It is currently Sun Nov 18, 2018 9:52 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 49 posts ]  Go to page Previous  1, 2, 3  Next
Author Message
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #21 Posted: Thu Nov 08, 2018 11:52 am 
Judan

Posts: 5287
Location: Cambridge, UK
Liked others: 295
Was liked: 2776
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Little test of #188 on a Macbook Pro (CPU only version of LZ), was a 15s game so that's only about 50 playouts per move if it's unexpected, got up to a few hundred when it can reuse tree with expected moves. Lost due to creating a ladder which didn't work (it wanted to play the failed atari but I played another choice tenuki to see if there was any chance to get winrate climbing up but there wasn't so resigned at about 6%. Opponent was a 65 win, 2 loss player who was also a computer as I found out via google translate chatting in Chinese after :lol:



Game chat:
Quote:
leelazero7 [8D]: Hi, I am LeelaZero #188 on macbokpro
商业围棋 [8D]: 苹果电脑?= Apple computer?
leelazero7 [8D]: 是的,没有GPU所以无法读取梯形图 = Yes, there is no GPU so the ladder cannot be read.
商业围棋 [8D]: 你输多了 = You lost more
商业围棋 [8D]: 投降吧 = Surrender
‘商业围棋’ W+ Res ‘leelazero7’
leelazero7 [8D]: Thanks
商业围棋 [8D]: 没有gpu打不上9d吧 = No gpu can't hit 9d.
leelazero7 [8D]: 你是人还是电脑? = Are you human or computer?
商业围棋 [8D]: 我也有电脑哈哈 = I also have a computer haha.

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #22 Posted: Thu Nov 08, 2018 6:55 pm 
Beginner

Posts: 7
Liked others: 3
Was liked: 4
Uberdude wrote:
Well, that was weird: just had my first loss, and LZ thought it was winning the whole game but lost according to Fox by 17 points, and that was a no komi game as black against an 8d. It lost winrate from the lower side fighting, but then thought it recovered. Counting the game independently with CGoban and myself I get a black win by 17, a Fox bug? At the end of the game I clicked the Count button and then a dialog for automaric counting came back saying "W+17 'leelazero7'" so the colour of winner was wrong but name right. I, perhaps stupidly, clicked ok because another game I won on time displayed the wrong B/W colour so I assumed it was a similar UI glitch and wouldn't actually make the game result incorrect. Is it possible in Chinese client for a player to change the winner field?That's annoying to lose the chance to double promote to 9d.



Attachment:
loss winrate.PNG


This was a game using a special rule set of captures counting double the points.

If you look down your game list, there is a column that says "none". This game has "一子千金", meaning this special rule set applied to this game.

If you don't want to play these type of games anymore, careful not to accept auto matches with the word "一子千金" at the top of the box


This post by hitlab was liked by 3 people: dfan, Gomoto, Uberdude
Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #23 Posted: Fri Nov 09, 2018 1:57 am 
Judan

Posts: 5287
Location: Cambridge, UK
Liked others: 295
Was liked: 2776
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Thanks hitlab! I was aware there was that new mode, but didn't realise you could get automatch challenges from it if you never clicked it yourself. I'll look out for that word in automatches in future.

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #24 Posted: Fri Nov 09, 2018 11:30 am 
Judan

Posts: 5287
Location: Cambridge, UK
Liked others: 295
Was liked: 2776
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Another #188 with low playouts (bit more this time as 30s byo), easy win vs 7d (luckily the ladder on opening was good for white!) and even just managed at end of byo-yomi period to find the shortage of liberties tesuji with move 84 as superior to capture (but both >97%). And then a nice kill at top side with another of shortage of libs, a classic tsumego that I probably wouldn't spot unless you told me it was a tsumego problem but 188 spots with just a dozen playouts, so quick to see the vital shape points.


Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #25 Posted: Fri Nov 09, 2018 3:46 pm 
Lives in gote

Posts: 549
Liked others: 61
Was liked: 247
Rank: AGA 5 dan
The life and death problem on top was nice, but perhaps not classic. B123 was suicide, pretty much forcing W to take the killing liberty :) B could play F18 instead to live.


This post by mitsun was liked by: Uberdude
Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #26 Posted: Fri Nov 09, 2018 4:24 pm 
Judan

Posts: 5287
Location: Cambridge, UK
Liked others: 295
Was liked: 2776
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Nicely spotted mitsun! Analysing now on my more powerful PC LZ 188 does find the f18 atari after about 400 playouts, but it still doesn't become the #1 playouts move after 30k. It does have a higher winrate of 2.7% vs 2.3%, but seeing as everything loses it doesn't really matter. I mean more a classic technique of tsumego of creating the shortage of liberties to keep the eye: it's something I'm aware is a bit of a blindspot for me (when reading many moves before I'm thinking this space is an eye and this space is an eye and don't think about that some stones are in self-atari in the future).

