Roughly nine out of ten video games made before 2010 are not legally available to buy today. Not expensive — unavailable. The servers are off, the publisher is gone, the licence expired, the hardware has failed. Against that background, a loose international network of archivists, emulator developers, libraries and hobbyists has spent three decades keeping interactive history playable. This is how game preservation actually works, where the law sits, and how to browse these collections without ending up somewhere you should not be.
The preservation problem in numbers
A widely-cited 2023 study by the Video Game History Foundation found that around 87% of video games released in the United States before 2010 are no longer commercially available in any form. For context, that is a worse survival rate than American silent film, a medium routinely held up as a preservation catastrophe.
The causes are structural rather than negligent:
- Licences expire. Sports titles, music games and licensed tie-ins vanish on a schedule, because the right to sell them was always time-limited.
- Publishers dissolve. Rights end up in estates, holding companies and bankruptcies where nobody knows who owns what, and the cost of finding out exceeds the value of the game.
- Storefronts close. Digital-only titles disappear entirely when a platform shuts down its shop.
- Servers switch off. An online-only game is simply gone the day the backend stops.
- Media decays. Optical discs rot, cartridge batteries die, magnetic media demagnetises.
What an archive actually contains
People imagine a pile of games. A serious archive holds considerably more, and the surrounding material is often the part that cannot be recreated.
| Category | Examples | Why it matters |
|---|---|---|
| Software images | Cartridge dumps, disc images, executables | The game itself; useless without the rest |
| Paratext | Manuals, box art, maps, feelies, warranty cards | Often essential to actually playing older games |
| Development material | Source code, design documents, prototypes, build logs | The only route to understanding how something was made |
| Marketing and press | Adverts, magazine coverage, review copy | Reconstructs how a game was received at the time |
| Community record | Forums, fan sites, speedruns, mods, server logs | Documents how a game was actually played |
| Hardware | Consoles, controllers, peripherals, CRT displays | Timing, latency and image quality depend on it |
Emulation: the technology that makes it possible
An emulator reimplements original hardware in software: the processor, graphics and sound chips, timing, input and quirks. Done well it is an act of reverse-engineering scholarship. The hard part is never the processor; it is the undocumented behaviour that games depended on, sometimes accidentally.
- Cycle accuracy. Some games rely on exact timing. Approximate it and they break in ways that take years to diagnose.
- Audio chips. Reproducing a sound chip's analogue character is genuinely difficult, and enthusiasts can hear the difference immediately.
- Display expectations. Art was drawn for CRT displays whose scanlines and colour bleed were part of the technique. On a modern flat panel, it looks wrong because it is wrong.
- Input latency. Modern displays and USB polling add delay that did not exist. For precision games this changes how they play.
Crucially, emulators themselves are legal in most jurisdictions. Writing software that behaves like hardware is not infringement, a position supported by US case law on reverse engineering for interoperability. The legal complexity lies in the files you feed it, and in firmware.
The law — and the five distinctions that matter
| Activity | General position | Caveat |
|---|---|---|
| Writing or using an emulator | Generally lawful | Distributing copyrighted firmware with it is not |
| Dumping a game you own | Often permitted as a personal copy | Varies by country; may conflict with anti-circumvention rules |
| Downloading a game you do not own | Infringement | “It is abandonware” has no legal meaning — copyright persists regardless of availability |
| Distributing copies | Infringement, and the most enforced | This is what takes archives offline |
| Institutional preservation | Narrow exemptions exist | Frequently limited to on-site access, which is why libraries cannot simply publish collections |
The unresolved tension is obvious. Copyright terms run for decades past the point at which a work is commercially available, and preservation exemptions are narrow enough that lawful archives often cannot let the public actually play what they hold. Libraries and archivists have petitioned repeatedly for wider remote-access exemptions; progress has been incremental at best.
This is general information, not legal advice. If you are running a project in this space, get advice in your own jurisdiction.
Who does this properly
- National and university libraries with formal games collections, cataloguing to library standards and preserving development material alongside the software.
- Dedicated foundations that acquire source code and design documents, and publish research rather than ROMs — the 87% figure above came from exactly this kind of work.
- Museums maintaining playable hardware, which is far harder and more expensive than maintaining files.
- Emulator projects documenting hardware behaviour in code and in prose, producing technical records that outlive the machines.
- Publisher-run rereleases, which preserve commercially valuable titles and, by definition, ignore the ones nobody will buy.
The honest summary: institutions do the careful work, communities do the broad work, and neither can do the other's job.
The hardware clock nobody can stop
Software can be copied forever. Hardware cannot. Cartridge save batteries reach end of life; optical media delaminates; capacitors in 1980s and 1990s consoles dry out; CRT displays are no longer manufactured and the remaining stock is finite. Every year, the number of machines capable of running these games natively falls.
This is why accurate emulation is not a convenience but the actual preservation strategy. At some point, the software reimplementation will be the only surviving version of the hardware — which places a serious documentation burden on the people writing it today.
Online-only games: the hardest case
A single-player game from 1994 can be preserved. A multiplayer game whose logic lived on a server switched off in 2014 largely cannot. What remains is video, documentation and, where communities have reverse-engineered the protocol, private server reimplementations — which are legally contested and technically incomplete.
The live-service model makes this worse every year. A game that exists only as a client talking to a proprietary backend has a built-in expiry date, and consumer-rights campaigns in several jurisdictions have begun arguing that publishers should be required to leave something playable behind. Nothing has been settled.
Browsing responsibly
- Prefer institutional sources. Library and foundation collections are lawful, catalogued and do not serve you advertising.
- Buy the rereleases that exist. The commercial case for preserving more is made by the sales of what has already been preserved.
- Dump what you own. If you still have the cartridges, the equipment to read them is cheap and the result is unambiguously your copy.
- Treat “abandonware” sites with suspicion — technically as well as legally. Sites distributing executables outside any legal framework are, unsurprisingly, a common malware vector. Never run an installer from one.
- Support the documentation work. Manual scans, oral histories and hardware notes are the part nobody funds and nobody can recreate later.
Verdict
Game preservation is a race between a slow legal system and fast physical decay, run mostly by volunteers. The law currently protects works long after anyone is willing to sell them, which converts cultural history into an accidental casualty of rights management. Until that gap closes, the practical position for a reader is straightforward: use the lawful archives, buy the rereleases, dump what you own, and give the people writing emulators and scanning manuals the credit they are owed — they are doing the work a functioning system would have funded decades ago.
