Skip to content

Case study: online gaming platform eliminates 89% of bonus abuse with device hardening

How a gaming platform reduced bonus abuse by 89% using DeepID's device fingerprinting, emulator detection, and root detection. $3.6M annualized savings.

2 min readDeepID teamCase studiesUpdated 15 March 2026

In this article

The Challenge

An online gaming and fantasy sports platform with 25M+ registered users was losing $4.1M annually to bonus abuse and multi-accounting. Fraudsters created hundreds of accounts using emulators, app cloners, and device farms to exploit welcome bonuses, referral rewards, and promotional credits.

The platform's existing defenses — email verification, phone number checks, and IP-based rate limiting — were trivially bypassed. Attackers used disposable email services, virtual phone numbers, and residential proxies.

The Scale of the Problem

Device farms: Investigation revealed organized operations running 50-200 emulator instances simultaneously, each creating accounts and claiming bonuses. A single fraud ring extracted $180K in promotional credits over 3 months.

App cloners: Individual users ran 5-15 cloned instances of the app on a single rooted device, each with a separate account claiming referral bonuses.

Referral loops: Fraudsters created circular referral chains where Account A referred Account B, which referred Account C, which referred Account A — all on the same device or emulator farm.

The DeepID Implementation

Persistent device fingerprinting: DeepID creates a device identifier that persists across app reinstalls, data clears, and even factory resets. When the same device creates multiple accounts, they're linked automatically.

Emulator and app cloner detection: DeepID detects BlueStacks, Nox, LDPlayer, Genymotion, and 15+ other emulator platforms. App cloner detection identifies Parallel Space, Dual Space, and custom cloning tools.

Root detection with Magisk coverage: Multi-layered root detection catches the rooted devices that app cloners require. Even Magisk with MagiskHide is detected through behavioral and filesystem analysis.

Device graph analysis: DeepID builds device relationship graphs that expose referral loops and coordinated multi-accounting. Devices sharing network patterns, timing patterns, or hardware characteristics are linked.

Results After 120 Days

89% reduction in confirmed bonus abuse. Monthly bonus fraud dropped from ~$340K to ~$37K.

$3.6M annualized savings from eliminated promotional credit abuse, reduced investigation workload, and recovered referral program integrity.

Multi-accounting detection accuracy: 97.2%. The combination of device fingerprinting + emulator detection + root detection produced significantly fewer false positives than IP-based or email-based detection.

Referral program ROI improved by 215%. With fraud removed, the referral program's actual cost-per-acquisition matched projections for the first time.

Legitimate user impact: <0.1%. Only users running emulators or root/cloning tools were affected. The platform offered an appeal process for the rare false positive.

More from the blog.

Dated articles on the rules, the attacks and how DeepID is built.

  • One decision id, three places

    Every DeepID decision carries one id: the same in the app that acts on it, in the event your FRM receives and in the decision log an auditor reads. Here is why, and how it is wired.

    2 min read

  • SMS OTP is not banned. Three things that are changing

    No regulator in DeepID’s markets has banned the SMS one-time password. What is changing is where it is enough: at login, for high-risk transactions, and as the floor under every payment.

    2 min read

  • Why a coached payment passes every check

    In a coached payment the real customer signs in on their own phone and approves the payment themselves. Each check does its job, and the money still leaves. What changes the outcome is reading the session as a whole.

    2 min read

See DeepID on a real phone.

We show your fraud, security and engineering leads how DeepID would run your sign-up, login or payment journey, then plan a pilot on your own traffic.