Clutch says it suspended 211 accounts on September 26 after identifying coordinated activity that it believed was intended to obtain gift cards through multiple accounts. The prediction-game operator reported reversing 114,606 points and 11,950 credits, and said no gift cards had been paid out to the group. [S01]

The figures are different units within Clutch's system. Credits are used on predictions, while points can be exchanged for rewards. The company did not assign a cash value to the reversed balances or disclose a monetary loss.

Patterns cited by the operator

Clutch described repeated, near-simultaneous predictions with the same choice and stake across several accounts, as well as synchronized advertising activity. It also identified 176 referred accounts that rarely went on to play.

The operator says the combination of patterns led it to conclude that one person controlled the network. That is the company's interpretation of its own activity data. It has not named an individual or reported a police investigation, criminal charge or court finding in the announcement.

The group was blocked before redeeming gift cards, according to Clutch. The distinction matters because the event concerns attempted extraction of rewards and the reversal of internal balances, rather than a disclosed completed theft of card value.

The disclosure puts numbers behind the referral pattern: 81% of ordinarily referred accounts placed at least one prediction, compared with 3% of the suspected farm’s invitees. Clutch does not disclose the ordinary referral sample size or observation window, so the comparison describes its incident analysis rather than an independently measured industry fraud rate.

The data behind the story

Clutch referral accounts that went on to play

The operator’s comparison of referred accounts placing at least one prediction in its phone-farm investigation.

%

Source
Clutch incident disclosure, September 26, 2026

Operator-reported observations, without an independent audit, disclosed ordinary sample size or measurement window. These are activity rates among referred accounts, not percentages of users committing fraud.

Show data tableHide data table
Scroll left or right to see all columns
Clutch referral accounts that went on to play (%)
CategoryPlaced at least one prediction
Ordinary referrals81%
Suspected farm referrals3%

Protecting a finite reward budget

Clutch, developed by France-based Rodger Studio, runs a free prediction game in which successful activity can lead to gift-card and prize redemption. Daily credits, rewarded advertising and referral bonuses create several ways for an account to accumulate value within the game.

The company says multiple-account farming diverts its reward budget and distorts the leaderboard. Its stated rule is one account per person; it also says ordinary shared Wi-Fi use or frequent play alone is not prohibited.

The company also reported an average of 13.6 rewarded videos per active day for a farm account, compared with 8.5 for a typical player. That is an activity comparison inside the reward system; it does not establish advertising revenue, a cash loss or the monetary value of the points removed.

The data behind the story

Rewarded-video activity in Clutch’s investigation

Reported daily rewarded-video viewing by the operator’s typical-player and suspected-farm account groups.

videos per active day

Source
Clutch incident disclosure, September 26, 2026

Averages reported by Clutch. Group sample sizes and observation dates are not disclosed. More videos do not establish more gift-card redemptions; the operator reported none paid to the suspended group.

Show data tableHide data table
Scroll left or right to see all columns
Rewarded-video activity in Clutch’s investigation (videos per active day)
CategoryVideos per active day
Typical player8.5 videos per active day
Suspected farm account13.6 videos per active day

No independent audit of the account findings was provided. The confirmed public development is the operator's dated suspension and balance reversal, with the reported detection outcome limited to its own description of the incident.