What's in the launch build

The launch ships with the full core loop: strategic tower placement, multiple tower types with upgrade paths, and an endless wave structure that scales indefinitely. Boss waves break the rhythm with high-HP single targets that punish splash-only setups, and swarm waves do the inverse — punish single-target-only setups with sheer numbers. Mouse and touch controls are both first-class, and the build runs cleanly on mid-range mobile.

Why TD fits the ArcadeDeck format

Most browser TDs are either too short to feel meaningful or too long to fit a casual session. We tuned wave pacing so that the first ten waves take 4-6 minutes — enough to learn the map, commit to a tower mix, and see whether your read was right. If you misplay the early game, you feel it by wave 8 and can restart without losing 30 minutes. If you play well, the run scales up cleanly past wave 20 and the leaderboard becomes a question of how far you can push before a single mistake compounds.

Opening strategy: the first ten waves

Three habits separate first-session players (typically peaking around wave 5-7) from players who reliably clear wave 10 on their second or third attempt.

1. Read the chokepoint before you place anything

The map has one or two natural chokepoints depending on enemy path. Don't drop your first tower until you've identified them. A first tower placed at the chokepoint kills more enemies than two towers placed mid-path. The early-game economy is tight — a single misplaced tower in wave 1 is the difference between affording an upgrade by wave 4 or scraping through to wave 6 under-levelled.

2. Mix tower types early, not late

The most common new-player mistake is going all-in on one tower type because it felt strong on wave 2. Single-target towers melt tanks and starve on swarms; splash towers shred swarms and stall on tanks. If you reach wave 6 with a pure single-target setup, the next swarm wave will leak. Commit to at least one of each type by wave 4, even if it means delaying an upgrade.

3. Save resources for emergency upgrades

Don't spend down to zero between waves. Keep a reserve roughly equal to the cost of a mid-tier upgrade. When a boss wave starts and you realise your damage is short, that reserve is the difference between a clutch upgrade mid-wave and a leaked life. The leaderboard rewards Peak Wave above all else — surviving wave 12 by one HP is worth more than dying cleanly on wave 11 with a full bank.

What we're watching post-launch

Two things on our radar over the next two weeks. First, wave-progression telemetry: where do players consistently die, and is the curve smooth or does one specific wave act as a wall? If we see a spike at, say, wave 14, we'll rebalance that wave specifically rather than flatten the whole curve. Second, tower-mix data: are players converging on one optimal mix, or is there genuine strategic variety? If everyone runs the same loadout by wave 10, that's a tuning problem and we'll address it in a patch note here.

Surviving wave 12 by one HP is worth more than dying cleanly on wave 11 with a full bank.

Get on the board

The leaderboard is empty as of launch — early entries will set the bar for the April-May Hall of Fame cycle. If you crack the top three this week, expect a feature in the next monthly recap. Bug reports and tuning feedback go to help@arcadedeck.net; the most useful reports are wave-specific ("wave 13 felt unfair because…") rather than general impressions.

See you on the leaderboard.

Hardware

You need a USB or Bluetooth keyboard with full N-key rollover. Membrane keyboards typically support only 2-3 simultaneous key presses, which becomes a problem in a four-player game where everyone is mashing direction keys at once. If your keyboard ghosts, players will lose inputs and accuse each other of cheating. A cheap mechanical board solves this for under $40.

Bracket structure

For 4 players, run a round-robin: every player faces every other player once in 1v1 mode, then a final 4-player free-for-all. With six 1v1s plus one final, total runtime is ~25 minutes for typical match lengths. Award 3 points for a 1v1 win, 5 points for the FFA winner, 3/2/1 for second/third/fourth in the FFA. Highest total wins.

For 5-8 players, split into two pools of 3-4, run round-robin within each pool, then take the top two from each into a four-player FFA final. Keeps total runtime under 40 minutes.

Map and rule selection

Triangle Battle has three arena layouts and four hazard settings. For tournament play, we recommend:

• Arena: Symmetric Wedge. The other two layouts have spawn-point asymmetries that cumulate over a long set. Wedge is the only fully mirrored map.
• Hazards: Off for the round-robin, On for the FFA. Hazards spice up the final without polluting head-to-head results with random deaths.
• Best of one per matchup. Best-of-three drags total runtime past 45 minutes and tournament fatigue starts costing players inputs.

Common arguments and how to pre-empt them

"They got the better starting position." Coin-flip starting positions before each round. We use a physical coin because it ends arguments faster than a digital one.

"My key didn't register." The N-key rollover keyboard solves this almost entirely. If a key truly didn't register and you can demonstrate it by replaying the input, replay the round. Use this rule sparingly or it gets abused.

"Final FFA isn't fair to the leader." The 5/3/2/1 weighting means the round-robin leader is favoured but not guaranteed. We've kept the format because the tension is the point — close finishes happen most when the late game can still flip the standings.

Submit your tournament

If you run a Triangle Battle tournament with 6+ players, send results plus a short write-up to help@arcadedeck.net. We feature the most interesting brackets in the monthly Hall of Fame recap.

