Pokemon Go thread: If i miss that zubat one more time

NeloVergil · August 3, 2016, 1:27am

You know that nigga with a Articuno sucked some hella dick.

Cyntalan · August 3, 2016, 2:25am

The website based ones may be down, but this thing still seems to work just fine.

ArtVandelay · August 3, 2016, 3:30am

I like that Pokemon Compass app.
Still find my Pokemans and its crappy enough to not feel like cheating.

Amazing_Funbags · August 3, 2016, 3:58am

I use Pokemaps.

I’m mostly trying to farm coins for more incubators.

jebop · August 3, 2016, 4:11am

Everyone wants a Charizard when Blastoise is the op one.

Pimp_Willy · August 3, 2016, 6:33am

Yeah, people who set up person use ones are still in business, so I think thats just gonna be the future of these things. Makes me glad I didn’t actually spend all that effort into getting my site up and running, and instead can just keep it personal use :tup:

I wonder if we’ll get the new Sun/Moon pokemon before the other pokemon to help cross promote the game

Zatalcon · August 3, 2016, 7:20am

Yeah, but that’s for android only. The type of apps that you can get on Andriod just makes it better than Iphone, IMO. Iphone is ridiculously strict.

Does that link Cyntalan posted work for Iphone?

Cyntalan · August 3, 2016, 7:55am

I haven’t tried it myself, but as I understand it, you set it up to be a ghetto webserver that you direct connect to via the iPhone. There’s even a toggle for following location for that purpose.

jebop · August 3, 2016, 10:49am

One more Candy for Machamp, Venusaur, and Charizard. Then I think I pretty much have everything that isn’t retarded rare. Got the fossils too. I need like 30 for Dragonite and and Lapras to power up since it’s stupid strong. Planned to go nest hunting the weekend gone but the update rekt me. 129 in the Pokédex.

fishjie · August 3, 2016, 11:49am

Pimp_Willy:

fishjie:

Pimp_Willy:

fishjie:

Pimp_Willy:

AlphaCommando:

Amazing_Funbags:

https://www.reddit.com/r/pokemongo/comments/4vcyra/pokémon_go_0310_update_changelog/

New update for PoGo. Not out on Android yet =/

“Removed footprints of nearby Pokémon”

LOL. They can’t fix the fucking thing, so they’re just going to straight up remove it. Looks like I’ll be installing a Pokevision app then.

Its not really about fixing, but it shows how incompetent they are…

All the pokemon locations are calculated server side, but what happened was that for each users heartbeat request, when it passed back a list of all the pokemon within 100m near them, the server took each location and calculated how far the pokemon is in footsteps. This means taking the location of the user, and the location of the pokemon, then running some sort of mathmatical calculation figuring the distance between the two points, and then transforming that into a footstep count. Pretty easy, not particulary heavy in calculations likely. But the problem is, it has to do it for EVERY pokemon for EVERY client for EVERY heartbeat check (which is like 15-30 seconds or so). That adds up FAST.

They disabled the feature (by returning 3 feet always) to make sure their servers were not crashing. Like, they figured the most strain on their whole server was these calculations, and it was the biggest bottleneck, so they removed it, which was a fine emergency band aid fix. But not a long term fix.

The easiest solution, which makes the most sense, is to move that calulcation to the client. The app should get back the location of the pokemon, like it already does, and then it can run those calculations locally, figuring out how many footprints to show. This shifts the load from the server, to being distributed across each player. In fact, some players hacked the game to do just this to fix the 3 step bug, by intercepting the traffic, then changing the value in the packets to the accurate step values.

But instead of doing this, Niantic decided to just remove the step count all together. Its just kind of, wow. Like, I don’t really understand this change. My fear is that they plan to add a new Premium item to the game, that increases the range that you see pokemon on the map, so that instead of needing footprints, you can just use “Lucky Binoculars” and have your pokemon vision radius increased to like 70m so that all pokemon “nearby” are plain visible on the map.

Also, to help thrwart trackers, the range of pokemon returned from you is down to 70m from 100m. Which seems like just a 30% decrease, but when you’re talking about circle area, thats a huge decrease, meaning the poke tracker sights have a harder time working because they have to do more scans. In addition… they were messing with the APIs so they have to be reverse engineered again, which is why many of the tracker sites aren’t even working right now.

I just wish I understood what their ultimate vision is, besides trading. Wild battles, trainer battles, etc. Hopefully they’re too busy creating features to take the time to go back and fix the server optimization problems they have.

At least with the new update, its WAY more stable. Where previously you’d get schrodingers ball syndrome when it froze and you had to force restart, it now recovers from these situations and works seamlessly, which is amazing.

