Talk:BF2142 Statistics
Contents |
Note about EA and the AUTH key for stats-queries.
i asked John Hargelid about the AUTH section, here is what he said.
S@tCoM zegt: Question : the 2142 stats queries are encrypted, is this encryption info that will be available to the public? ( http://bf2142tracker.com/ seemed to have the solution. ) John Hargelid zegt: We yet have to find a solid way of allowing people to request data from GameSpy S@tCoM zegt: so does bftracker has a deal with EA ? John Hargelid zegt: no S@tCoM zegt: ok John Hargelid zegt: (at least not what I am aware of =)
So thats EA's definitive stand on the subject. --Satcom 07:02, 25 October 2006 (MDT)
PID is included within auth
while playing around i found that bf2142 sent a request using only AUTH and it returned my current players details
some basics about the auth
it does seem to be an base64 encode but it doesn't look like it. minimum char for the auth is 3 and max is 27
ok.
Based on Base64 the base64decoded value is 14 chars.
the PID is 8 chars.
12345678901234 - fictional decoded auth 44652121 - Gamespy PID ........?????? - Possible timestamp.
any suggestions?
Version handling
All the discussions about the auth key etc. If EA don't want us to access the game stat servers and done so much to avoid us being able to. They would have put in code that can/will be changed when new patches are out. I believe there would be some information in this string that tell EA/GameSpy what version the client runs...? Just a thought...
- they will change the auth token, that much I can confidently assure you. Supposedly GameSpy/IGN will provide public feeds so that we don't have to keep cracking the new auth token each time the game is updated. As I see it, they are limited as to what they can do though, either they provide a different url for each version of auth, or a new parameter name, or they change the underlying bits that make up the auth token (otherwise old clients' stats will not work). Whatever it is (sorry Bill) machine code has to be shipped w/ the game and we can always find out whats going on to produce that (unless they move to a single sign on model <wink>) --MadHatter 13:24, 3 November 2006 (MST)
RE: Version
That very well may be the case. There are 2 sections of the auth key that are static across all queries on accounts(for the most part). These could be the version identifier. But thats a pretty stupid way to attempt to block it, since we can just grab the auth key when the version changes and get the new number for the version.--Craigins 21:35, 2 November 2006 (MST)
History (Problems section from BF2142_Statistics article page)
Im having trouble trying to understand the 24-character auth parameter all of these functions seem to require. I know it represents your user account or player identity in some way but what confuses me is that each time Battlefield 2142 requests one of these functions a brand new auth string is generated. How is the auth string generated??? --Danielphin 01:11, 25 August 2006 (MDT)
same here, don't know how to solve the "problem" with the encypted auth-parameter :( --arnonuehm 15:55, 10 September 2006 (CET)
Did Battlefield 2 ever have any auth parameter before or just after it was released? It could be possible this value is only required for the betas to keep the data private. To an extent. --Danielphin 07:00, 11 September 2006 (MDT)
I noticed that the old query method that worked during the beta now returns an error message "E 104 $ 4 $" --Danielphin 10:11, 21 September 2006 (MDT)
I don't have any valid pid to test it out. It's only returning: "None: No error" --DestinedTConquer 17:41, 30 September 2006 (GMT+1)
It appears that they have changed the protocol yet again ready for the demo. Here is a link that worked during the second phase of the beta and now gives an expired token message. --Danielphin 09:16, 2 October 2006 (MDT)
A reason could be that they've simply "retired" http://stella.xxx as this was used for beta testing and nothing more. Now we're close to demo and live it could be worth trying the host "bf2142web.gamespy.com" (which does resolve correctly) to see if that returns the expected response, eg "http://bf2142web.gamespy.com/getawardsinfo.aspx" like in BF2. --Kurt 10:08, 3 October 2006 (MDT)
Hmm, that domain wasnt valid during the beta. I thought like yourself to try that domain but with no luck.
I just wanna heads up that this link now works correctly --Danielphin 12:39, 3 October 2006 (MDT)
Installed the released version of BF2142 yesterday and turned on a packetspy utility before starting the game.
FYI the auth= parameter still exists on every stats request against GameSpy (server is stella.prod.gamespy.com).
I think this means that it will be hard to fetch statistics from BF2142 using the old BF2-way.
I have some thoughts on how to fix this, and it includes the use of WinPCap on every clan members computer.
