I Sold 3 Japanese Stocks 15 Years Ago. They'd Now Be Worth $200K More — Here's What I Got Wrong

The short version: I lost ¥2.2M in cash, then watched ¥30M walk away

Fifteen years ago I sold three Japanese stocks: Sony, Koei Tecmo, and IIJ.

The realized losses were ugly enough at the time:

• Sony (6758): ¥300K loss, about $2,000 at today’s exchange rate
• Koei Tecmo (3635): ¥100K gain, about $670
• IIJ (3774): ¥2M loss, about $13,300

Net realized: roughly ¥2.2M ($14,700) gone. For a guy in his early 30s, that hurt.

But that’s not the part that haunts me.

If I’d just sat on those three positions for fifteen more years, each one would now be up roughly ¥10M ($67K). All three combined: about ¥30M, or $200K. The total swing from “what I did” to “what I should have done” is closer to ¥32M — $215K — when you add the realized loss to the missed gain.

Two hundred grand. Because I couldn’t sit still.

I’ve written about the day I cut a ¥2M loss on IIJ and the moment I finally stopped trading individual stocks in separate posts. This one’s different. This one lines all three up next to each other and asks: what was the actual pattern? What was I doing wrong, every single time?

Quick Japan Context (skip if you already know NISA)

If you’re reading from the U.S. or Europe and haven’t dealt with Japanese investing accounts, the 30-second version:

• NISA is Japan’s tax-advantaged investing account. Closest analogue: a Roth IRA. Gains and dividends are tax-free, but there’s an annual contribution cap (currently ¥3.6M, around $24K). The current version was launched in 2024 as a major overhaul of the original 2014 program.
• Sony Group (6758) is the company you already know — PlayStation, cameras, music, semiconductors. Listed on the Tokyo Stock Exchange, also has an ADR on the NYSE if you’d rather buy it in dollars.
• Koei Tecmo (3635) is a Japanese game publisher. If you’ve played Dynasty Warriors, Nioh, or Atelier, that’s them.
• IIJ (3774) — Internet Initiative Japan — is one of the oldest ISPs in the country. They also run IIJmio, a budget mobile carrier.

So none of these are obscure picks. I wasn’t gambling on penny stocks or pre-IPO startups. These were three big, recognizable Japanese names, all profitable, all paying dividends. Which is part of why the regret stings so much. The thesis wasn’t wrong. The companies didn’t blow up. The holding period was the problem. It was always the holding period.

One more piece of context worth noting: Japanese retail investors at the time mostly traded in 100-share or 1,000-share lots, depending on the stock. IIJ before its splits was a 1-share unit but traded at over ¥300,000 per share. That’s a ~$2,000 minimum position, which is part of why my IIJ loss got so big so fast — you couldn’t really nibble in, you had to take the whole bite.

The math, with the splits baked in

Here’s what makes Japanese stocks tricky to look back on: stock splits. A lot of them.

Sony did a 1-for-5 split in October 2024. Koei Tecmo did four separate splits between 2014 and 2022. IIJ went absolutely nuts — a 1-for-200 split in 2015 followed by two more 2-for-1 splits, for a cumulative 1-for-800 ratio.

If you don’t adjust for splits, the numbers look insane in the wrong direction. So let me show the real math.

Stock	Code	~2011 price range	May 22, 2026 close	Cumulative splits since 2011
Sony Group	6758	¥1,300–¥2,500	¥3,525	1→5 (Oct 2024)
Koei Tecmo HD	3635	¥400–¥700	¥1,531	1.2 × 1.2 × 1.3 × 2 = 3.744×
IIJ	3774	¥280,000–¥365,000 (pre-split)	¥3,071	200 × 2 × 2 = 800×

Once you adjust for the splits, here’s what one share I sold back then is actually worth today:

Stock	Sale price (assumed)	Today’s value of that one share	Multiple
Sony	¥2,000 (~$13)	¥3,525 × 5 = ¥17,625 (~$118)	~8.8×
Koei Tecmo	¥550 (~$3.70)	¥1,531 × 3.744 = ¥5,732 (~$38)	~10.4×
IIJ	¥320,000 (~$2,133)	¥3,071 × 800 = ¥2,456,800 (~$16,378)	~7.7×

Roughly 8 to 10× in fifteen years. Annualized, that’s 15–18% a year. Better than the S&P 500’s long-term average. Better than basically any sane “long-term hold” projection I would have made in 2011.

And these aren’t moonshot tech startups. Sony is a 79-year-old hardware conglomerate. IIJ is a B2B telecom company. Koei Tecmo makes games about medieval Japan. Boring, knowable businesses. They just kept compounding.

