reader comments 54
earlier this week, showing screenshots where two different accounts in the same exact location showed different Pokémon on their “nearby” lists. The site estimates that tens to hundreds of thousands of accounts may have been blinded in this way, based on reports from inside the Pokémon Go hacking community.
reports that many new accounts seem to be blinded quickly and automatically, signaling a change from the more manual ban waves Niantic has issued to bot makers periodically. Some suspect Niantic is making use of machine-learning algorithms to detect bots quickly while limiting false-positive punishments on legitimate accounts (the company was publicly searching for a Machine Learning Engineer last year).
“This is may very well be the beginning of Niantic’s machine learning approach to active bot countering,” user Dronpes writes on The Silph Road subreddit. “If the parameters for a shadowban are constantly adjusted server-side, as they can now easily be, then Niantic’s machine learning engineers can train their detection (classification) algorithms in ever-improving, ever more aggressive ways, and botters will constantly be forced to re-evaluate what factors may be triggering the detection.”
Users on Reddit are busy crowdsourcing a list of the Pokémon that are hidden when using these “blinded” accounts. Some tracker makers, meanwhile, are crowdsourcing user experience data to determine what, precisely, is setting off the blinding.
since shortly after the game’s release. Some bot-makers even use GPS spoofing technology to create apps that can play the game for you, virtually “warping” players to lucrative locations and snapping up Pokémon. Niantic has tried cutting off access to these third-party apps altogether in the past, but bot-makers quickly found a way around that block last August.
In a way, the new “blinding” punishment could be seen as a form of poetic justice for users that want to use third-party tools to see more Pokémon than the base game allows. Unlike a straight ban, the blinding can force bot makers and users to waste time playing a severely limited version of the game or devoting resources just to figuring out if they’ve been blinded in the first place. And if machine learning algorithms are really being brought to bear for more automated bot detection and enforcement, Pokémon Go‘s third-party tool makers may be facing their toughest challenge yet.