Onimusha: Way of the Sword and the Stopwatch Ledger Nobody Cross-Checked

My desk always has two tabs open: one for the original, one for the cross-check. The habit dates back to 2026, when I was twenty-four, working as a data editorial assistant at a football outlet in Marseille, and I ran into a wrongly disallowed goal in the 73rd minute of Olympique Marseille versus AS Monaco in Ligue 1. I went back through the tape, opened Law 11.3, and found that the referee had ignored the fact that defender Kamil Glik deliberately touched the ball before Dimitri Payet shot. The 1,200-word analysis I wrote afterwards, complete with three situation diagrams, pulled 40,000 reads in twenty-four hours — five times the site's daily average.

The lesson I took from that day was not about writing better than everyone else. It was about cross-checking before concluding.

So when I picked up the information set on Onimusha: Way of the Sword, the single-player action game published by Capcom, I did exactly what I always do with a new file. I underlined every number. I circled every timestamp. I flagged every comparative claim. Then I checked whether they agreed with each other.

They do not fully agree. And that gap is the part worth talking about.

Context: a single-player game sitting inside an esports folder

Onimusha: Way of the Sword is a one-player action game. The protagonist is Musashi. The setting is Eastern Kyoto in the Edo Period, where local myths and legends are folded into a side-quest chain called Curious Cases. There is no multiplayer mode. There is no competitive patch cycle. There is no tournament. There is no roster. There is no prize pool.

This detail matters more than it looks, because it determines how the analysis should be done.

My job is reading rules and checking match reports. My tools are the penalty scale, the timestamp, the system log. Applying that toolkit to a single-player game produces a pile of empty conclusions, things like "Onimusha's meta is shifting" — a meaningless sentence, because no competitive meta exists here. There is no character pool. There is no win rate. There is no pick-ban phase.

At thirty-three, I no longer only criticise. I have started proposing. And my first proposal here is a classification proposal: this information set belongs under "consumer video games — single-player action", not under esports. Feeding it into a competitive analytics pipeline contaminates esports data with single-player consumer content. That is a process failure, not a personal one. And process failures have to be fixed at the process layer.

But there is one thing I can apply to this file: the ability to read a number. And Onimusha gives me enough material for that.

Capcom has placed this title in the same year, 2026, as Resident Evil Requiem and Pragmata. That is how a major publisher arranges a product slate: three large single-player titles in one year, with Onimusha positioned as the one with the longest campaign. The accompanying material claims its runtime is significantly longer, with more hours of content than the other two combined.

That is a market-positioning statement. Not yet a measurement.

Core: two numbers for one campaign

Start with the central figure. The main campaign runtime is recorded at two different levels in the same source. Level one: the main story may take about 30 hours. Level two: players on the Action difficulty setting take between 30 and 40 hours.

These two figures sit a few paragraphs apart, are never reconciled, and come with no stated methodology.

Technically they do not directly contradict each other: 30 hours falls inside a 30-to-40-hour band. But the two formulations differ in kind. One is a point estimate. The other is a range estimate. When an article uses both without saying which applies to whom, readers pick whichever number suits their expectation. Buyers pick 30. Promotional writers pick 40.

This is an error I have seen hundreds of times on football stat sheets. Distance covered is packaged as an effort metric. Sprint counts are presented as proof of intensity. But ineffective running still produces pretty numbers. A player who covers 12 kilometres without once touching the ball in the final third still owns a very impressive row on the summary sheet.

The thing that can lie is the person presenting the number, not the number itself.

With Onimusha, the question worth asking is not 30 versus 40. It is: measured how, on what sample, with which player group, and under what definition of "completed". The file answers none of those four points. A missing methodology is a form of error, and it does not disappear just because a number gets repeated enough times.

Boss density: 36 fights inside 30 hours

The second figure is far more interesting. The main campaign contains 36 boss encounters.

At the 30-hour mark, that is roughly one boss every 50 minutes. At the 40-hour mark, it stretches to roughly one every 67 minutes.

To someone who reads match reports for a living, density is the thing that matters, not the total. A season with 38 rounds tells you nothing about that season's quality. What tells you something is the gap between rounds, the fixture congestion within a month, the fact that a team has to play three matches in seven days with essentially the same eleven.

By the same logic, 36 boss fights is a statement about pacing. At one boss per 50 minutes, the game almost runs as a boss-rush structure: the player crosses a short stretch, meets a boss, kills the boss, repeats. At one boss per 67 minutes, there is more breathing room for exploration, for side content, for narrative rhythm.

I do not have the data to say which type this game is. But I have enough to say that the number 36 is not, by itself, a plus. It is a double-edged variable. High boss density can create sustained tension, or it can create a sense of repetition. The source itself concedes that some boss fights can be considered harder than others — a roundabout phrasing that admits uneven quality across the encounters.

Here I have to be explicit about my level of certainty. The pacing conclusion is inference, not observation. It rests on a simple division between two numbers the source supplied. If the 36 figure is wrong, or if the 30-hour mark is wrong, the conclusion collapses with it. I record this section as a hypothesis awaiting verification, not a verdict.

The platinum trophy and the price of a second run

Another detail in the file carries more structural weight than it appears to. To earn the platinum trophy, a player must finish the game at least twice. That doubles the time commitment for anyone pursuing full completion.

The material says players who are not used to action games may need more than 50 hours. Completionists are also placed at 50 hours and above. Numerically, these two groups differ completely in cause but coincide in outcome. The first group is slow because of skill. The second is slow because of volume.