The ¥30M ($200K) figure isn’t dramatized. I had real, non-trivial positions in each. If I’d kept all three at their sold sizes, the missed gain is in that range.

Why I sold each one (it’s embarrassingly similar each time)

The buying logic was actually fine. I bought what I knew.

Sony first. I’ve owned a PlayStation since the PS1. My TV is a Bravia. My headphones are Sony. Buying Sony stock was basically a “I like this company, might as well own a piece” decision. Held it three months. They had a bad earnings print. The stock dropped. I sold. Down ¥300K.

Then Koei Tecmo. I grew up playing their strategy games — Nobunaga’s Ambition, Romance of the Three Kingdoms. The shareholder perks (a Japanese thing — companies send you stuff if you hold their stock) were decent. Held this one about a year. Got into the green by ¥100K. Sold to “lock it in.” The only winning trade of the three. And in hindsight, the worst one — I sold the actual compounder.

Then IIJ. This is the wound that still aches. Their subsidiary had just launched IIJmio, a budget SIM service, and I had this clear narrative in my head: budget mobile is going to explode in Japan, IIJ is positioned for it, this rides for years. Held three months. One day the stock dropped hard. I panicked. I clicked sell. Down ¥2M. About $13K. Roughly one and a half months of take-home pay at the time, gone in an afternoon.

And the thesis? The IIJmio thesis? It played out. Almost exactly the way I predicted. Budget mobile did explode in Japan. IIJmio became one of the major MVNO brands. IIJ’s stock did what I thought it would do, on basically the timeline I thought it would do it on. I was right about everything except the part where I had to sit through a quarter of bad price action to collect on being right. I exited because of noise. The signal arrived on schedule. I just wasn’t in the position anymore to see it.

Notice the pattern? Three months. Three months. Twelve months. The longest I held any of these “long-term convictions” was twelve months. For a thesis that needed at least five years to play out.

I wasn’t an investor. I was running a slow casino on myself.

The actual pattern: I had no scenario where the stock went down

Lined up side by side, the trio gives away the game:

• IIJ: stock dropped sharply → reflex sell
• Sony: stock dropped on earnings → reflex sell
• Koei Tecmo: stock went up → reflex sell

Every single sale was a reflex. Not one was a considered decision. There was no review of fundamentals, no check on whether the original thesis was broken, no math on valuation. Just a finger on the sell button.

Here’s the thing I only see clearly now: when I bought each of these, I had no mental model for “what if it goes down.”

I’d built a thesis for why it should go up. That’s it. There was no plan B. No “if it drops 20% but the business is fine, I’ll add.” No “if it drops 30% I’ll cut, otherwise I won’t touch it.” Nothing.

So the moment reality diverged from the upside-only mental model — even slightly — I had no script. And when I have no script, I do the most emotionally satisfying thing in the moment. Which is to make the discomfort stop. Which means selling.

The Koei Tecmo case is the most revealing. I bought it expecting it to go up. It went up. And I sold. Because once I had a paper gain, a new, louder voice took over: “what if you give this back?” Loss aversion is wild. The asymmetry is unreal. Even a small unrealized gain felt fragile enough to “lock in.” But an unrealized loss felt unbearable enough to “cut.”

Both reflexes destroy compounding. They’re the same reflex, actually. Just pointed in opposite directions. The reflex isn’t “buy low sell high” — it’s “make the current emotional state stop.” Up too much? That feels precarious, sell. Down too much? That feels terrifying, sell. Either way, the position closes. Either way, the compounding clock resets to zero.

Looking back, I think this is the real reason almost nobody actually owns a 10-bagger. Not because 10-baggers are rare — they’re not, even Sony was one in this window — but because holding one requires you to ignore both fear-of-loss and fear-of-giving-back-gains, simultaneously, for a decade. That’s not a portfolio decision. That’s a temperament that only a small percentage of people have, and I am definitively not one of them.

I wasn’t building wealth. I was day-trading cash.

The second thing I see now is that I never thought of stocks as a thing that grows.

To 2011-me, a stock was a number on a screen that could turn into a slightly bigger number or a slightly smaller number. Buy, wait, sell. That was the whole shape of it. The frame in my head was “did I make money on this trade.” Not “did this company create value over the period I held it.” Not “what’s a reasonable holding period for this kind of business.”

The phrase “let your winners run” didn’t compute. The phrase “10-bagger” sounded like marketing nonsense. The phrase “reinvest dividends for 20 years” sounded like advice for retirees, not for a 30-year-old.

