I joined the GeoGuessr Police

GeoGuessr creator code: matepotato
Socials:
Discord - / discord
Twitch - / matepotato
Twitter - / matepotato
Instagram - / mate_potato
#geoguessr

[@alexanderpedersen1401]

It should be displayed if the person we are reviewing is making the first guess, or the point where the opponent guesses.

[@matepotato]

oh yeah absolutely good point, having the timer once the opponent guessed on screen would also help

[@Icy.3P]

Thank you for keeping the community safe officer potato 👮

[@trulybuzzy]

13:54 also some people queue up for all modes and forget, i've had times where i thought i was in a nmpz duel but it was nm

[@NathanLipetz]

I sometimes play moving as if it’s NM because of long NM queue times (I’m in gold 1). My moving is also pretty slow so sometimes just playing it as no move or barely move and putting in a fast guess wins games

[@matepotato]

Oh yeah good point, i have also certainly played nmpz on nm games for the first few rounds before haha

[@FrankGevaerts]

That Oslo round looked to me like they wanted to go find more information to get the best possible guess, but then their opponent instaguessed so they had to give up on that and go generic Oslo.

[@matepotato]

yeah probably something like that.. or maybe they wanted to look around a bit more to make sure that sign wasn't bait

[@Leaferr]

i think it should specify rank aswell as the country they're from, then you basically have all the tools to be able to understand how a player is going to think.

[@matepotato]

yeah that could work.. although people can put up fake countries in their profiles too of course

[@Leaferr]

@@matepotato this is true, how much do you think it would matter? i could see a point in say switching your country flag to say russias and then exclusively banking on cheating on russia rounds due to it's large size to seem legit

[@TehOnlyAnd1-pw8ci]

One thing you didn't notice in the Negros Occidental round is that this was apparently a team duel. The player you are reviewing plonks on the town at 17:55 in your video and the team mate plonks four seconds later (different avatar). In my view we can't really judge team duels from just one vantage point (although in this case I still think it was googling). There was another round in Frankfurt yesterday that looked somewhat suspicious until you noticed that the teammate plonked in Frankfurt before the player that was being reviewed. So at the very least it should be indicated that it is a team duel but as you don't see the chat (and don't hear a potential VC obviously) I don't think team duels should be included in the review unless we can review both players.

[@matepotato]

oh wow nice spot! yeah that is insane, reviewing 1 round from the pov of 1 player in a moving team duel makes absolutely no sense, even more so if they dont even specify that it is a team duel anywhere.. totally agree with what you said

[@trulybuzzy]

bro has 1million unanswered messages

[@matepotato]

bro i swear i cleared the backlog like 1 week ago but it built up again

[@michaelwisniewski6047]

As you got basically the same rounds as GeoPeter, it’s probably pilot / testing phase

[@astrosoup]

Even in a lower division, center clicking Ecudaor is not a rare occasion. Showing player rank would not be good because it would lead evaluators to assume that the player cannot have certain knowledge that they might indeed possess, even if they don’t know enough overall to be high ranked.

[@matepotato]

yeah that is a good point, you cant really ever know what someone might or might not know but even still you can sort of gauge whether it is suspicious or not based on how they play. for instance, someone zooming in on a sign for 30 seconds and then proceeding to zoom onto the place is usually more suspicious than someone who instantly zooms onto it after seeing a sign (for if you know the place, you should not need an additionally 30 seconds to start zooming onto it usually).. there are always ifs and buts of course but reviewers having more information (such as rating) will certainly help make a better judgement than with less information

[@oZiinCx]

Yeah I kinda agree with that, not long ago, I played against a bronze guy who was in a 30+ winning streak (I'm in master division) and he absolutly destroyed me. At first I wannted to report him, but I looked at his profile, and he just have an enormous number of solo games so he's just way to good for his rank so that was just bad luck for me

[@arrowheaded]

14:32 It was reported because they 5ked without zooming in. You can zoom in on the map to see exactly where they guessed after, it's really helpful. Another huge telltale sign that i found funny is that the cheating players never line up the road angle with the conpass, but still manage to 5k a rural round...

[@matepotato]

Ah interesting, I missed that.. always still possible that it was just a lucky plonk though i guess.. yeah that's a good point seeing who uses the compass could certainly help

[@cropsini]

13:30 Kigali shows on the map only if you zoom really hard and it can feel wierd

[@matepotato]

yeah that's true.. also dont really think that map zoom is too suspicious for nmpz gameplay

[@fursuits]

I did my cases earlier today and we have the exact same ones lol. I'm guessing they're trying to see how accurate everyone is in their verdicts.

