How I made $1,000,000 with my First Game

Storyline / Setting

• The speaker’s game, Tang GTD / Tang GTD2 (later released on Steam as Tang GTD), is a tower-defense / lane-defense game set in themed fantasy environments.
• Its presentation evolved over time:
  • Early: an “orange” identity
  • Later: chibi/hero characters and a more RPG-like presentation
• The core premise shifted multiple times during development:
  • Early on: enemies attack the player’s units, and you must block/position correctly.
  • Later prototypes: enemies behave more like Bloons-style forward movement, with units punished on late rounds.
  • Final direction: a clearer loop where enemies are the pressure, and player defenses/jobs create survivable progress, including meta progression (farm gold, upgrade outside levels).

Gameplay Loop (as it Evolved)

• Place units/towers
  • Early builds: sometimes restricted by slots
  • Later: mostly free placement
• Earn currency over time/combat, then spend it to:
  • place more units,
  • upgrade/level heroes,
  • gain abilities via skill trees / items,
  • and later, add meta progression between runs.
• Survive multiple rounds
  • Winning: level up / choose upgrades
  • Losing: still gain some progress via meta systems in later versions
• Boss rounds appear in demo versions, and the game starts to resemble a Bloons-like progression, with deeper build customization.

Key Gameplay Highlights Mentioned

• Tactics/automation system (early + intermediate versions)

  • Units could trigger abilities based on conditions (e.g., health thresholds, status effects, cooldown rules).
  • Later, the speaker judged this approach too confusing and unworkable for clear, fast tower-defense gameplay.

• Enemy pathing improvements

  • Uses Bezier curves for smoother enemy movement (the speaker still recommends this approach).

• Team/hero systems

  • Early builds required selecting heroes into a “team” before placing them.

• Abilities vs. items rework

  • Early: abilities chosen from skill trees and/or paid with money.
  • Mid: abilities/conditions became too complex.
  • Late: abilities are tied to items, making upgrades feel immediately powerful and easier to combine.

• Zoom/clarity and UI rewrites

  • Multiple UI redesigns to fix:
    • confusion and cramped menus,
    • unclear “what can I do now?” moments,
    • too many small/overlapping elements.

Development Timeline & Major Strategy Decisions (High-Level)

• 2021 (engine + first playable prototypes)

  • Built a custom rendering engine using Vulkan, then made small clones:
    • Vulkan “introduction” (learning vertices/quads)
    • a Pong-like prototype
  • Developed early TD foundations:
    • text rendering, quad rendering, textures,
    • early enemies following waypoints,
    • early pixel art (blurred due to linear filtering).

• Late 2021 – Early 2022 (tower-defense becomes visible)

  • Added pixel art and started custom art pipelines.
  • Built an editor and level tools early (later described as huge value).
  • Noted that engine work created major overhead (busy-work like renderer/asset uploading).

• 2022 (first demo direction)

  • Gameplay loop prototypes:
    • unit placement (often limited by slots),
    • basic tower upgrade/level-up flow,
    • early AI tactics menu (still too hard to understand).

• 2022 – 2023 (itch.io + “prototype bloat” / Unity-like problems)

  • Reworked the game repeatedly, but players struggled with:
    • how to earn money,
    • how abilities unlock,
    • what to do next.
  • Pain point: too many clicks/menus for upgrades versus Bloons/KR.
  • Engagement issues: few wishlists and almost no Steam traction.

• Late 2023 – 2024 (Steam attempts + bigger reworks)

  • Changes accidentally harmed onboarding:
    • Removed placement slots and allowed placing “everywhere,” causing UI/clarity issues and drag/drop placement confusion.
    • Made skills cost money at some points (later called this “dumb” and unfun).
  • Added meta progression ideas and aimed for a more “5 minutes of fun” experience.