I’ve written before about how the first three years of indexing feel like nothing happens. That’s true. But it’s worse than that — if your time horizon is three months, “compounding” isn’t a slow thing. It’s an invisible thing. It does not exist in your worldview at all.

I was running a cash-trading game and calling it investing. They’re not the same. They’re not even adjacent. They use different muscles and they have opposite incentive structures.

If you’re sitting there thinking “but I trade individual stocks for the long term” — okay, honest question: when’s the last time you didn’t check the price for a full week? Two weeks? A month? If the answer involves a number of hours, not days, you might be running my 2011 game without realizing it.

I’m not saying individual stocks are stupid. I’m saying they require a discipline I never had and probably never will have. There’s a small group of people who can really do this — read 10-Ks for fun, watch a position drop 40% and add to it, ignore noise for a decade. They exist. They’re rare. I assumed I was one of them because I had opinions about Sony’s product line. I was confusing “I know this company” with “I have the temperament to hold this company through a decade of news cycles.” Those are completely different skills. One of them I had. The other I did not.

What finally rewired me: NISA, and outsourcing the decision

The turn happened around the time NISA launched.

Japan rolled out tax-advantaged accounts. Low-cost global index funds became available. I started reading. And somewhere in there — I can’t pin down a single moment — the words “buy and hold” stopped being abstract slogans and started being a workable plan.

But I want to be honest about something. The intellectual conversion didn’t fix the reflex. Not right away.

A few years into automatic monthly NISA contributions, the account showed an unrealized gain of about ¥2M ($13K). TV pundits were saying the next year would be down. And the old voice came back, louder than ever: lock it in, take the family on a trip, you’ve earned it, what if you give it all back.

I sat on my hands that time. Not because I’d developed iron discipline. Because the ghosts of Sony, Koei Tecmo, and IIJ were standing in the room. Three reflex sells. Three stocks now worth roughly $200K more than the day I sold them. The math of “I always sell too early” had finally, finally accumulated enough evidence in my own data set to override the impulse.

These days, before I touch any individual security, I ask exactly one question:

Can I hold this for ten years without checking?

Not “should I.” Not “will it go up.” Just: can I, the actual human being I am, with the actual reflexes I’ve documented in three different stock sales, sit on this position for a decade without flinching?

If the answer is anything other than a clean yes, I don’t buy it. Index funds answer yes immediately, by construction. They’re a basket — there’s nothing to second-guess. No earnings call to react to. No CEO to lose faith in. No competitor to panic about. The fund just keeps rebalancing in the background and I keep doing my actual job.

So that’s what I buy. Almost exclusively. There’s a tiny bucket — under 5% of the portfolio — where I let myself own a couple of individual names for fun, like a controlled vice. The rest is boring, automated, global index exposure. And I do not touch it. I don’t even open the app most weeks.

The unsexy truth: I stuck with indexing because I gave up

Here’s the part that doesn’t make a great investing essay.

The honest reason I’ve been able to stick with index investing for years is that I’ve offloaded the decision. There is no decision. Money leaves my bank account on the same day every month. A computer buys a global index fund. I find out the price later, if I bother to look.

When you own individual stocks, every single day is a decision day. Earnings dropped — what do you do? Competitor announced something — what do you do? Currency moved — what do you do? Even if your answer is “nothing,” you’ve still had to make the call. Multiply that by a portfolio of five names. By 250 trading days a year. By however many years you plan to invest. The cognitive load is brutal, and the cumulative drag from getting any of those daily calls wrong is what eats real-world investor returns.

The index just sidesteps the whole thing.

There’s also a weirder benefit nobody talks about. When the market drops 20%, I genuinely don’t feel responsible. I didn’t pick anything. The market did the market thing. I can blame Powell, blame the Fed, blame China, blame whatever — none of it touches my ego. Because none of it was my call.

That’s almost embarrassing to admit. The investing literature wants you to “stay the course out of conviction.” Mine is closer to “stay the course because I have no skin in any specific bet.” But it works. It has worked for years. It would not have worked if I’d kept picking individual names, because I’d have re-lived the 2011 sequence over and over.

¥30M of opportunity cost. About $200K. That’s a big number. It’s also kind of a useless number — you can’t run that experiment again, and “what if I’d held” is a parlor game in hindsight.

What’s actually useful is this: my old self looked at charts every morning and felt his stomach tighten — win or lose, the tension never let up. My new self checks the account about once a month, sometimes less. The number’s been going up. Slowly, boringly, without me.

I’ll take boring.

That trade, I’d make a thousand times in a row.