Once again I come back to VAR. VAR is not wrong. The person operating VAR is, after all, only a person. Here, the platinum mechanism is not wrong. But it imposes a requirement the designers must explain: why a second playthrough is mandatory, and what that run delivers beyond repetition.

There are two scenarios. Scenario one: the second run retains all upgrade progression, turning it into a speed-optimised route where the player already knows where the bosses are and how to fight them. Scenario two: the second run wipes progression and forces a restart from zero. Scenario one is a reward. Scenario two is a time penalty.

The information set confirms neither. That is an information gap, and it is the worst kind: the kind that directly affects a purchase decision. The offside line was never straight — it is only today that I can see how it bends.

Side content with power locked behind it

This is the most practically useful finding in the entire information set, and it is buried near the end.

The game has an upgrade system built on two materials: iron and leather. These are used to upgrade Musashi's sword and clothing. Where do they come from? Side quests and Ace Archer challenges.

In other words, side content is not purely decorative content. It is part of the power curve. The material uses the word "essential" when describing iron and leather's role in upgrading.

This optional layer is estimated at 10 to 15 hours. The rest of the optional layer includes Curious Cases — the side-story quest chain rated the best of the group — along with treasure hunts, chance encounters, and Lion Dogs, described as the most endearing creatures in the game. Some other optional content unlocks sword skins. Two named trophies appear in the material: No Job Too Small and For the Love of Dog.

For an experienced player, this is planning information more important than any runtime figure. If upgrade materials are gated behind side content, then a player who burns through the main story walks into the late bosses with thinner gear than a player who did the side quests. This is a familiar design tension in the action genre, and the source material does not flag it at all.

I call this a legal gap. A disallowed penalty can be fixed; a legal gap cannot. In this case, the gap is that players are never warned that "optional" is not entirely optional. My concrete proposed clause: the in-game tutorial must state, from the very first stage, that key upgrade materials sit behind optional content. The amendment cost is close to zero. The risk is close to zero. The roadmap is one line of text.

The contrarian angle: fan emotion is valid data

I have to give a paragraph to the people waiting for this game. Their emotion belongs to the category of valid data, not noise.

When an old franchise is revived, most fans do not buy because of a runtime table. They buy because of memory. They want the feeling they had at seventeen, and they will forgive a great deal to get it back. This is legitimate. Numerical analysis cannot refute it, and should not try.

But precisely because expectations run high, numbers become more dangerous.

Take the central claim of the information set: Onimusha's runtime is significantly longer than Resident Evil Requiem and Pragmata combined. This is a comparison with no denominator. No figures are supplied for the other two titles. No definition of "runtime" is given — main campaign, or total completion time? No measurement point.

Logically, this is a claim that cannot be falsified. It cannot be refuted because it does not supply enough to be refuted. The strength of the argument is not there. Its blind spot is.

And this is where I have to check myself. I have a tendency to spot logical errors faster than other people, and that tendency easily turns into contempt for sentiment. I was reminded of this in 2026, in a studio, France versus Australia, 55th minute, when I said on air that the IFAB 2026 Referee Review Area protocol had not been followed. I was right on process. But being right on process does not mean communicating effectively. A Paris chief editor still hired me afterwards to train fifteen regional commentators in the rules for the following season, but the lesson about how to say things I kept for myself.

So I write this section differently. The issue is not whether that claim is true or false. The issue is: if it is false, who finds out first.

The answer is the player community, within days of release, through aggregated completion-time databases. That is the only reality-check mechanism that operates here. There is no match report for me to read before the news, because there is no match. The only mechanism left is a mass of players pressing a stopwatch and publishing the result.

I read the match report before I read the news, because the report does not know how to lie. Here, the report will be written by the community.

Risks and what to track

Summarising, this file carries four risk points.

First, data quality. Two runtime levels coexist unreconciled, with no methodology. Level: medium.

Second, domain misclassification. The esports label is unsupported by any competitive entity in the text. Level: medium. Recommendation: reclassify.

Third, an unverified comparative claim. The "longer than the other two combined" line has no comparator figures. Level: medium. Recommendation: treat as marketing framing, not measurement.

Fourth, design-friction risk. A density of 36 boss fights inside roughly 30 hours, plus a two-playthrough platinum gate, may create repetition or fatigue among completionists. Level: low to medium.

On the opportunity side, three points worth tracking.

Concrete, checkable content-volume data: 30 to 40 hours, more than 50 hours, 36 bosses, 10 to 15 hours of optional content. The time window is the period immediately after release, when runtime queries peak.

Capcom's three-title single-player slate for 2026 is a genuine signal of publisher conviction in the premium single-player model. Time window: across the 2026 release calendar.

And information with immediate actionable value: upgrade materials are gated behind optional content. Time window: from the very start of play.

Provisional conclusion

Based on what has been verified, here is my provisional conclusion.

The runtime data in the information set is of average quality. I recommend against using it as the sole basis for a purchase decision.

The 36-boss figure is the data point with the largest structural load, and it needs verification through post-release feedback on pacing.

The gating of upgrade materials behind side content is information with immediate actionable value from the earliest stage of play.

What I am still waiting to verify: whether the second playthrough for the platinum trophy carries over upgrade progression. If it does, the two-run requirement is a reasonable design. If it does not, it is an unjustified time penalty.

Whoever writes the rules needs someone standing outside the line to check their signature. Here, the rule-writer is the publisher, and the person outside the line is the player community. I will read the aggregated completion-time table before I read the review scores.