<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[Angry Old Wizards Dev Log]]></title><description><![CDATA[Angry Old Wizards Dev Log]]></description><link>https://angryoldwizards.hashnode.dev</link><generator>RSS for Node</generator><lastBuildDate>Wed, 18 Nov 2026 13:17:21 GMT</lastBuildDate><atom:link href="https://angryoldwizards.hashnode.dev/rss.xml" rel="self" type="application/rss+xml"/><language><![CDATA[en]]></language><ttl>60</ttl><item><title><![CDATA[Building alliance co-op for a mobile strategy game with a team of almost nobody]]></title><description><![CDATA[By Cinder, developer at Angry Old Wizards. Disclosure: this is about our own game, Crownfall: Kingdoms, a free iPhone kingdom strategy game. Alliances are…]]></description><link>https://angryoldwizards.hashnode.dev/building-alliance-co-op-for-a-mobile-strategy-game-with-a-team-of-almost-nobody</link><guid isPermaLink="true">https://angryoldwizards.hashnode.dev/building-alliance-co-op-for-a-mobile-strategy-game-with-a-team-of-almost-nobody</guid><category><![CDATA[GameDev]]></category><category><![CDATA[Game Development]]></category><category><![CDATA[iOS]]></category><category><![CDATA[Indie]]></category><category><![CDATA[Mobile Development]]></category><dc:creator><![CDATA[Cinder (dev)]]></dc:creator><pubDate>Wed, 18 Nov 2026 05:05:44 GMT</pubDate><content:encoded><![CDATA[<p><em>By Cinder, developer at Angry Old Wizards. Disclosure: this is about our own game, Crownfall: Kingdoms, a free iPhone kingdom strategy game.</em></p>
<p>Alliances are the heart of a persistent kingdom game. They're why people come back on day 30, and they're the scariest system for a tiny studio to build. Alliance play is social, concurrent, and full of edge cases, and when it breaks, it breaks in front of twenty people at once.</p>
<p>We launched Crownfall: Kingdoms on iPhone on November 10, 2026, with alliances, rallies, and <strong>Epic Battles</strong>: timed, multi-phase co-op fights that a whole alliance takes on together. Here's what we'd tell another small team, ordered by how much pain each lesson would have saved us.</p>
<h2>1. Decide what "co-op" means before writing any code</h2>
<p>"Alliance features" can mean twenty different things. We narrowed it to three verbs and built only what supported them:</p>
<ul>
<li><strong>Talk.</strong> Alliance chat, roles, and recruitment. Without a place to coordinate, the other features don't matter.</li>
<li><strong>Gather.</strong> Rallies, where members pool effort against a target that's too strong for one kingdom.</li>
<li><strong>Endure.</strong> Epic Battles: enemies with several phases and a timer, designed so that <strong>preparation and sustained contribution</strong> beat one lucky strike.</li>
</ul>
<p>Everything else (alliance tech trees, shared buildings, territory) went on the "later, maybe" list. A small team that ships three solid alliance verbs beats one that ships eight broken ones.</p>
<h2>2. Make the server the only source of truth</h2>
<p>In a solo game you can trust the client a bit. In co-op you can't trust it at all. Every attack, every point of damage, and every reward in an Epic Battle is decided on the server. The phone sends what it wants to do and shows what the server says happened.</p>
<p>This costs some responsiveness. There's a small wait between tapping and seeing the result. It saves you from the problem that ends small multiplayer games: two players' phones disagreeing about whether the boss is dead. Pick server authority on day one. Retrofitting it later is a rewrite.</p>
<h2>3. Assume every request will be sent twice</h2>
<p>Mobile networks drop requests constantly. A player taps Attack in an elevator. The request reaches the server, the response doesn't reach the phone, the app retries, and now the attack has happened twice.</p>
<p>The fix is simple, and we wish we'd started with it:</p>
<ul>
<li><strong>Every action gets a unique ID made on the client.</strong> The server stores it. A repeated ID returns the original result instead of running the action again.</li>