I found an .NET API against WinPCap, and created a command-line application that I start before fireing up BF2142. This .NET application uses WinPCap and starts a trace with a filter looking for requests that BF2142 sends to http://stella.prod.gamespy.com. Every request (not that many) for the current user will then be sendt to a webservice on our clan website that handles the stat-fetch part. The requests against gamespy are at least vaild for 20 minutes or so (I have not testet the exact expiry time).
I have run a test on this today, and it actually works. Not the easiest way around though. Would be great to know how/what to encode in the auth-parameter. I think it is the PlayerID (PID) that is the base for encryption, but the auth-key changes after som time for the same request, so it has to be something more.
I Really hope that someone finds a easier way to gather the BF2142 stats.
Good luck :o) --Rusty(NO) 07:58, 19 October 2006 (MDT)
did anyone noted the 'token' ?
/getplayerinfo.aspx?auth=NPh1eX91SAeUsDoUam7xvg__&mode=base&pToken=OxSWQ6oylBUPRhync6yubBS7ORzVvYwLMKkDUfSXqvzzHXL5oDShxEZu0EaHXH2lwK32vy3D6a1J52raU8Glvg__
I also noted that the last 2 UNDERSCORE signs, are NOT part of the auth. the appear the same with 10+ accounts.
Update : the sheme looks very like a RSA SecurID Token. this sheme has time hashes, that are similar to what we expirience.... ofcourse, this is all speculation. --Satcom 14:34, 20 October 2006 (MDT)
the token is an encrypted stat query. I'm sure the auth value is equates to some datetime stamp since using one generated by the game works for about 15 minutes after I capture it in ethereal saying the time has expired. Looking deeper into the game, it seems they're using PACE ESD for authorization, and I can only assume that they're using this on the stats servers as well, so it may be that we can never access the servers w/ out a complete authentication transaction... --MH 21 October 2006 (CST)
GetBackendInfo seems to return the awards / unlocks criteria used for the games sats. --MH 21 October 2006 (CST)
The website 2142-stats.com seems to have worked out how to query bf2142 statistics. Would anyone care to contact the webmaster and ask for a solution to our problems? --Danielphin 09:13, 22 October 2006 (MDT)
I am the owner and coder of 2142-stats.com but at the moment I am not allowed to say anything about how I retrive the stats. You have to find it out! --1ApRiL 10:55, 22 October 2006 (MDT)
- Does that mean it's doable or do you mean that EA have told you how and asked you not to say? --Streaky 19:47, 22 October 2006 (MDT)
with users like 1ApRiL,Dont we like the BFcommunity? anyways: Team ABR has issued a email regarding this mechanism to EA. i'll keep u informed.--Satcom 12:49, 23 October 2006 (MDT)
BF2142 stats are now up on http://bf2142tracker.com. Anyone know if they have access to a public stats feed or if they solved the auth part? --Rusty(NO) 02:07, 24 October 2006 (MDT)
bf2142tracker is owned by a guy who does tracker sites like this for AA and all the bf titles. he's worked closely w/ ea on the stats feeds and gets it from a special stats feed that they provide for him. --MH 14:52 24 October 2006 (CST)
I have spoken with a guy, who has done alot of work with stats for bf2 and he says that decrytping the auth key is the (no pun intended) key to the whole thing. He said it was a fairly simple encryption that anyone with a basic knowledge of key decryption could break it. Thats all i could get out of him though.--User:esparkman 04:41, 24 October 2006 (EST)
well, lets get started then shall we?
the auth and token parameters are base 64 encoded. 3 characters are replaced they are:
| Base 64 char | EA's URL char replacement |
| '=' | '_' |
| '/' | '[' |
| '+' | ']' |
so that "4Eq5FDcUJyWw6NR3urE/z4hN75o04bCaHe8Cin9FiHmt1lpAbQlj+1GTUZiBdGPgL8QsqzsNRKex37JJa/dVUQ==" (a query token from get player info) gets transformed into "4Eq5FDcUJyWw6NR3urE[z4hN75o04bCaHe8Cin9FiHmt1lpAbQlj]1GTUZiBdGPgL8QsqzsNRKex37JJa[dVUQ__"
auth is always 16 bytes long (there are a few encryption algorithms that use a 16 byte key, though I dont think this is a key). I tested this by generating a random base 64 string from a 16 byte array, and sent it into http://stella.prod.gamespy.com/getbackendinfo.aspx?auth=TWFkSGF0dGVyMjE0Mg__ and the message recieved is: "DecryptionFailure: Authentication token decryption failure" which may be a generic response, but leads me to believe that this 16 byte value is an encrypted value (I'm guessing a time stamp because it changes in a manner that seems consistent w/ this).
