Rocket Map admits defeat: account blinding is super effective, it will take months to defeat it

0
27
@everyone Hi guys, it’s time to talk about the Niantic in the room. This is about the blind accounts.

Ever since the blinding waves, we haven’t stopped researching every single aspect of RM, pgoapi and PoGo. We did this privately to focus on our work rather than trying to deal with endless speculation and opinions.

So to answer the important question “Is there a final solution in the short term?”: No.

If you want to make sense of what has been happening, imagine that Niantic has installed an AI to analyze and flag accounts. Whether this is actually the case is irrelevant for the example, the end result is the same: it’s not a single or straightforward thing that gets accounts flagged, and the effect on accounts is based on multiple aspects of the accounts.

So what now?

Internally, we’ve set up a list of items that we need to address that all have a possible (or clear) effect on the flagging rate. We’ll be continuing our work with the goal to make RM’s scanning behavior the least identifiable in the long term. Finishing all of these items will take several months, but in the meanwhile we’ll continue to work on (and release) updates that will reduce the effect and the importance of account flags/bans.

We can confirm that:
1. New accounts work for 60 to 140 hours (depending on config).
2. Blinded accounts can get unblinded after 6 to 10 days.
3. Reusing unblinded accounts in any 3rd party app (including RM) will get them blinded again faster than before (sometimes in just a few hours).
4. If you buy accounts to scan, do so at your own risk. These have often been getting blinded in a matter of a few hours (~4h), most likely because they had already been flagged before.
5. Once flagged, the blinding is inevitable.
6. All 3rd party apps/scanners are affected in the exact same manner. We’ve spent extra time to confirm this because some people were pretty convinced we were wrong, although it usually ended up being because they hadn’t even realized their accounts were already blind.
7. There is NO recommended way to test accounts for blindness. The current 3rd party implementations are incorrect and could get your accounts flagged if they weren’t already. This is a WIP and is being added to RM itself.

Right now, here are 3 approaches for your configs to maximize the scan time per account. Results depend on a lot of things, so test and experiment for yourself until you find what works best for you:
1. Burn through accounts: no sleep, no account rotation. For some whose accounts usually get flagged very early on, this will increase the scan time of the account.
2. Use a basic constant rotation: e.g. asi 8h (8h of scanning) for ari 4h (4 hours of sleep).
3. Use more realistic scan times: low asi and high ari (scan in bursts: short period of scanning for a realistic resting time), or low asi and low ari (what you would call “not too active players”) but use enough spare accounts to fill 24 hours with realistic schedules for all accounts.

The ideal will depend on your own results, we’ve found that all three approaches had positive effects for at least one testing setup. A person who needs #1 will be the direct opposite of someone running #3, but both are equally valid.

The blinding issue will only be solved in the long term by addressing all points, and that’s what we’re doing now thumbsup It’s a solvable problem, but it takes time.

 

Leave a Reply