Games that play in your browser — phone or desktop.

Game review

Curveball

6.5 / 10

Developer
Joshua Skeels (attribution corroborated by three independent secondary sources — a 2022 fan remake's own credits page, period forum links to the developer's domain, and Shockwave.com's later listing — but no primary company record or original credits screen was directly inspected, so treat this as well-supported rather than certain)
Publisher
Paragon Digital Media (pdmedia.net)
Released
1 January 2002
Platforms
Browser (Flash)
Genre
Pong Variant / Arcade Sports
Monetisation
No source found describes an explicit monetisation arrangement. The game was hosted directly on the developer's own domain rather than placed with a portal, which rules out a standard sponsorship or site-lock deal; the honest position is that this looks like plain self-hosted freeware, not a confirmed ad-funded or sponsored release
Players
Single-player only, against a fixed AI opponent — no two-player or online mode found in any source checked

The short version

Curveball takes Pong's oldest idea — a paddle, a ball, keep it from getting past you — and adds the one mechanical wrinkle that makes it worth a second look decades later: hit the ball with a paddle that's moving, rather than one held still, and the ball curves away from a straight line on its way back, exactly the way the name promises. It's a small, honest piece of design, well suited to a mouse-driven Flash game with no other input to spare, and it's the reason period forum posts from 2002 and 2003 describe it as "addicting" and "insidiously addictive" rather than as one more disposable ball-and-paddle clone.

What keeps the score from climbing is that the curve mechanic is the entire game — there is no second idea layered on top of it, no power-up, no two-player mode for a design clearly built around out-thinking an opponent, and the level ladder's answer to "how do we keep this interesting" is simply to speed the ball up and narrow the margin for error until, by multiple independent accounts, the final stage is close to unbeatable. That is a defensible shape for a free, developer-hosted single-idea game in 2002; it also means there is nothing here to bring a player back once the trick has been learned and the wall has been hit.

Biggest strengths

  • The curve-on-contact mechanic is a real, well-executed idea: paddle movement at the moment of the hit changes the ball's flight, turning a static reflex test into something with an actual skill ceiling
  • It reached a wide audience without a portal sponsorship or an exclusivity deal, spreading through direct links on forums and personal sites from as early as December 2002 — an unusually organic distribution story for a Flash game of its period
  • Free of any confirmed ad-funding or site-lock scheme; what survives points to a plain, unencumbered self-hosted release

Biggest weaknesses

  • The curve mechanic is the only idea in the game; there is no second system, power-up, or variation introduced across the level ladder to develop it further
  • No two-player or online mode exists despite a mechanic built entirely around outmanoeuvring an opponent — the single-player AI ladder is the only way to play it
  • The difficulty curve is reported, consistently and independently, to end in a final stage that is effectively unbeatable rather than merely hard, which reads less like a deliberate "and now the true test" design choice and more like a difficulty curve nobody re-tuned once it stopped serving the player

Who it's for: Fans of Pong-lineage arcade games who want to see the format's one truly novel Flash-era twist, and Flash preservationists curious about a title that reached a wide audience entirely outside the portal system that defined most of this era.

Our recommendation: Worth ten or fifteen minutes through an emulated build rather than a long session — the curve mechanic is genuinely satisfying to learn, but the level ladder is built to defeat you rather than to reward extended play, and there is no second system waiting once you've felt the trick work.

6.5 / 10

Curveball is one of the small number of Flash-era games that never seems to have launched through a portal at all. There is no Newgrounds page for it — none turns up in the Wayback Machine's records, and no Newgrounds-style vote count or author comment survives to check reception against, because none was ever posted there. The earliest trace this review could find is a Flash file linked directly from pdmedia.net, the domain of a small outfit called Paragon Digital Media, being passed around a boating forum on 26 December 2002 and a Pontiac forum a month later, in both cases as a plain link rather than an embed on a games site. That is an unusual way for a Flash game to reach an audience in 2002, and it is worth sitting with for a moment before getting to what the game actually does, because it shapes what evidence exists to judge it by.

Who actually made it

The attribution this review settled on is Joshua Skeels as developer, working with (or for) Paragon Digital Media as publisher. That comes from three independent sources converging on the same names: a 2022 Unity remake of the game, published on itch.io under the handle "rab," states outright that "the original game was made in Flash, by Joshua Skeels and published by Paragon Digital Media around 2002"; Shockwave.com's own later listing for the game credits it as "Developed by PDMedia.net"; and the earliest period forum links both point to a .swf file living at pdmedia.net/games/curveball/, the same domain. None of that is a primary credits screen inspected directly, and no independent biography or other credited work for a "Joshua Skeels" turned up anywhere else this review checked — so the name should be read as well-corroborated, not certain. What can be said with more confidence is what it is not: nothing in any source checked supports "Shawn Charlton" as the developer, a name this review was asked to verify and found no trace of connecting to this game, or to Flash development generally, under any spelling.

The one idea, and it's a good one

Strip away the "3D" framing and Curveball is Pong: two paddles, one ball, keep it from passing yours. What earns the name is the single mechanic layered on top — hit the ball with a paddle that's moving, rather than one sitting still, and the ball comes off at an angle and with a spin that bends its flight path, rather than bouncing back on a straight line. Multiple period descriptions of the game agree on this in near-identical language: catch it on the edge of the paddle in motion, and the curve makes it harder for the AI on the other end to line up a return. It's a small addition to a decades-old formula, but it's a genuinely well-chosen one — it turns a reflex test into something with an actual decision buried in it, timing not just where you meet the ball but how you're moving when you do.

