A player reaches level 12 and quits. The pass rate on that level is 23%, while previous levels had 60%. A classic symptom of poor balance: the difficulty curve is too steep and monetization squeezes the wallet. This situation kills retention and lowers ARPU. According to our data, fixing such a mistake early increases LTV by 15–25%, which for a game with 50k DAU means $30,000–$50,000 monthly. We design balancing systems that prevent such failures. Our experience includes balancing for 50+ mobile projects – from hyper-casual to MMORPG. We have 8+ years of experience in game analytics and have shipped over 60 successful games.
Game balancing is the tuning of the progression curve and game economy, where the mathematical model determines mobile game monetization and retention. Each parameter – from enemy damage to potion price – affects these metrics. Without a systematic approach, even a small error in the difficulty growth coefficient can crash day-7 retention to 15%. Losing 5% of day-7 retention reduces LTV by 20%, which could cost $50,000 at 100k DAU.
Mathematical Progression Model in Game Balancing
The foundation of any balance is the progression curve. For RPGs, strategy games, and most casual games, power or exponential dependencies are used:
cost(n) = base_cost \times growth_factor^n
For example: base_cost = 100, growth_factor = 1.5. Then:
- Level 1: 100
- Level 5: ~759
- Level 10: ~5766
- Level 20: ~332,525
With growth_factor > 1.6, progression becomes too steep – the player hits a wall and either pays or quits. With < 1.3, it's too flat – no sense of achievement.
Level difficulty curve:
enemy_hp(level) = base_hp \times (1 + level \times difficulty_scale)
player_dps(level) = base_dps \times (1 + level \times power_scale)
The key parameter is the ratio difficulty_scale / power_scale. If the player gains power faster than difficulty grows (power_scale > difficulty_scale), the game becomes trivial by mid-game. If slower, a pay-wall appears. An error in this coefficient can cost up to 40% of future revenue.
Tooling and process
GameAnalytics / PlayFab Analytics
We analyze retention by level, pass rates, and first exit points:
// Log death with context
GameAnalytics.NewProgressionEvent(
GAProgressionStatus.Fail,
"world_1", "level_07",
score: remainingHP // health at death as difficulty proxy
);
// Log completion time
GameAnalytics.NewDesignEvent("Level:CompletionTime:world1_07",
(float)completionTime.TotalSeconds);
The score = remainingHP on Fail shows how close the player was to winning. remainingHP = 5% means "almost passed" – slightly reduce difficulty. remainingHP = 80% means "didn't even start" – significant balance gap.
Google Sheets / Airtable as balance sheet
Enemy parameters, items, abilities – in a table with formulas. Changing one cell recalculates all dependent values. Then JSON/CSV is exported into the game via Remote Config or Addressables.
Remote Config for hot-patching balance
Critical parameters (drop rates, store prices, damage multipliers) via Firebase Remote Config – changed without an update:
var remoteConfig = FirebaseRemoteConfig.DefaultInstance;
await remoteConfig.FetchAndActivateAsync();
float bossHealthMultiplier = (float)remoteConfig.GetValue("boss_health_multiplier").DoubleValue;
float goldDropRate = (float)remoteConfig.GetValue("gold_drop_rate").DoubleValue;
Economy in Mobile Game Balancing: Three Pillars
Sources of currency: quests, levels, daily bonuses, achievements, sometimes ads. They should be predictable – the player plans accumulation.
Sinks: upgrades, consumables, content unlocks, cosmetics. They must actively consume currency, otherwise it loses value.
Exchange rate: how much real time is needed to obtain a unit of in-game value without paying. This is the main monetization lever.
| Metric | Recommendation |
|---|---|
| Sources/Sinks ratio | ≥ 1.2 for a healthy economy |
| Time to next goal | 1–3 days without payment |
The economy is healthy if the time to achieve a goal without payment remains reasonable (1–3 days for the next significant goal), and paying speeds up but does not block progress. Exception: cosmetic-only monetization – there the balance is different. With a properly tuned economy, ARPU grows by 10–30%, which for a project with 50k DAU brings an additional $10,000–$30,000 monthly.
How does A/B testing improve balance accuracy?
Iteration must be fast. The scheme:
- Hypothesis: "Level 12 is too hard – pass rate 23%, target 55-65%"
- Change: Reduce enemy HP on the level by 20% via Remote Config
- Rollout: To 10% of audience (A/B test via Firebase)
- Measurement: After 3 days, check pass rate and retention in the experimental group
- Decision: If pass rate is 58% and retention hasn't dropped – roll out to 100%
Without A/B testing, balance iterations are blind flights. A change may improve one metric and kill another.
| Parameter | Recommended value |
|---|---|
| Pass rate per level | 55-65% |
| Day-7 retention | >35% |
| Time to first purchase | <1 hour of gameplay |
| Average session length | >5 minutes |
PvP and multiplayer: matchmaking and rating
In PvP games, balance is complicated by the matchmaking system. ELO-like systems (TrueSkill, Glicko-2) are used as a base, but with mobile constraints: you can't keep a player in queue for long. The usual compromise is a tight skill range in the first 15 seconds of queue, then wider range after:
float skillRange = Mathf.Lerp(50f, 300f,
Mathf.Clamp01(queueTime / maxQueueTime));
var opponent = MatchmakingService.FindOpponent(playerRating, skillRange);
Match data must be logged and analyzed: if win rate of top 10% players > 75%, the rating system is not working.
Common mistakes in PvP balance
- Too narrow skill range – queues >30 sec.
- Ignoring ping – leads to negative experience.
- Not accounting for team composition (e.g., 5 tanks without healer).
What's included in the work
We deliver comprehensive documentation and provide ongoing support:
- Audit of current analytics data: retention by levels, drop points, economic flows
- Development or audit of the mathematical progression model
- Setup of analytics events for balancing data (custom events in GameAnalytics/PlayFab)
- Moving balance parameters to Remote Config / Addressables for hot-patching
- Designing a balance sheet with dependencies (Google Sheets or Airtable)
- Setting up A/B testing for iterations (Firebase)
- Matchmaking recommendations for PvP (if applicable)
- Final report with all findings and recommendations
- 1-month post-launch support and iteration
Timeline
Audit and balance recommendations for an existing game: 1 day. Full design of a balance system from scratch plus analytics: 2–4 weeks. Cost is calculated individually, but as an example, an audit for a mid-core game costs $3,000 and typically identifies issues that can boost ARPU by 15%, adding $15,000 monthly at 100k DAU. A full custom balance system starts at $15,000. On average, a project pays for itself in 2 months of ARPU growth, bringing an additional $15,000–$25,000 monthly at 100k DAU.
We recommend Game balance as a starting point for study.
Contact us for an audit of your project. Order balancing now – we guarantee quality based on experience from 50+ successful projects and 8 years in the industry.