Progress in the fight against bots: The New Anti-Cheat Bot Detection

Hey travelers,
Hope your Adventure Week is yielding you troves of buddy candy. :)
This weekend Niantic rolled out a new major anti-cheat measure to help level the playing field, only 16 days after the anti-spawn-sniping measures were added. But as details were hazy, no megathread emerged on the topic with accurate information. We're going to attempt to share what is known, however, and the Silph Road team's analysis of what may be going on to help curb the spread of misinformation on the Road.

Bot Accounts Are Receiving 'Shadowbans'

Midday Saturday EST, reports started emerging that something was amiss in botting communities. Some accounts suddenly stopped finding Pokemon at certain known spawnpoints. It turned out that they were incapable of seeing anything more than very common species.
These bot accounts weren't being "banned" outright, but they were apparently 'flagged' as illicit.
Huge numbers of bot accounts were being flagged, though many were still operating normally. It was initially suspected that accounts were flagged due to using deprecated API calls to access server resources from Niantic's private API, which may have been true, but it turned out not to be the only reason illicit accounts were flagged.
As of this moment, it is still unclear what behavior is tipping off Niantic that an account is not legitimate.
We're calling this flag a 'shadowban.'

So are Scanners down or up?

As bot accounts (which power 'scanners') are being flagged, some scanners are only able to show common species. Others have shut down temporarily, pending a workaround to the anti-botting measures, to preserve their accounts from being shadowbanned.
But many are still operation, as not all bot accounts are being flagged.
Bot operators can simply create new accounts (though this can cost time or even money from black-market vendors) to keep things operational for now. Though notably, even many of these brand new bot accounts are being flagged.

So Why is this so Notable?

Previously, Niantic has banned bot accounts in large banwaves.
This system marks the first rolling, 'automated' ban system we've seen. It signals a clear, continued effort by Niantic to level the playing field against bots and bot-provided advantages.

Are there any false-positives?

Despite potentially hundreds of thousands (or more) of bot accounts being reportedly caught in this net, we've seen very few false positives caught. These have typically been travelers who granted IV Fly access to their account, some of which have ignored the official warning message for some time.
We are not able to say whether IV Fly has been using accounts for botting purposes, but it is always a possibility when granting 3rd-party tools access to your account. Don't give 3rd parties access to your account, travelers!Plenty of awesome IV raters do not require illicitly accessing Niantics servers and can still greatly speed up IV checking.
In the (very unlikely) event that your legitimate account has been shadowbanned (which is easiest to detect when playing with a friend and ensuring you see the same spawns), PM the Silph Executives and we will forward your account information on to Niantic to examine. Be sure to include whether you've authorized illicit 3rd party applications or manipulated your GPS (spoofed) in your message. we recommend giving this page a read.

But ... Spoofing!?

This anti-bot system is tackling a different issue - illicit bots. But that does not mean spoofing is not also being tackled. While GPS manipulation (spoofing) is still a large hurdle the geolocation gaming genre needs to overcome, Niantic has made increasingly explicit statements that they're aware of the problem and plan to tackle it. Pair that with the fact that they've now more than tripled their team size now (100+) and the future looks bright. :)

Long Term Impact

So, will this curb botting forever? Probably not. Once the reverse-engineer groups nail down what it is that's tripping this shadowban, they'll simply be able to adjust their bots to avoid it.
But it is our opinion that we may be witnessing the beginning of a Machine Learning approach to bot detection. 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.
Botting and black markets will always innovate and grow. But Niantic's increasingly aggressive moves to even the playing field for all players send an unmistakable signal of their intention to continue the pursuit until botting becomes prohibitively expensive or laborious.

Parting Words

We've been excited to see Niantic's momentum increase. Paired with the anti-sniping feature rolled out on May 6th,that's two large anti-cheat measures in 16 days - one that impacted spoofers and bots, and one for bots. We've never seen momentum like this. :)
Hopefully this helps prevent the spread of misinformation, travelers.
Finally, we want to thank the good travelers in /new who've been answering newbies' questions reliably. We've seen a marked decrease in snark and rudeness and a big boost in friendly, courteous, help. Thank you for helping keep the Road an awesome place to come learn and enjoy the game with us. :) You do us proud.
Travel safe !


Popular posts from this blog

Bubblestrat 3.0 +++ confirmed

Pokémon Go Raids - Level requirement, how Raid Battles, Raid Bosses, counters, Raid Rewards and Raid Passes work