Here are the five patterns we see consistently in top-five replays. Each one is a cheap habit to practise; together they shave 30-60 moves off a typical full-board clear.

1. Reserve the corners

Corner cells only have two neighbours, so any gem you place there can only contribute to two potential merges. Keep corners empty until the late game and use them as overflow slots for low-tier gems you don't yet want to commit to a chain.

2. Read the queue, not the board

Beginners place each new gem next to the matching colour they can see. Top players pre-plan two to three placements ahead by reading the queue. If you see a tier-2 red incoming after the current tier-1 blue, dropping the blue away from the red prep zone is almost always the right call.

3. Build chains diagonally

Triangular merge layouts cascade better than linear ones. A diagonal stack lets a single high-tier merge resolve into the surrounding tier-1 cluster you've been seeding. Linear stacks tend to dead-end after one promotion.

4. Don't fight the RNG

Sometimes the queue gives you four blues in a row. Place them. Forcing a chain of greens by leaving blues in awkward spots costs more moves than just absorbing the blue run and resetting your plan. The leaderboard punishes wasted shuffles, not flexible plans.

5. End on a tier-jump, not a clear

If you can finish the round by promoting an existing tier-4 to tier-5 instead of clearing a half-merged cluster, take the promotion. Tier jumps are worth more on the primary score and don't add to your move count. Several record runs end with a single decisive promotion rather than a full-board sweep.

Putting it together

Try one pattern per session. Don't try to internalise all five at once — the cognitive load makes you play worse, not better. Most players see a 20-30% move reduction after a week of focusing on diagonals alone.

Most players see a 20-30% move reduction after a week of focusing on diagonals alone.

If you climb into the top ten this month with a strategy not on this list, send us a replay link and we'll feature it in the next post.

Three kinds of posts will land here:

Patch notes

Every meaningful balance change, content drop, or bug fix on a published game gets a short write-up. We will explain what changed and why, and where relevant we will share the data that drove the decision (heatmaps from Galaxy Launch, win-rate splits from Spirit Duel, click-distribution charts from 15s Click Challenge). If a game gets harder or easier, we want you to understand the reasoning rather than guess.

Design deep-dives

How do you tune a roguelike's difficulty curve when every run is finite? How do you keep an idle game from collapsing into a cookie-clicker grind? These are the conversations we have internally every week — and they tend to be more interesting than "version 1.4 is live." Expect a long-form post roughly once a month, focused on one specific design problem in one specific game.

Hall of Fame recaps

At the end of every month we crown the top three players across the entire ArcadeDeck catalogue. The Hall of Fame page archives the names; this devlog will tell the stories. How did the leaders earn their points? Which games trended this month? Were there any controversies on the leaderboards? If you finished in the top ten anywhere, expect a shout-out.

Why we're doing this

Most browser game portals are content aggregators. ArcadeDeck is the opposite — every game is built in-house, and the platform reflects a particular set of opinions about what makes web games fun. Sharing those opinions in writing keeps us accountable, gives players a clearer view of where the platform is heading, and (we hope) starts conversations with developers and players who care about the same things.

See you on the leaderboards.

How the points broke down

Of the 16 playable games on ArcadeDeck during March, the leader board contributed points across nine of them. The biggest single contributors to the top three were Galaxy Launch (idle, three first-place finishes), GEM MERGE (puzzle, two first-place finishes including one record-setting low move count), and Triangle Battle (multiplayer, contested across many sessions).

Playerkt's 12-point total came from a remarkable five top-three finishes spread across genres — including a top-three on the deeply technical Polygon Invasion TD, which most players bounce off in their first session. GG concentrated on action and arcade titles (Gate of Hell, 15s Click Challenge, Skyfall), and BEN posted a strong showing on the strategy side (Kingdom's Last Stand, Endless War, Spirit Duel).

Trends we noticed

March's leaderboards skewed heavily toward arcade and casual titles. The longer-form strategy and roguelike games (Dicefall Chronicles, Heroes of the Last Nexus) had top-five activity but lower volume — typical for genres where a single run can run 30-60 minutes. We expect April to swing more strategy-heavy as players who finished March pivot toward longer sessions for the second monthly cycle.

The only game with no scored entries this month was 40th Century: Void Survivor's hardcore mode, which remains the hardest single ranking on the platform. If you crack the top 10 there, expect a featured profile in the next recap.

Playerkt's 12-point total came from five top-three finishes spread across genres — including a top-three on the deeply technical Polygon Invasion TD.

Spotlight: GEM MERGE move-count record

Playerkt's GEM MERGE first-place run finished a full board with 27 fewer moves than the previous record. We pulled the replay and the route is unusual — heavy diagonal stacking with deliberate corner reservations, which lines up with several of the patterns covered in our recent strategy post. If you're chasing the GEM MERGE leaderboard in April, that replay is worth studying.

April outlook

The April cycle is live. New monthly archives are added on the first of each month, and last month's standings stay on the Hall of Fame page indefinitely. If you didn't finish in the top three in March, the cycle resets cleanly — every player starts April at zero points.

Browse the live standings on the Hall of Fame page. See you in the next archive.