<li><strong>The app tells the truth about uncertainty.</strong> If we don't know whether an attack landed, the app says so ("Attack not confirmed. Retry safely.") and the retry is safe by design.</li>
</ul>
<p>This one pattern removed a whole category of "I got charged twice" and "my damage vanished" reports before they happened.</p>
<h2>4. Realtime updates should mean "go check," not "here's the new state"</h2>
<p>Co-op screens want to feel live. When an alliance-mate lands a hit, your health bar should move. The tempting design is to broadcast the new state (HP, phase, rankings) to everyone watching.</p>
<p>We do the reverse. A realtime message is just a <strong>signal to re-fetch</strong> from the server. Broadcasts can arrive late, twice, or out of order. If they carry state, an old message can overwrite a newer one, and the bar jumps backwards. If they only say "something changed," the worst case is an extra fetch.</p>
<p>We also number our own requests, so a slow, older response can't overwrite a newer one that already came back. It's a small detail, and it prevents the most visible kind of co-op bug: a health bar that jumps backwards when several people attack at once.</p>
<h2>5. Time-box everything, and show the clock</h2>
<p>Epic Battles are timed, and so is our world boss. Timers do three jobs for a small studio:</p>
<ul>
<li><strong>They create the "now" that co-op needs.</strong> "Tonight, before the timer runs out" coordinates people better than "whenever."</li>
<li><strong>They limit the damage from bugs.</strong> A mistake in a timed encounter ends when the encounter does.</li>
<li><strong>They make content last.</strong> A multi-phase fight with a deadline can be replayed and tuned. A permanent one just gets solved.</li>
</ul>
<p>Show the clock everywhere the encounter appears. Our event cards show the start time in a fixed zone ("9:00 AM ET") and a live countdown, because "starts at 9" means nothing to a player in another time zone.</p>
<h2>6. Don't make solo players second-class</h2>
<p>The biggest design call for our first live event: <strong>the world boss is open to every kingdom, including players with no alliance.</strong> Alliance play is our core, but a brand-new player who hasn't found an alliance yet shouldn't be locked out of the one thing everyone is talking about.</p>
<p>The event becomes a lobby. Solo players fight the same boss as the big alliances, see the same worldwide attack count, and get a reason to look for an alliance afterwards. Recruiting happens naturally instead of being forced.</p>
<h2>7. Test the co-op flow with fewer people than you expect</h2>
<p>Small studios rarely have twenty testers on call. You can find most co-op bugs with three accounts and a script:</p>
<ul>
<li>two players attacking the same target in the same second;</li>
<li>one player leaving the alliance during a rally;</li>
<li>a phone going offline mid-attack;</li>
<li>a phase changing while a request is still in flight.</li>
</ul>
<p>Give every ugly case a repeatable database test, then run them on every change. You don't need a big QA team. You need the four or five races that matter, written down and automated.</p>
<h2>8. Ship it, then listen to the chat</h2>
<p>We're a week past launch. Our alliances are small, our Epic Battles are still being tuned, and our first live event has just started. Analytics tell you <em>where</em> players stop. Alliance chat is where they explain <em>why</em>, in their own words. If you build chat first (lesson 1), you get a free, honest, always-on playtest. Read it.</p>
<h2>The short version</h2>
<ul>
<li>Three alliance verbs, done well: talk, gather, endure.</li>
<li>The server decides everything.</li>
<li>Every action has an ID; retries are safe.</li>
<li>Realtime means "re-fetch," not "here's the state."</li>
<li>Timers and fixed time zones on everything.</li>
<li>Let solo players into the big moments.</li>
<li>Automate the four or five races that matter.</li>
</ul>
<p>If you're building co-op on a tiny team and want to compare notes, I'm happy to. Reach me at <a href="mailto:crownfall@angryoldwizards.com">crownfall@angryoldwizards.com</a>.</p>
<hr />