Making the calculation available on the client would mean client would have access to the location of the pokemon, which means ppl can reverse engineer app and then use it to locate pokemons. I’m guessing why the calculations were being done on the server. So I can understand their reasoning

Where I think they show their incompetence is not transacting basic things such as transferring pokemon. Can’t keep track of the # of times I transferred a pokemon only to lose the pokemon and not get a candy. Essentially, the transfer of a pokemon should be rolled back if the user did not get a candy. Or if they don’t wanna transact it cause of server load, they should increment candy count FIRST, then decrement pokemon count. much better user experience, and users occasionally get free candies

They already do, though. The API returns the exact location of the pokemon; thats how the client shows it on the map. But in addition to returning where the pokemon exactly is, it also calculated what step count to show, and the client just displays whatever is shown. Thats how the pokemon tracker sites work, because it simulates the client and gets a list back of all pokemon with gps coordinates to them.

oh the server was already returning the location of the pokemon, and what footprint distance to show? yeah that’s really silly, they could just offload that calculation to the client. i’m not good with low level networking stuff, but how did people intercept the packets being sent back and forth between phone and server (assuming this is how they figured all this out)? genuinely curious, its pretty neat that people figured this out on their own with no actual API from niantech.

was this information being sent in the clear unencrypted, or did they have to figure that out as well?

anyway i knew about pokevision but didnt use it. part of the fun of this game was walking around parks in real life with the wife and having that sense of anticipation of seeing what new pokemon would show up. pokevision pretty much ruined that spirit of discovery.

Well, I was trying to crack it and gave up. The app uses https, so its hard to sniff the traffic and intercept it; it might be possible, but it’d be approaching levels of hacking I wasn’t going to breach lol. So I gave up.

Turns out some other dudes figured it out, and posted their code open source. I think they decompiled the .apk file, and were able to reverse engineer what webservice calls were being used from inside the client; then, you just write your own app, and include all the same information that the app would when you call the webservices. In essence, they were just rewriting the client after decompiling it. I think, I’m not actually sure, I can’t imagine they were able to trial and error their way into a private API, I think the decompiling route makes the most sense. Once they were able to simulate the calls, everything else was built on top of that. The clients even go so far as to have to log in to the pokemon service, which is why they say make sure that you use a fake account not tied to your main one when using these, because you don’t want to get your main account banned if they start looking for anomalies in the data. I know they were able to do this routinely with Ingress, if you did 2 actions that were too far apart to be legal (not just like across the country, but if you tried to hack 2 portals – like using a pokestop – where you were going faster than like 12 mph, then it would prevent it from happening). So they have the tools to recognize spoofers and these pokevision client type things. Now, Im sure they can’t handle it with the amount of raw data coming in, but its only a matter of time anyway before they are able to crack down.

I think in the meantime, until they get the new tracking method up, they should just display all pokemon on the map on your nearby list, as a show of good faith

lol that is really cool. so they decompiled code, reverse engineered it, and then were able to write their own client. If there are alt accounts being used to gather all this data dont see how they could stop this. They’d have to constantly update their API in non backward compatible ways to break existing clients. Pokevision websites can get around the detection by dividing up the country into specific regions, and having multiple instances of the client, each with its own alt account, responsible for patrolling that region by simulating what API calls would look like for a person walking around an area. This would get past any anomaly detection. The pokevision websites can then aggregate all this data and draw a map that wouldn’t be real time, but good enough.

Pimp_Willy · August 3, 2016, 11:55am

I mean thats basically how it was working before. The big hit is that they made two changes:

Scan Radius reduced from 100m to 70m. This is a pretty big percentage drop, requiring more scans to cover the same area.
I think they increased the “heartbeat” timer, so that the time between one client being allowed to scan and another is higher; but yes this could be gotten around with more accounts and smarter distributed checking

I think the main reason the sites have stopped, is because of cease and desist letters, since they are accessing the API against terms of service

jebop · August 3, 2016, 6:39pm

New tier list is out. Vaporeon A Tier lol.

Zatalcon · August 3, 2016, 7:12pm

Ah man. I don’t know shit about that stuff…

I’ve been planning to switch to Android a long time ago so i guess this will be an added reason for me to make the move. Android just gets the best apps.

Pimp_Willy · August 3, 2016, 8:57pm

Actually, I was wrong about the time between scan. It used to be 1 second, then it was upped to 5… and apparently now they’ve upped it to 10 seconds between updates. Which is fairly reasonable for a player on food/bike, to be sure, but sometimes when we’re driving in a car and we see a snorlax pop up, it makes tracking it a bitch lol.

NickRocks · August 4, 2016, 1:46am

My hacked version that lt me walk around wont open up anymore and other ppl w the same hack have confirmed its down

REST IN PEACE DEAD GAME

po_pimpus · August 4, 2016, 1:53am

Now niggas are making tier lists?

***OMFG. ***

jebop · August 4, 2016, 2:19am

Tier lists been out since the very start. It’s accurate, you try killing a 2k+ Dragonite.

It’s basically for attacking/defending gyms, some are better than others and some (Goldbat) are just garbage.

Wasted all my stardust evolving/leveling an Exeggcutor. Then I find a 700 CP Exeggcute. FFS!

Amazing_Funbags · August 4, 2016, 4:13am

Everybody knew Dragonite would be top tier anyway. It’s like Gen 1 all over again except without all the stupid shit.

Cyntalan · August 4, 2016, 5:51am

I don’t recall Dragonite doing all that hot Gen 1, considering how prevalent Ice Beam was. Starmie alone made his life hell.

Infernoman · August 4, 2016, 5:54am

To be fair the devs (like gamefreak) didn’t make it fair from the get go. Therefore tiers were inevitable due to the fact that no all pkmn are capable of being trained equally.