Sweat official logo
Sweat logo with rounded background
Token
arrow icon for external link
Sweat logo with squared background
Wallet
arrow icon for external link
Learn iconLearn
arrow icon for external link
Blog
arrow icon for external link
home iconHome
arrow icon for external link

Sweat Economy Blog

Step verification and fraud detection

Updated

SWEAT is a cryptocurrency minted directly by movement. We have therefore taken great measures to ensure the authenticity of movement that goes into minting SWEAT.

Over the last seven years, we have developed a very sophisticated movement validation system with multiple layers of defense. It is impossible to completely prevent any kind of fraudulent movement activity, but the validation system we have developed is very adept at detecting fraudulent behavior, while simultaneously implementing blocks to fraud, such as minting limits to make fraud very expense.

It is worth noting that the Sweat Wallet — your new crypto wallet — is not handling fraud detection itself when it comes to movement data.

We have designed Sweat Economy as such: Sweat Wallet is the hub where minted SWEAT is held and used. The movement data that is used to mint SWEAT directly on-chain comes from another application. This facilitates the onboarding of new Movement Validators to decentralize the network and expand the kinds of movement people could mint SWEAT for. This, however, is a few years away.

For now, the Sweatcoin App is the Movement Validator for steps.

First defense — system design

Users can only mint SWEAT with their first 5,000 steps in a day. Given that the rate of minting difficulty is going to increase exponentially after TGE, those 5,000 steps will mint progressively fewer SWEAT. This is hard coded in our smart contracts.

This model forces aspiring fraudsters to scale horizontally into hundreds or thousands of accounts to justify the operation. Horizontal scaling is expensive, as the Movement Validator, Sweatcoin, only allows for genuine Apple and Google accounts, and each account requires a unique device.

Second defense — movement data and device telemetry

Phone telemetry is essentially the transmission of data points from your phone’s movement to SweatCo Ltd who run the fraud detection service. Telemetry gives strong signals to our team for the need of manual intervention and review of the account. Details can not be provided, for this would only aid attacks!

Third defense — movement validation model

The Sweatcoin Step Verification Algorithm (™) validates steps to filter out the shakes and bumps registered by your phone — which are not real steps — while also recognizing cheating. The algorithm analyzes your movement data and takes out the invalid steps. Once it’s done, all remaining verified steps are converted into sweatcoins or SWEAT.

While some fraudsters — and even regular users — may see an increase in their ‘step count’, it does not necessarily equal an increase in the conversion to sweatcoins, or SWEAT.

This is a constantly-evolving part of the defense. The model evaluates every single step that arrives from every user and, in case of any doubt, flags the accounts for manual investigation. If fraud is confirmed, the case is added as input into the model, and fed into the anti-fraud model (below).

Fourth defense — anti-fraud model

The anti-fraud model takes input from the first, second and third levels of defense, and looks for patterns and connections between accounts that might indicate fraudulent activity. Accounts and networks of accounts are flagged for forensic review and, based on this, the review model is further refined.

Think of the third defense as a single-account analysis. The fourth is a multi-account analysis — the forest in which the trees are misbehaving.

And the fifth…

The fifth and the most important defense is you — our community. It is in all of our interest to make sure that SWEAT is minted by genuine physical activity, and is never ever issued without evidence of sweating.

For the latest news, make sure to join us on Twitter, Discord, Telegram, and Reddit.