1 trillion and 10 trillion are not "both big companies"

Your brain reads numbers as ratios, not absolute values. The horizontal intraparietal sulcus (HIPS), deep in the parietal lobe, handles approximate quantity, and according to Dehaene (1997) it responds to the ratio between two numbers rather than to how big they are. Comparing 10 and 20 lights up the brain almost identically to comparing 1 million and 2 million. So 1 trillion and 10 trillion, a clean tenfold apart, never feel as cleanly split as 1 and 10. Both get crammed into a single "trillions" slot.

This is logarithmic compression, better known as the Weber-Fechner law: the bigger the stimulus, the duller our read on the same difference. In the world of prices and sums, that dullness turns straight into judgment errors. NVIDIA's market cap sits around $3.3 trillion and Tesla's around $800 billion, yet because both register as "giant tech," plenty of people peg the gap at roughly 2x instead of 4x. In reality, NVIDIA alone is four Teslas stacked together.

Why you should read headline figures by orders of magnitude

The news hurls big sums at you nonstop: "Company invests 5 trillion," "12 trillion for the acquisition." Without a sense of magnitude, 5 trillion and 12 trillion congeal into one blob of "a huge amount of money." But 12 trillion is more than double 5 trillion, and that gap alone equals the annual revenue of dozens of mid-sized firms. Even inside the single phrase "large company," gaps of tens of times are routine. The distance between the No. 1 and No. 50 company by market cap, or between Apple and a bottom-ranked name in the S&P 500, isn't one digit but two or three.

Lined up side by side, it hits you: the single phrase "large company" is holding both 0.5 and 4,000 at once. Ignore the orders of magnitude and that entire gap evaporates. Hsee et al. (2003) and their evaluability research land on the same conclusion. Put two numbers side by side and people feel a 100x difference clearly; evaluate them separately and they rate them almost the same. Comparison itself is the device that breaks the desensitization.

Stereotypes overwrite magnitude

Narratives like "tech firms are always bigger" or "old companies are small" paper over the actual ratios. Thanks to the availability heuristic, the extreme cases that pop to mind most easily quietly become your reference point. Ask about CEO pay and people reach first for Musk's option packages worth tens of trillions, while the median S&P 500 CEO compensation sits around 16 billion KRW. Salaries, home prices, market caps, it's the same trick: one or two easy-to-recall cases pose as the whole distribution. A sense of magnitude is the work of dragging that narrative back to data, swapping "both are expensive" for the habit of nailing it down as a ratio: "this one is about 7x the other."

✍️ Operator's note — I used to be the person who lumped it all together — "Samsung or whoever, they're all big companies." Only after actually ranking market caps did I find firms whole orders of magnitude apart sharing that same "big company" label. If the order of magnitude is different, it's simply a different world. Treat 5 trillion and 50 trillion as the same "big money" and every news headline lands with the same flat thud. So now, whenever a large figure shows up, the first thing I run in my head is: "If 1 trillion equals 1, what is this?"

How this game builds your sense of magnitude

Higher or Lower forces two assets side by side and makes you rank them on the spot. That's the whole point: a ratio that goes dull in isolation is dragged into sharp focus by the act of comparison. On a wrong round, both real figures drop together, so you instantly see by how many times your intuition was off. Because one miss ends the run, the habit of conservatively checking the ratio gets paid back in score. Repeat within the same groups (market cap, salary, real estate, commodities) to find your weak spots, and every time a big number shows up, convert it in your head: "If 1 trillion is 1, how many times is this?" After only a few rounds, you've wired the circuit to read headline figures not as a single lump but by their order of magnitude.