there are some certs in the bf2142 folder in a file named confbar.mdf which may help shed some light as to what the auth parameter is comprised of. --MH 21:43, 24 October 2006 (CST)
Ok i have been trying to work out an base64 encode using a unix time but it doesn't seem to work. I used "stellaXXXXXXXX" with x being replaced with the integer of the unix time of a forwarded 1 hour in advance and it gives me a 24 char base64 hash and i just replace the == with __ but it still doesn't work. I have tried to unencode the auth that bf2142 sends but there is no success, I believe that there is a alpha numeric replacement and then you can decode the base64. I believe there is a need to decompile some .dll to find a function that does the trick. Nitr021 04:14, 25 October 2006 (MDT)
- Nitr021, i'm wondering what triggered you to try stellaUNIXTMSTMP?
If the auth string was simply base64encode(SOMETHING + UNIX_TIMESTAMP+3600) then SOMETHING would have to be some piece of information that uniquely identifiers the player, as each auth string is somehow unique to that player. --Danielphin 10:32, 25 October 2006 (MDT)
that something can only be the gamespy playerid --Satcom 10:34, 25 October 2006 (MDT)
hmm, inserting a gamespy playerid in various places makes the base64 hash more than the typical 24 characters. --Danielphin 10:51, 25 October 2006 (MDT)
One obvservation I made was that each request to the stats servers had a new auth value (always 20 characters) except if when I requested the leaderboard stats and checked off filters (like dog tag filter and so on), then for those requests it used the exact same auth value. the leaderboard requests were made in a vary short time of oneanother, but the others were spaced out a little. this is why I think that that value (if not directly) has something to do w/ a time stamp. perhaps not directly. I ran seferal queries (Except for getplayerinfo) with auth values from the other queries (generated by the game) and it still returned valid results. One thing we should try is to replace an auth value not sent out w/ the getplayerinfo query to see if the Token parameter is at all related to the auth parameter.--MH 13:41, 25 October 2006 (CST)
Danielphin - are you certain that the SOMETHING has to be unique to the player. Have you tested multiple accounts from the same IP or even Machine(MAC address)? --Craigins 12:55, 25 October 2006 (MDT)
- I'm sure its unique, ive put a request URL into the google translator :P and it gives the same result. I also did a wget on my remote server. --Danielphin 02:09, 26 October 2006 (MDT)
did anyone try out different date/time formats? we should make a list of what´s tried, and with what thought behind it --Satcom 13:16, 25 October 2006 (MDT)
You guys said that the unix time stamp and the PID don't fit in the 16 byte auth key. However:
// Unix timestamp: 1161812842
// PID: 44652121
string s = "stella" + ((char)2);
s += (char)169;
s += (char)86;
s += (char)89;
s += (char)69;
s += (char)63;
s += (char)219;
s += (char)106;
If you break the 2 numbers down into their separate bytes, then cast them to chars and encode the resulting string, you get:
c3RlbGxhAsKpVllFP8Obag== which is 24 characters. Don't assume that it is in the straight number->text format. Who knows.--Craigins 15:01, 25 October 2006 (MDT)
except that the auth value is pretty consistently 20 (base 64 encoded) characters (actually 16 bytes padded to 20 base 64 characters).--MadHatter 15:22, 25 October 2006 (MDT) EDIT: sorry, it is 24 characters... my bad
MadHatter- take out the word stella from the above string and you get AsKpVllFP8Obag== which is 16 characters. Pad it to 20 somehow and see if it works. The reason i went to 24 was because of the above post which had an example authkey of auth=NPh1eX91SAeUsDoUam7xvg__ which is 24 characters long.
If you unzip the pylib-2.3.4 file in your bf2142 dir. You got a lot of neat files. Especially in the test directory. Like testbase64.py All of them are in python but can be opened with notepad. All of them pretty much explain the code. Also the testmd5.py & stat.py & statcache.py and the TOKEN.PY is a cool read !! Hope this helps.. --LDD Bozo 16:18, 25 October 2006 (MDT)
how exactly does this help? I know nothing about python, but test_base64.py is a unit test to test the base64 encoding. test_md5.py is a unit test for md5, stat.py are file attribute tests, and statcache.py seems to be some sort mechanism for loading the stat py code. token.py seems to be the way to tokenize things for python serialization. these dont provide much more insight to me atleast. --MadHatter 17:18, 25 October 2006 (MDT)
Havent found anything in pylib-2.3.4 that I can use, but it might still exist in this library. The unittests vil not give many answers.
