Japan stock prediction (Nikkei / TSE): the honest numbers and why it's a coin flip
On a full public log of about 2,405 verified predictions, our model scores roughly 49.9% directional accuracy on Japan — almost exactly a coin flip. Here's the honest reason the Nikkei and TSE land right at even for our AI model.

I'd rather show you a market where we sit right at a coin flip than dress it up. Across roughly 2,405 verified predictions on Japanese equities — the Nikkei and the broader Tokyo Stock Exchange — our model has landed about 49.9% directional accuracy. That is almost exactly even (50%): on this market the model has essentially no edge, up or down. The sample is now large enough that this isn't early-days noise — it's a stable read, and 49.9% on thousands of calls is the honest verdict. Every one of those calls is public at /predictions, the wrong ones included, because that's the whole point of this company.
For context, the same model runs about 49.7% blended across all 2,405-plus verified calls we've logged spanning 16 markets. Our two strongest markets — Canada and the US — sit around 53%, a thin but real edge. Japan, at 49.9%, is right in the middle of the pack: not our worst, not a weakness, just a market where momentum-style signals come out a wash. I find that genuinely more interesting than the markets where we squeak ahead.
So let me be honest about why Japan lands at even, instead of pretending the number means more than it does.
Our edge is momentum, and Japan mutes momentum
Most of our features are some flavour of price and volume momentum — the model learns from how a stock has been trading and extrapolates the short-term tendency. That signal gets diluted in a market where price formation is driven by things that never show up in a price chart. Japan is that market, for at least four structural reasons, and together they're enough to cancel out whatever small edge our features find elsewhere.
Cross-shareholdings and keiretsu ties. A large slice of Japanese share registers is still held by affiliated companies, banks and group partners — stakes held for relationship reasons, not because anyone is trading a view. When a meaningful fraction of the float doesn't move on fundamentals or sentiment, clean price discovery gets muted. The signal my model is trained to read is partly absent by design.
Lower retail participation. Historically, Japan has had thinner retail involvement than the US, and far thinner than Korea, where individual traders dominate flow. Retail activity is noisy, but it's also where a lot of the short-term momentum behaviour my features are built to detect actually lives. Less of it means less for the model to grab onto.
Years of Bank of Japan ETF buying. For a long stretch, the BOJ was a mechanical, price-insensitive buyer of Japanese equity ETFs as a matter of monetary policy. That is a giant flow that no technical indicator can anticipate — it doesn't respond to momentum, valuation, or news the way a normal participant does. A model trained to read market-driven supply and demand is, in effect, reading a tape that a central bank has been leaning on. The distortion is invisible on the chart and corrosive to a momentum model.
Governance reform you can't see in the technicals. The TSE is in the middle of a serious corporate-governance push — most visibly the campaign pressuring companies trading below book value to fix it, the so-called "PBR above 1" effort. When a re-rating is driven by policy and management reform rather than by trading behaviour, it arrives as a step-change that the price history simply doesn't foreshadow. My features see the past; this kind of move is about a future the past doesn't contain.
Why I'm not dressing a coin flip up as skill
The easy move would be to round 49.9% up to "about half right, basically as good as the pros," or to keep publishing calls while quietly burying the hit rate. Plenty of tools do exactly that — I wrote about the incentives in why most AI stock-picking tools are lying. The dishonest version of this business is very profitable. I'm trying to run the other one.
An even number tells me something real: our feature set is a wash against this regime, not that the market is unbeatable. Getting an edge in Japan probably requires inputs we don't yet have — flow and ownership data, policy-event tagging, governance-reform signals — rather than more momentum features layered on top of a market that mutes momentum. That's an honest research problem, and it's a different model than the one we ship today. One more caveat worth stating plainly: a clean backtest on Japan would flatter these numbers, because a backtest gets to learn on history it can also see. The live, forward-only record is the real one, and it's the lower, even number — 49.9%. You can read exactly how the current model is built, and what it does and doesn't look at, on our methodology page.
Here's the practical takeaway, stated plainly: on Japan, weight your own research more heavily here. We label every call Bullish, Neutral, or Bearish — never Buy or Sell, never a guarantee, never as instructions — and on this market in particular a coin-flip hit rate means that label is close to no information at all. I'd rather you knew that than discovered it the expensive way. A tool that can't tell you where it has no edge isn't a research tool — it's marketing. We publish the 49.9% so you can hold us to it.
See the evidence for yourself — download the full resolved-prediction dataset, read the live public self-audit (hit-rate confidence intervals, live-vs-backfill split), inspect every model card, or run the research tools on your own data. No hype, just the receipts.
This article is educational content about machine learning and market structure. It is not financial advice, not a recommendation to buy or sell any Japanese-listed or other security, and not directed at any individual's circumstances. Trading Agent is a quantitative research tool operated by WU Capital Limited (New Zealand).