[@matepotato]

yeah this is also what i was thinking.. it's probably just a bunch of rounds that geoguessr already has their own anticheat for (so they already know the answers) and are maybe just trying to figure out how good the community is overall at detecting cheaters/which players are better at detecting cheaters

[@eriklindberg-wysocki6642]

i currently dont think its an active system your vid, peters vid and my testing it have a lot of repeats, i even got some repeats within my own testing. if it was actually pulling from an actual list of reports i wouldnt expect that

[@trulybuzzy]

if it's a repeat you don't stay still on the loc for 15+ seconds. the googlers mate voted as guilty are 100% guilty

[@eriklindberg-wysocki6642]

​@@trulybuzzy i mean a repeat investigation round, i got the same round to review as peter and mate

[@trulybuzzy]

​@@eriklindberg-wysocki6642 oh my bad, yep we're all getting the same reports so that there are a lot of votes, which is good to ensure fairness

[@matepotato]

yeah certainly, feels like some overall testing phase before they make the actual thing

[@БранимирМилошевић]

I think you can only vote cheating if it's conclusively obvious. Like, the SA 5k after zooming WA and SA, they never stopped moving to google and they might've just been very ballsy and lucky going in. It happens. I've definitely made even closer and luckier 5k's with somewhat random zoom commits. I just feel like time spent and movement cues are key.

[@matepotato]

yeah certainly.. and there will still always be some ambiguity because while reviewing we hyper analyze every action of theirs but the players might not necessarily be taking it as seriously (eg maybe they're eating while playing so they happen to pause or not move while taking a bite).. you never really know the full context

[@mariuswinkler2790]

As a gold II player sometimes I stare for long time at simple signs like that Rio sign e.g. to figure out if it's spanish or maybe Portuguese. Hope that I won't get found guilty soon for my slow and random gameplay😂

[@matepotato]

fair enough haha.. i assume if they mention the division/rating of the player it would certainly help gauge these things better

[@spunj]

The biggest and first thing I noticed was what you said, just include the division and it becomes a lot easier to give my opinion lol.

[@ipranayy]

the 8th unread message is me lol.

[@matepotato]

oops i'll get to it soon haha

[@OsefKincaid]

It takes a decent amount of time to google something, like the first guy did. The second guy didn't really have a pause like this in his gameplay, I don't think it's very clear that he googled from this.

[@cheeseburgercheese2]

i’m from south australia, and that sa round really looked like sa to me, there is a chance they were from there but idk

[@matepotato]

yeah i suppose it's always possible, the zoom into wa was still odd then in that case

[@DadgeCity]

You have to bear in mind that some people think everyone is cheating. Gg should delete reports from bad players rather than waste reviewers' time.

[@matepotato]

haha that's true

[@Ice_B3ar]

I didnt even knew this feature existed

[@thegees]

first video i have watched from you it was pretty good (that one round when you submitted your guilty verdict you left the scripting on 'insufficient evidence'. i think by default none should be selected so that doesn't happen)

[@matepotato]

that's a good point, was certainly unintentional from me

[@arrowheaded]

I did like 10+ of these earlier to try it out and it's interesting to see you review mostly the same cases that I had. We had pretty much the same verdicts, so I wonder how unanimous community consensus will be 🤔

[@matepotato]

Most of these cases did feel rather conclusive.. there probably will be some that are more borderline/hard to figure out where opinions might split i suppose

[@faaafv]

the first one is when i wished they also gave the elo because it was obviously ecuador but maybe not if it was a low rank i give them the benefit of the doubt so when they look at a sign for some time i try to see what you can google with the information and time they have so i did that for ecuador and didn't get anything in ecuador in the first few results, maybe they are good enough to know it was ecuador but wanted to get a better score by googling and didn't find anything conclusive or maybe they looked at the sign didn't google and went with a random ecuador guess some of them are hard to be sure, elo and flag would help in some cases

[@matepotato]

ah interesting! yeah certainly was a weird round to have to judge haha

[@dominikheinz2297]

It would be great if we see the mouse cursor movements the player is doing. E.g. they might not be scanning the map or panning around, but just hovering with the cursor over letters on a sign. For example a to them forgeign language - let's say cyrillic, and they take a long time to translate each letter. With the current system, starring at a sign for a long time, tabbing out and googling the location would appear no different than starring at the sign for a long time, while slowly hovering with the cursor over the letters to translate them one by one.