A standard base64decode (UTF-8) of auth "NPh1eX91SAeUsDoUam7xvg__" gives us: "4uyuH\a:�jn" (after replacing __ with ==) so there exists som form of encryption that has to be solved after the base64 decode has run. The ByteArray looks like this before encoding it to the string "4uyuH\a:�jn":
{Dimensions:[16]}
[0]: 52 [1]: 248 [2]: 117 [3]: 121 [4]: 127 [5]: 117 [6]: 72 [7]: 7 [8]: 148 [9]: 176 [10]: 58 [11]: 20 [12]: 106 [13]: 110 [14]: 241 [15]: 190
Btw, EA has released their own BF2142 stats site now: http://stage.battlefield.ea.com/battlefield/bf2142 --Rusty(NO) 06:30, 26 October 2006 (MDT)
I agree with rusty about the python code. If someone wants to dig through each file, find out where the Host object is defined at. It is used to communicate with the stats server I believe(after reading through all the python stat code).
I also agree that there is a lot of encryption going on after the decoding. I monitored my packets last night and came up with the following list of auths, decuded using standard base64decode (UTF-8) and came up with the following:
AuthCode Ascii of the characters after decoding. iQEBWfVFxglMh81vF9i75g__ 1 1 89 69 9 76 111 23 59 0VhKfQfV1gXjOtrPLj43eA__ 88 74 125 7 5 58 46 62 55 120 2p5aIMGI[1Dya8H57Zh8Dw__ 158 90 32 80 107 124 15 38dpI6LJC1vOdCpi814teQ__ 105 35 11 91 116 42 98 94 45 121 55uq4a[IOxRVMnrV88JLDg__ 234 59 20 85 50 122 75 14 5aDb47bhk48FLA[cv]OBjg__ 207 5 44 15 63 78 5fDxSUnhqSdJLfmR7m5dGg__ 73 73 39 73 45 110 93 26 62PjcWg8cEDJmYiyI2CG1w__ 99 113 104 60 112 64 89 35 96 7pyFAHHDtbPNYZQCDaRwmw__ 5 0 113 245 97 2 13 112 FT1WvqcI8AHItWr5NzkyUw__ 21 61 86 8 1 53 106 55 57 50 83 g6q8nlLzs2vHTE3MzNYd9g__ 82 107 76 77 29 H6S[iGJwIY]42IbnIz33Ag__ 31 98 112 33 6 35 61 2 idss]0lsSaKBuWQydP]hgw__ 44 73 108 73 100 50 116 MJB9[9GYU6ms2NeFbkquiw__ 48 125 88 83 197 110 74 mRHQTvZQIMhi4U4Op24SoQ__ 17 78 80 32 98 78 14 110 18 pMgSo2t5CxjuJjhBCrv77g__ 18 107 121 11 24 38 56 65 10 q00qnIEsryEtLskzb5HwIg__ 77 42 44 33 45 46 51 111 34 q00qnIEsryEtLskzb5HwIg__ 77 42 44 33 45 46 51 111 34 ReOBI8sQaAFIf3pCIRKsvQ__ 69 35 16 104 1 72 127 122 66 33 18 TIvOgj[eLqCJ7JECW52yLg__ 76 130 63 46 2 91 46 VCfK25Yxd4zW8ZD593ek2Q__ 84 39 214 49 119 119 Xejcq[TgTT]FhU]4cLD0SQ__ 93 43 77 63 79 112 73 igg06CRnbSAoVSmQsIuMnA__ 8 52 36 103 109 32 40 85 41 0YJVfMVj74oCiNAqCPgsjA__ 66 85 124 99 2 42 8 44 5Expic1e1POyODluFCxfPA__ 76 105 94 56 57 110 20 44 95 60 7d[AnhHiku5dAKX9Xsjw6A__ 17 93 0 94 s0JSWPmFjo[Tu9d[GnKqIA__ 66 82 88 251 127 26 114 32 GXszUK0tTC7[kbgCtoBjMQ__ 25 123 51 80 45 76 46 2 99 49 s0JSWPmFjo[Tu9d[GnKqIA__ 66 82 88 251 127 26 114 32 3KLvL5bYb6AvLA5NuBXyAw__ 34 47 111 47 44 14 77 21 3 0YJVfMVj74oCiNAqCPgsjA__ 66 85 124 99 2 42 8 44 DHA1JibVoqqzpOHQjBvcSQ__ 12 112 53 38 38 98 12 27 73 igg06CRnbSAoVSmQsIuMnA__ 8 52 36 103 109 32 40 85 41 0]Wsje7QbJSTkCtK66Etgw__ 13 108 43 74 45 iQEBWfVFxglMh81vF9i75g__ 1 1 89 69 9 76 111 23 59
If anyone wants to give decrypting a shot, go for it. --Craigins 07:44, 26 October 2006 (MDT) edit: yes i ran the conversions listed in the table above
- Did you do the underscore and square bracket conversion? maybe this is just getting to my head --Danielphin 09:02, 26 October 2006 (MDT)
Quick note: the HOST object is NOT in python. python imports it, but the actual code is c not python. dont forget that the stats is EXACTLY the same as bf2, with little addon authcode. i have ´ranked server files´ and also there is nothing on the auth, after i checked them. --Satcom 08:27, 26 October 2006 (MDT)
Bear in mind that the client does not necessarily have to decode, only encode as the results of the queries are plain text so I'd be looking for encoding functions first as they may be no decoding functions in the client. Also, Craig it'd be useful to know your PID as we need to get from PID to auth if we can find relevant py stuff. --Kurt 08:29, 26 October 2006 (MDT)
- bf2142pid=81273635 you can also search the stat sites for the player Craigins. Also the authcodes were from around 2006-10-26 12:15am (CDT) +/- 5 minutes if you want to check for timestamps as well --Craigins 09:44, 26 October 2006 (MDT)
@Danielphin - base 64 only uses abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ+/ characters for conversion ('=' is used for padding). so if he decoded it, he would have had to convert the underscore and brackets.