That's also, as far as any source checked could establish, the entire game. There's no second mechanic, no power-up, no obstacle beyond the AI opponent and an escalating level ladder that turns the screws by speeding the ball up and shrinking the margin for error as the stages climb. Judged as a piece of design built to fit inside a small, mouse-only Flash file with no server, no account system and no portal deal behind it, that's a reasonable, even elegant scope — one good idea, fully explored, nothing left half-built around it. Judged as a reason to keep playing past the first sitting, it's thin: once the curve trick is learned, the ladder's only remaining question is how many levels of increased ball speed a player can survive, and multiple independent period accounts — forum posters comparing how far they'd gotten, landing consistently in the level 6 to level 9 range — describe the later stages as close to a wall rather than a fair climb. A blog post from May 2003 calling it "insidiously addictive" and a boating-forum poster comparing it to Tetris a few months earlier both point to the same thing: the curve mechanic is strong enough to hook someone for a run at that ladder, even though the ladder itself isn't built to be won.

No opponent worth the mechanic it was built for

The single clearest missed opportunity here is that a game engineered entirely around out-manoeuvring an opponent's paddle angle never got a second human to actually out-manoeuvre. No source checked found any evidence of a two-player or online mode, local or otherwise — Curveball is, and apparently always was, one player against a fixed AI. That's not an unusual limitation for a 2002 Flash game with no server infrastructure behind it, and building real-time netcode for a free self-hosted release would have been a serious undertaking for what reads as a small, independent project. But it does mean the mechanic's most interesting implication — reading another moving paddle's angle in real time, the way real ping-pong spin play actually works — was never put in front of a player capable of bluffing back.

Playability today

The original pdmedia.net hosting is gone, and this review found no live portal page carrying an original, undisturbed version of the game. What survives is preservation-side: an Internet Archive item titled curveball_flash, uploaded on 28 October 2020 by the Flash Archive account associated with the Flashpoint preservation project, configured to run through Ruffle emulation — checked 5 September 2026. This review could not independently confirm that the Ruffle build renders and plays without glitches, since no bug report or compatibility note for this specific title turned up in either direction; the honest position is that it is configured to run, not that it has been verified to run cleanly. Beyond that single archive.org item, the game is also still served today as an embedded Flash emulation by a scattering of long-running mirror and retro-game sites (Springfrog, MaxGames and others), any of which likely depend on the same kind of in-browser emulation rather than a native Flash Player. Anyone searching for "Curveball" today should also expect to run into work that isn't this game at all: a 2022 Unity fan remake that explicitly disclaims any official connection to the original, a "Curveball 2" hosted at curve-ball.org with no credited developer or stated lineage, and a handful of generic "Curve Ball 3D" browser and mobile titles trading on the same name. None of those are covered by this review, and none should be mistaken for the 2002 original when judging what has or hasn't survived.

Final verdict

Curveball earns its place in the Flash canon on the strength of one idea, executed cleanly, in a game that reached a real audience without ever needing a portal's front page to do it — a distinct enough distribution story on its own that it's worth remembering alongside the mechanic itself. The curve-on-contact trick is a genuine, well-judged addition to Pong's ancient formula, and it's clearly what period players meant when they called the game addicting. But it's also the only thing here: no second system, no multiplayer for a design built around reading an opponent, and a difficulty ladder that by every independent account stops being a fair climb well before it ends. A 6.5 reflects a small, well-made single-idea game that never asked to be more than that, honestly limited by a scope that suited a free, self-hosted 2002 release and hasn't grown since.

Players interested in the wider Pong-descended lineage of Flash sports arcade games should look at Bubble Trouble for a different single-screen arcade format from roughly the same period, or Sports Heads: Football for a later, more elaborate take on the same "one screen, one opponent, one core trick" design shape.

Pros and cons

Pros

  • The curve-on-contact mechanic is a real, well-executed idea: paddle movement at the moment of the hit changes the ball's flight, turning a static reflex test into something with an actual skill ceiling
  • It reached a wide audience without a portal sponsorship or an exclusivity deal, spreading through direct links on forums and personal sites from as early as December 2002 — an unusually organic distribution story for a Flash game of its period
  • Free of any confirmed ad-funding or site-lock scheme; what survives points to a plain, unencumbered self-hosted release
  • Simple enough — one mechanic, mouse-only control — to load and understand in seconds, which suited exactly the audience finding it through a friend's link rather than a portal's front page

Cons

  • The curve mechanic is the only idea in the game; there is no second system, power-up, or variation introduced across the level ladder to develop it further
  • No two-player or online mode exists despite a mechanic built entirely around outmanoeuvring an opponent — the single-player AI ladder is the only way to play it
  • The difficulty curve is reported, consistently and independently, to end in a final stage that is effectively unbeatable rather than merely hard, which reads less like a deliberate "and now the true test" design choice and more like a difficulty curve nobody re-tuned once it stopped serving the player
  • No contemporary press review or Newgrounds-style vote tally exists to check reception against — what evidence survives is informal forum chatter, not a measured critical reaction
  • The developer's identity, "Joshua Skeels," rests on corroborating secondary sources rather than a primary credits screen or company record this review could directly inspect

Category scores

Category scores are the reviewer's read on each area. They inform the overall score but do not average into it — the number at the top of the page is a judgement, not a calculation.

  • Gameplay 7 / 10
  • Visuals 5.5 / 10
  • Progression 5.5 / 10
  • Originality 7.5 / 10
  • Replayability 5 / 10