[@matepotato]

certainly that's true.. and especially so for nmpz because we really will have no idea what theyre paying attention to in those games.. although im not sure how feasible that is because i assume seeing the cursor would require geoguessr to record your screen (which probably has some privacy issues) while the current system afaik doesnt record your screen but instead just pulls your movements, panning and zooming in that game from their internal data

[@WannesDeBacker]

I find you where too hard on the guy in Australia that clicked in SA, I would totally have clicked exactly the same with the same clues and I'm pretty decent at Australia.

[@matepotato]

yeah i might have been, that's fair.. although there are also other comments in here saying it was the most obvious script of them all hahaha interesting to see all the perspectives.. i dont think as much that their final guess was too suspicious but the way the reached that guess was more suspicious (aka spending much time on that sign + zooming into wa first for some reason)

[@WannesDeBacker]

@@matepotato yes ok I do get it (although he didn't really look at the sign that long, looked like the time to read an entire sign and just went straight onto moving). But the scripting thing could be real. Can this be a team duel? I imagine playing with a team member that calls out australia but the guy doesn't know where, so he just looks at WA, and then his teammate just says, naah just a little away from adelaide. (i might be overthinking this, he could have just been scripting).

[@Canadagraphs]

LOL, your first example was one I got too. Its the only inconclusive I've handed out yet. Like, its suspicious how long he was on that sign, but didnt click in the immediate area of where it is. I didnt know what to make of that. FOLLOW UP - I didnt let the video play far enough.... I am in Master I and got the same replay. So I dont think they are handing them out based on division (or, if they are, they are also just handing out assignments for lower division games to everyone as well).

[@matepotato]

interesting.. curious to see what they plan to do in the future with this system

[@Greenslime300]

Looking at only isolated incidents rather than whole rounds doesn't really tell you much about how the player normally plays. I wonder how many false positives are going to come up just because players simply got a lucky guess

[@matepotato]

yeah certainly.. the more info to the reviewer the better but you can never really be 100% sure

[@leoa.1571]

I think that cambo plonk was way too good and the player didnt even zoom in. Clear scripting to me

[@matepotato]

certainly possible but we all can have some lucky plonks sometimes.. a bit ambiguous for me personally

[@jimlawton4184]

It just so happened the other day I got a location right in my hometown which is only about 400,000 in Ontario. I honestly didn’t know the neighborhood but it looked so eerily familiar I just guessed and got it right. I’m definitely gonna be included in these 😅 Also, a couple others I could see a city name in VERY tiny writing, Lagos and Kansas City, it’s likely other people might not be able to read it and assume I’m cheating

[@AnthonyRochester]

Do they give guidelines on how to determine if it's cheating or not, or a standard about what level of confidence you need to have for the decision etc?

[@AnthonyRochester]

I guess the second part of my question is covered by the verdict input thing

[@michaelwisniewski6047]

They are sending this to very highly rated players. They know better than the dev team how to determine this.

[@AnthonyRochester]

@@michaelwisniewski6047 Good point

[@matepotato]

yeah i think for now they are probably just leaving it up to our judgement to maybe test how well our opinions agree with whatever internal anticheat systems they might have.. this is probably still in the testing phase

[@CalebTheRat]

I think if geoguessr gives this to the right people it will be a great feature. I also just dropped my first ever good score in daily challenge! P.S (it should be given to known good players like you.)

[@matepotato]

congrats on your good score!

[@cribu_]

nice work!

[@JHK314]

Re-stream Max Velocity

[@matepotato]

u reckon

[@SidKneeGeo]

I think if its a consistent level of crazy guess then yes, they probably are cheating. However i think every one who has played this game for while would have made an insane guess that would def look like cheating. Its why i love this game to sometimes get that insane plonk. I guessed a field in Thailand no moving yesterday and was 1.5 km away. There was no meta like poles cars etc. Its hard to guess if someone is cheating off one round.

[@dominikheinz2297]

That doesnt work. One could play legit the entire round, lose more and more hp, and then for example google with the info they have in the last round to insta win. A lucky guess and cheating often appear very different.

[@matepotato]

yeah hopefully reviewers keep in mind the possibility of lucky plonks happening because they certainly do happen and just vote unsure.. i think guilty should only be voted when you are like at least 95% sure it looks like cheating

[@SidKneeGeo]

@@matepotato Sounds good!

[@SidKneeGeo]

@@dominikheinz2297 yeah thats quite possible for sure if they do get googleable info in the latter rounds.

[@RACECAGD]

Its actually crazy the amount of cheaters ive ran into on a daily basis for example my first game of geoguessr duals i ran into a cheater

[@NathanLipetz]

I suspect the games are of any level. I wish they showed the player’s elo and division in the reports you are reviewing though

[@matepotato]

definitely

[@treezerik]

You are so famous and cool

[@trulybuzzy]

lfgggg