% | 59% |
| Aug | +0.37% | 57% | +1.38% | 63% |
| Sep | +0.07% | 50% | −3.99% | 32% |
| Oct | −0.36% | 55% | +2.11% | 67% |
| Nov | +1.68% | 67% | −0.45% | 52% |
| Dec | +1.76% | 78% | +0.33% | 62% |
| All | +0.89% | 63% | +0.18% | 53% |
The 200-day line is a real, useful filter — the post nailed that. Follow it one more step and September’s story flips from “secretly explosive” to “secretly boring, unless the trend’s already broken.”
Same idea, one rabbit hole deeper.Honestly? Not much about September itself — and that’s the fun conclusion. Seasonality this mild isn’t a strategy, and the post is totally right that a blanket “September scary, sell everything” is the lazy take. The thing we both land on is the good one: watch the trend, not the calendar page. A September that opens above the 200-day line is just an ordinary, sleepy month — hold through it. A September that opens below it (−4% average, red two times in three) is where you actually perk up. The month name was never the tell. That 200-day line the post reached for? That’s the tell.
$SPX), daily closes from Norgate Data, February 1928 through August 2026 — 1,183 completed months.Anyway — none of this makes the original post wrong, and that’s kind of the point. It grabbed the right variable (trend state, via the 200-day line) and asked a good question. Chasing that same variable across all twelve months just adds a fun coda: the explosiveness we all half-expect from September isn’t really there. It’s a quiet month that only gets interesting once the trend’s already broken. Thanks to @HostileCharts for the jumping-off point — that’s the good stuff about this corner of finance twitter.