also, I came across a possibly related connection (though I dont know how it is if it is). this may be part of the anti-piracy protection (because of the challenge / response api of PACE), but may be used in the auth. just before the game requests stats, it goes through a challenge / response session over tcp w/ stella.available.gamespy.com (havent figured out wich port just yet). the challenge / response sequence has a number of hex, and base 64 (looking) pieces to it (seperated by '/' forward slashes).
@Craigins - I've been trying out different permutations of those auth strings have a running todo list on my site and they seem to be query specific (switching the auth token from one query to another returns weird results).--MadHatter 09:24, 26 October 2006 (MDT)
@MadHatter - if you want I can give you my packet data filtered to stella during the run of the stats check. I noticed they seemed to change on each page it sent to, the packets include the pages they were sent to as well(obviously) --Craigins 09:44, 26 October 2006 (MDT)
sure, feel free to post it here or on my forum (I just put mine on my forum in case they end up exposing some data that I dont want to share w/ the world. so in the instance it does, I can delete it with out having it show up in the history). --MadHatter 09:51, 26 October 2006 (MDT)
GUYS SCREW FINDING THE HASH CODE, just grab the stats off: [[1]] just preg_replace, easy stuff, probably friday afternoon ill make a class for that in php, ill only give it to people who really wants it. Nitr021 14:15, 26 October 2006 (MDT)
Unfortunatly pulling the data off EA's site will leave you with partial data. I couldn't find things such as Time as recon/assault/engie/support skin and # of times you received pins/awards and the dates. It would be nice to have the complete dataset to work with so that the stats will be comprehensive, instead of what EA decided was important to show. --Craigins 14:22, 26 October 2006 (MDT)
Tubar from bfstats.info just posted the disassembly of the code that generates the auth parameter. still have a little work to do on it, but this is pretty huge: http://www.sanity-free.org/forum/viewtopic.php?pid=157#p157. --MadHatter 14:57, 26 October 2006 (MDT)
Seems EA are being a little more protective of their stats than before. http://www.bf2142tracker.com have disabled their feed after nasty words from them. EA have promised an "official" feed at some stage but later rather than sooner. --Kurt 06:56, 27 October 2006 (MDT)
From the news post on bf2142tracker.com it seems like the way they were getting stats was to run each query in game, capture the auth code and then use a script to scrape the data from the server. What leads me to believe this is that he was saying it only worked for 3-4 hours a day, since the auth codes expire quickly it would only work for a certain time after encoded. If we are able to figure out the algorithm used by EA to generate the auth key, they would not be able to distinguish between a call from the server to obtain the information and a call from in game.(ie think about 1 auth key slamming the server hundreds of times in a few minutes, sticks out a lot). If you do that sort of stuff you are suppose to do it in small bursts over the entire day, like 1 minute intervals, and NOT during prime time. The site I'm planning on will only update stats when you request it, it will then copy the stats into an internal database that way it won't hit the EA server each time someone views a page. In fact it would benefit EA's server since people wouldn't be hitting their servers as much, so I have no idea why they have a problem with it. --Craigins 08:23, 27 October 2006 (MDT)
bf2tracker.com and bf2s.com started when bf2 came out. they designed their stats to pull directly from the same servers that the game queries. a short while after that (because they were putting such a strain on the stats servers) EA told them not to use the normal game stat servers. instead what they did was to put up some server that had a replicated db of the stats, and both sites pulled their stats info from there. there are a couple of posts by homey (the guy who runs bf2tracker / bf2142tracker) about waiting for them to put up those servers so they could pull info from it. what I gathered from it when he started showing stats is that either they set up a server for him, or he started pulling stats from the normal game stats server.