<p><em>Crownfall: Kingdoms is free on iPhone (iOS 16.4+), no ads, with optional gem purchases. <a href="https://apps.apple.com/us/app/crownfall-kingdoms/id6830826391">App Store</a> Â· <a href="https://angryoldwizards.com/">Angry Old Wizards</a></em></p>
]]></content:encoded></item><item><title><![CDATA[Designing a 31-day world boss where the best reward can't be bought]]></title><description><![CDATA[By Cinder, developer at Angry Old Wizards. Disclosure: this is about our own game, Crownfall: Kingdoms, a free iPhone kingdom strategy game with optional…]]></description><link>https://angryoldwizards.hashnode.dev/designing-a-31-day-world-boss-where-the-best-reward-can-t-be-bought</link><guid isPermaLink="true">https://angryoldwizards.hashnode.dev/designing-a-31-day-world-boss-where-the-best-reward-can-t-be-bought</guid><category><![CDATA[GameDev]]></category><category><![CDATA[Game Development]]></category><category><![CDATA[iOS]]></category><category><![CDATA[Indie]]></category><category><![CDATA[Mobile Development]]></category><dc:creator><![CDATA[Cinder (dev)]]></dc:creator><pubDate>Wed, 18 Nov 2026 05:04:38 GMT</pubDate><content:encoded><![CDATA[<p><em>By Cinder, developer at Angry Old Wizards. Disclosure: this is about our own game, Crownfall: Kingdoms, a free iPhone kingdom strategy game with optional gem purchases.</em></p>
<p>On November 17 we started our first live event: <strong>The Eclipse King Rises</strong>, a world boss that every kingdom in Crownfall fights together for 31 days. We're a tiny studio. Our audience is small and new, and we really didn't want the first event to tell players "the people with the biggest wallets win."</p>
<p>This post covers the one decision that shaped everything else: <strong>we capped hits, not damage.</strong> I'll explain why, what it costs, and the numbers, so you can steal the idea or avoid our mistakes.</p>
<h2>The problem with "most damage wins"</h2>
<p>Most world boss events in mobile strategy games pay out by damage. That's simple, readable, and fits the fantasy. It also means the leaderboard is mostly a power ranking. In a game where purchases speed up growth, a power ranking is also partly a spending ranking.</p>
<p>That's fine for a leaderboard. People expect the strongest kingdoms at the top, and you can't tell a strategy-game player their power doesn't count. It's a bad fit for the <strong>reward most players actually want</strong>: a cosmetic that says "I was there for the first event."</p>
<p>So we split the event into two tracks:</p>
<ol>
<li><strong>Damage</strong>: shown on a worldwide top-10 leaderboard, with your own rank under it. Power matters here, as it should.</li>
<li><strong>Hits</strong>: the number of successful attacks you land. This is what unlocks the event's exclusive reward, two Founder avatar skins at <strong>1,600 hits</strong>.</li>
</ol>
<h2>Why a hit cap is the fairness lever</h2>
<p>A hit count alone doesn't solve anything. If hits were unlimited, the player who refills forces fastest would hit 1,600 on day two. The fairness comes from the cap: <strong>each player can land at most 10 successful hits per hour.</strong> The server enforces it, and nothing in the store raises it.</p>
<p>That one rule changes what the event rewards:</p>
<ul>
<li><strong>The pace is the same for everyone.</strong> The most hits anyone can land in 31 days is 31 Ã— 24 Ã— 10 = 7,440, and nobody can go faster.</li>
<li><strong>1,600 hits is about 160 capped hours out of 744.</strong> That's roughly 21% of the theoretical maximum. In practice it's about 52 hits a day, or five or six check-ins a day. Players who show up regularly get it; no single session can.</li>
<li><strong>Late joiners can still make it.</strong> Someone who starts a week in still has around 24 days, which works out to about 67 hits a day. That's harder, but possible, and we wanted it possible.</li>
</ul>
<p>The reward is basically a measure of <strong>showing up</strong>, the one thing in a free-to-play game that money doesn't buy.</p>
<h2>The trade-offs (so you don't think it's free)</h2>
<p><strong>It can feel like a chore.</strong> Capped, hourly resources can become "log in or lose out." We tried to soften this in three ways:</p>
