Code Veronica Remake: Eight Leaks, Four Confirmed Facts, and a 54 Percent Ledger

A Vietnamese gaming site ran the headline: “Resident Evil Veronica Remake leaks new script.” The wording suggests someone walked out of Capcom's office with a document. I opened it and did nothing but count. Twelve claims in total. Four of them can be checked against Capcom's own statements: the remake is in production, it uses a third-person view on RE Engine, the core team comes from the RE2 and RE4 remakes, and the story will receive “adjustments.” The other eight arrive from “unnamed sources.” The headline sold the eight and put the four on the poster as evidence.

My opening verdict is blunt: the game is real, the leaks are plausible, the headline is dishonest. Capcom has not said a single sentence about the Code Veronica script; an anonymous source said it, and the headline dressed it in an official voice. My job is not to predict the game. My job is to weigh eight claims separately and then land on a number.

Context: the last Dreamcast chapter and the remake ledger

Code: Veronica released on 3 February 2026 for the Sega Dreamcast. It is the last classic fixed-camera mainline entry, made just before Capcom decided the camera had to change. The story sits in December 2026 on Rockfort Island: Claire Redfield searching for her brother Chris and getting trapped in a private prison complex alongside Steve Burnside, the Ashford twins, and a returning Albert Wesker. It is the most neglected classic in the series, largely because the Dreamcast's commercial failure buried its reputation.

I have kept Capcom's remake ledger for years, because the dates alone tell a pattern. RE2 Remake on 25 January 2026: a full rebuild, modernised story, Mr. X's footsteps added. RE3 Remake on 3 April 2026: content-cut complaints, mixed reception. RE4 Remake on 24 March 2026: expanded story, Ada content, critical and commercial success. Three remakes, two outcomes. The variable was not talent. It was scope and time.

In January 2026 I built a checklist for RE2 Remake: no cuts, more than six hours of playtime, puzzles intact. Every box got a tick. Then I walked into that police station corridor and understood that fear does not show up on a spreadsheet. The spreadsheet said one thing. The stadium said another. With RE3 Remake it inverted: correct on paper, hollow in the hand. That double experience produced my verification method—feeling comes last, but feeling is never deleted.

Now the strongest signal in the entire story. Capcom has confirmed that the Code Veronica core team is the group behind the RE2 and RE4 remakes. That is the only part of this news that does not rest on inference. When one studio delivers two excellent remakes and one compromised remake in a single decade, its expected value on a fourth project leans toward the two. I do not trust a signal until it survives a cold Tuesday in February—meaning, until scope, budget and schedule all show evidence of aligning at once.

Core analysis: eight rows, eight different weights

In my day job I do this with football matches: build a shot map, count the goals, and decide who actually won. I built the xG model before I understood the market, and that taught me the difference between a raw number and a weighted number. Leaks work the same way. I judged each claim on three axes: what it costs to build in RE Engine, how it aligns with Capcom's previous remakes, and how specific the leak's language is.

The cheapest claim is the crafting bench in safe rooms—70 percent. RE4 Remake's attache case and Village's resource management show Capcom has installed that economy four times. A bench is not a new system; it is a rearrangement of an old one. The Bandersnatch return (65 percent) costs more, because a new enemy needs models, animation and AI, though Capcom's habit of restoring trimmed content was visible in RE2 Remake.

The most credible leak is the expanded Chris–Wesker confrontation at 75 percent. That fight is the original climax, and RE4 Remake made Verdugo longer and more brutal than the source material. The leak's phrasing—“longer, more brutal CQC”—matters. Someone who only plays games does not use the term CQC that way. Someone who has seen design discussions does.

Steve Burnside carries two separate claims: a character rebuild (60 percent) and a rewritten Claire–Steve dynamic (55 percent). Steve is widely regarded as the weakest-written character in the franchise, and Capcom does not hesitate to rewrite unpopular characters—RE4 Remake visibly rebuilt Ada's arc. But this is also the least falsifiable claim. Nobody can prove before launch whether a relationship changed.

Leon S. Kennedy's appearance is only 25 percent, and this is where my scepticism concentrates. Code Veronica takes place on Rockfort Island in December 2026; Leon is recovering after Raccoon City and entering training. Geography and timeline both resist it. This leak probably translated asset reuse or engine cross-pollination into a story rumour. Correlation is not causation, and this row is the clearest example.

Semi-linear progression (50 percent) and dynamic outfits (30 percent) sit at opposite ends of cost. A hub-and-spoke structure is normal modern Resident Evil: the RCPD hub in RE2 Remake, then linear sewers. Outfits that shift with weather, injury and scratch marks mean a separate texture set, blend and lighting pass for every state. Technically possible, economically expensive. A studio carries that cost only when the system is a marketing headline. That signal does not exist yet.

The weighted average of the eight leaks in my ledger is 54 percent. That is not “half true.” It is a ceiling: right now there is not enough evidence to act on any of it. The irony is that the claims closest to our expectations—Chris–Wesker, crafting—feel the most credible precisely because our priors do the work, while independent evidence for them is zero.

Contrarian angle: plausibility is not proof

Here I will admit the weakest joint in my own framework. Two of my three axes—Capcom's precedent and the leak's specificity—are reflections of my own expectations. Data is not the game. Data is the game confessing its patterns. But this data concerns a game nobody has seen. Pattern-matching cannot verify truth here; it can only rank probabilities.

The second danger is not authenticity but scope. The lesson of RE3 Remake is clear: the failure was not talent, it was cutting. If Code Veronica Remake loses its twelve-to-fifteen-hour structure and becomes an eight-hour linear product, it will not matter whether the leaks were true. The opposite risk is equally real: Capcom can expand so much that the original's isolation disappears.

The third problem is the economics of the news itself. The site that published this runs T1/Faker, PUBG Asia Stars and beauty content alongside it—a volume aggregator model where the headline's job is to pull readers, not to verify information. Specificity will never impress me on its own. If the leak really comes from a QA tester or a localisation team, Capcom's correct move is silence, because every denial raises the curiosity it wants to suppress.

The final risk belongs to the buyer. Pre-ordering on a leak means betting money on an unverified claim. A 25 percent Leon number does not mean the claim is false. It means that acting on it will be wrong three times out of four.

Takeaway: locking the number and the name

I am writing two lines into the ledger today. First, the 75 percent on Chris–Wesker stays untouched until an official studio announcement or gameplay trailer; if no official footage arrives within 90 days, I cut it to 55 percent. That is my kill criterion. Second, the 30 percent on dynamic outfits and the 25 percent on Leon will never cross 50 until Capcom shows them on screen.

My number for Code Veronica Remake today is 54 percent, and the name I am watching most closely over the next six months belongs to no executive—it is Claire Redfield. She is the only figure in this entire report about whom Capcom has said nothing wrong, and for exactly that reason her face will be the first piece of evidence.