I've never planned on setting up a stats server. I like to do my own sig and thats the only reason I'm interested in this (kind of sad eh?). one easy way to bypass EA trying to shut stats sites down (and something I've thought of) would be to create a windows service that can run in the background that watches for the bf2142 process to exit, after a few seconds it queries for the players stats and sends that data via a webservice to a "tracker" site which takes that data and puts it into a db. this way the webserver isnt pulling stats from ea, and there would be no central point to shut down if they dont like other sites doing game stats... just a thought.
in the interest of keeping this wiki up to date as we progress, Tubar showed us the section of code that does the auth token. Its format is as follows (lets break this out into its own section after we get the hashing algorithm psudo code).
it uses a Little-endian 16 byte array (the game uses 4 DWORD / 32 bit integers). the values of those bytes are (in hex):
| location | value stored there |
| bytes 1 - 4 | unix time stamp (in seconds) |
| bytes 5 - 8 | magic number 00000064 |
| bytes 9 - 12 | player id |
| bytes 13 - 16 | UNKNOWN: some sequence that we havent figured out yet |
so that
time stamp - 1161927440 -> 45419B10 magic # - 00000064 pid - 81246737 -> 04D7BA11 other # = 222 173 0 1 -> DEAD0001
turns into
byte[0] = 0x10 byte[1] = 0x9b byte[2] = 0x41 byte[3] = 0x45 byte[4] = 0x64 byte[5] = 0x00 byte[6] = 0x00 byte[7] = 0x00 byte[8] = 0x11 byte[9] = 0xba byte[10] = 0xd7 byte[11] = 0x04 byte[12] = 0x01 byte[13] = 0x00 byte[14] = 0xad byte[15] = 0xde
then the bytes are coverted (the xor bit we used to have here is most likely an init vector of the encryption algorithm),
then those bytes are base 64 encoded ([] for /+ and _ for =) as the auth value.
--MH 09:57, 27 October 2006 (MDT) (lasted edited 16:02, 2 November 2006)
For stats, I just want a page that my clan and I have access to. What EA is probably upset about is they lose out on all the advertising money. If they are the only place people can go to get stats then people will have to view their ads instead of all the other stat sites making money off of the EA content. What EA needs to do is set up an AD server as well, so that if you want the stats you have to take their ADs, then they would get credit for it anyways. That is the only reason I can see them not allowing external sites with stats. At this point its almost worth figuring out the encryption just for the sake of figuring it out. --Craigins 14:03, 27 October 2006 (MDT)
In case no-one has noticed yet, EA have put a public stats site up... http://www.battlefield.ea.com/battlefield/bf2142/
I don't know yet if there's a way to feed the stats into your own web site though.--Mort 00:42, 12 November 2006 (MST)
Yeah, the data returned is consistant in its format - I'm writing a stats retriever and basically using the public stats site as a feed. Of course the data is not designed for parsing, just for display but you can read the html and search for specific phrases to extract the values. Untidy but it seems to work fine. I'm using it for a basic client app but if you were to pull in volume it'd be a pretty untidy way to do it. --Kurt 05:00, 20 November 2006 (MST)