<ul>
<li>The cap is generous enough that you don't need to play every hour.</li>
<li>The target is well under the maximum, so skipping nights is fine.</li>
<li>Every hit pays out on its own, so a short session never feels wasted (more on that below).</li>
</ul>
<p><strong>Strong players get less to chase.</strong> A high-power kingdom can't turn its power into the cosmetic any faster. We accepted that and gave them the damage leaderboard instead. It's visible, it's competitive, and it's where power is supposed to show.</p>
<p><strong>It's more server work.</strong> A per-player rate limit on a shared, contested resource is the kind of thing that breaks under retries and flaky mobile connections. If a phone sends an attack, loses signal, and sends it again, you must not count it twice or charge the player's hourly budget twice. We gave every attack request a unique ID so a retry is safe. The app also says "Attack not confirmed. Retry safely." instead of failing silently. Getting that one message right was worth the effort.</p>
<h2>Making every hit worth something</h2>
<p>A long event with a far-off reward needs short-term rewards, or people stop by day four. So every successful hit pays:</p>
<ul>
<li><strong>Gold, every time.</strong> We tuned it so that Fight (with soldiers) and Assassinate (with spies) pay the same gold at the same kingdom investment. We didn't want one army type to be the obviously correct choice.</li>
<li><strong>A small chance at a rare drop:</strong> 1â€“100 gems or 1â€“10 health crystals. Never both on one hit. These are the same gems people can buy, so a free player who keeps showing up gets some premium currency without paying.</li>
</ul>
<p>We describe the rare drop in the app as a <em>chance</em>, not as something that happens every hit, and we did the same in the App Store event copy. If you run a drop like this, be careful about that wording. Players notice when "win gems!" really means "occasionally win gems."</p>
<h2>One shared bar, many phases</h2>
<p>Everyone hits the same Eclipse King. The health bar is huge on purpose (998 billion HP at launch): the community-scale number is part of the fantasy. When the server empties the bar, the phase counter goes up and the fight continues. The event runs its full 31 days either way. It doesn't end early just because a few strong kingdoms burned through it, which would have killed the hit track for everyone else.</p>
<p>The app shows the live HP, the worldwide attack total, the current phase, and your own hit and damage totals in one panel. In an event this long, "you've done 412 of 1,600" is the most motivating number on screen. Don't hide it behind a menu.</p>
<h2>What we'd tell another small team</h2>
<ul>
<li><strong>Pick which reward has to be fair, and protect that one.</strong> Not every reward has to be equal. Ours: damage can follow power; the cosmetic can't follow spending.</li>
<li><strong>Cap the action, not the outcome.</strong> A hits-per-hour limit is easy to explain and easy to check on the server.</li>
<li><strong>Set the target well under the cap.</strong> About 20% of the max worked out to "regular, not obsessive" for us. Check your own retention data before copying that number.</li>
<li><strong>Pay out every action.</strong> A 31-day goal needs a reason to show up today.</li>
<li><strong>Make retries safe before launch.</strong> Players will tap Fight twice on a train.</li>
<li><strong>Say "chance" when it's a chance.</strong></li>
</ul>
<p>We'll write a follow-up after the event ends on December 18 with what actually happened: how many players reached 1,600, where they dropped off, and whether the cap felt fair or felt like a chore. If you've run a capped event and have opinions, I'd genuinely like to hear them.</p>
<hr />
<p><em>Crownfall: Kingdoms is free on iPhone (iOS 16.4+), no ads, with optional gem purchases. The Eclipse King Rises runs until December 18, 2026, 9:00 AM ET. <a href="https://apps.apple.com/us/app/crownfall-kingdoms/id6830826391">App Store</a> Â· <a href="https://angryoldwizards.com/">Angry Old Wizards</a></em></p>
]]></content:encoded></item></channel></rss>