• 2025 (rendering + clarity + feedback-driven iteration)

  • OpenGL → DirectX
    • OpenGL: cross-machine issues (notably bloom/visual bugs) and weak error reporting.
    • DirectX: stabilized the game and produced actionable errors, including improved logging.
  • Community-feedback-driven loop improvements:
    • make threats (enemy attacks) feel central,
    • reduce confusion around training vs using skills,
    • improve “obviousness” and UI feedback,
    • faster map building using splines for road creation.

• 2025 Oct (Steam Next Fest target)

  • Goal: 1,000 wishlists during Steam Next Fest.
  • Big appeal changes:
    • global skill tree
    • items now grant abilities (removing abilities from confusing separate systems)
    • stronger instant gratification when upgrading

Strategies / Key Tips the Speaker Emphasizes

• Build the gameplay loop first, then iterate with feedback bursts

  • Make an MVP loop, release it, gather feedback, then ship meaningful patches.

• Avoid polishing systems nobody understands

  • Example: tactics menus/skill conditions were technically cool but not understandable enough.

• Reduce clicks and menu complexity

  • Upgrades should feel like “tap Upgrade,” not deep lists with hidden conditions.

• Make progression feel real

  • Early mistake: enemy scaling prevented meaningful outgrowth, making progression feel pointless.

• Design for “instant power fantasy”

  • Later: items visibly change units and add on-hit effects to create explosive combos.

• Test map-building speed

  • Road/map tools are critical; splines let the dev test gameplay without waiting on art.

• Choose the right renderer/toolchain for iteration stability

  • Vulkan was educational but painful; OpenGL caused cross-machine issues; DirectX enabled faster debugging and consistent visuals.

• Presentation matters (appeal beats theme-only ideas)

  • The “orange” theme wasn’t enough; the game evolved to more attractive chibi/hero visuals.

• Packaging/communication for demos and outreach

  • Itch.io pages needed clearer calls-to-action (wishlists/feedback forms).
  • Outreach emails should stay simple to avoid spam concerns.

“Important Fixes” Discovered from Player Feedback (Examples)

• Skills/abilities

  • Problem: players didn’t realize they had to train abilities before using them.
  • Fix: move abilities onto items and simplify the upgrade model.

• Placement clarity

  • Problem: free placement (“everywhere”) caused onboarding/UI confusion.
  • Fix: placement slots (or clearer placement restrictions) to prevent wrong builds.

• Enemy interaction clarity

  • Problem: players weren’t sure what would be attacked and when.
  • Fix: improve UI/visual feedback.

• Reward pacing

  • Problem: players weren’t receiving items quickly enough in early demos.
  • Fix: later builds improved the drop/earning flow.

• Enemy pressure timing

  • Problem: abrupt unit deaths felt bad.
  • Fix: the final loop makes enemy attacks a clearer, earlier, more teachable threat.

Sources / Gamers / References Mentioned (At the End)

• Churno (referenced for an application-layer event concept)
• Handmade Hero (inspiration)
• Jonathan Blow (inspiration)
• Stu Valley (visual inspiration for style/grass and art)
• Darkest Dungeon (inspiration for team selection / condition-based RPG flavor)
• Attack by more than N and similar condition systems (referenced conceptually)
• Bloons / Kingdom Rush (used for comparison)
• Vampire Survivors (named for upgrade systems / reroll-style inspiration)
• Primagen (raid and a Twitter repost referenced)
• Tommy (developer of Gnomes, discussed game design advice)
• Thornity (developer referenced; mentions Click and Conquer)
• Mr. Ramos (feedback strategy: write exact to-do steps with time boxes)
• AQ / “aq” (help with lighting/glow/shader)
• null exception jet (helpful feedback that drove the “abilities on items” redesign)

Explicitly Mentioned Creator Examples

• A Russian/Polish person who played the game in Next Fest (content creator mentioned, not named in subtitles).
• The video also references an earlier “YouTube guy” who rage-played the Steam demo (no name given in subtitles).