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #27 Posted: Fri Nov 09, 2018 5:16 pm 
Judan

Posts: 5287
Location: Cambridge, UK
Liked others: 295
Was liked: 2776
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
So I played a 20s game on my 1060 with #188, so that's about almost 2k playouts per move, plus an extras from pondering and reuse from previous moves when things go as expected. It played an atari for a ladder which didn't work in another 3-3 invasion joseki, and when the 8d pulled out then it had enough playouts to realise the ladder didn't work and compromised for a bad result, but then a dozen moves later was winning again (wth was white e10?! says q10 would keep him at 60%). My understanding of this joseki is if the ladder doesn't work black's hane at n18 is questionable as it's white right to choose this variation and if black can't ladder it's hard to deal with the cutting stone (moreso as white already has r12). But reading the ladder back at the time to play n18 or another choice like p18 seems to be beyond LZ 188 with this many playouts. Another interesting thing: before playing the o8 ladder LZ was looking at forcing with r8 first (higher winrate but low playouts), and as it happened with the game continuation that would've been a very nice kikashi.


Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #28 Posted: Fri Nov 09, 2018 7:13 pm 
Lives with ko

Posts: 170
Liked others: 126
Was liked: 97
Uberdude wrote:
wth was white e10?!


Misclick?

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #29 Posted: Sat Nov 10, 2018 5:12 am 
Dies with sente

Posts: 109
Liked others: 27
Was liked: 11
Rank: KGS 6 dan
@Uberdude
I noticed that in many of your games there is at least one ladder issue. Have you considered trying modification of LZ that knows ladders? It recognizes ladders in less than 50 playouts and uses a bit of cpu for that recognition (and gpu is used for position evaluation). The network itself is 20 blocks and 192 filters. It has ~60% winning score against the strongest 15 block network. All credit goes to ttl (github), jio (discord), ancalagon (discord). It works with Lizzie v0.5 (not v0.6 because it was built before the latest release).
You can download it here: https://we.tl/t-VKZh5CQq9d (wetransfer.com)

Example of your latest game:
Attachment:
sshot-1.jpg
sshot-1.jpg [ 169.78 KiB | Viewed 538 times ]

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #30 Posted: Sat Nov 10, 2018 6:27 am 
Lives in sente

Posts: 889
Liked others: 0
Was liked: 145
At this point, experimentation with the neural nets seems centered on "zero" versions of training. We need to keep in mind that the goals/objectives of neural net experimenters is not necessarily to as soon as possible come up with the strongest go playing nets as it is to study how nets learn something as difficult as go.

It is am interesting property of neural nets that prior (even damaged prior) training affects subsequent learning/(re)learning.

The open question is what WOULD constitute "ladder training". And at what stage of zero training should the zero training be interrupted for a course of "ladder training"

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #31 Posted: Sat Nov 10, 2018 8:51 am 
Judan

Posts: 7812
Liked others: 2143
Was liked: 2742
Mike Novack wrote:
It is am interesting property of neural nets that prior (even damaged prior) training affects subsequent learning/(re)learning.

The open question is what WOULD constitute "ladder training". And at what stage of zero training should the zero training be interrupted for a course of "ladder training"


My thinking, which is hardly original, going back decades, is adversarial training instead of or in addition to self play training. If LZa is weak on ladders, its adversary, LZb, could learn to set up ladders that LZa misevaluates.

Edit: Also, go requires a number of skills. Go AI of the future may have different modules for different skills. While in theory neural networks can learn all about ladders, given infinite time, it may be more efficient to have a module that reads ladders, probably among other things.

_________________
The Adkins Principle:

At some point, doesn't thinking have to go on?

— Winona Adkins

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #32 Posted: Sun Nov 11, 2018 4:04 am 
Judan

Posts: 5287
Location: Cambridge, UK
Liked others: 295
Was liked: 2776
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
@johnsmith
Thanks for the link, I'll check it out.

So I got to 9d, and it's a lot tougher to win there, but that's because lots of the other 9ds are bots too! I just lost 3 games in a row by half a point using #188 on my 1060 GPU (pretty sure at least 2 of those were using bots, many moves matched LZ's, plus game chat). As LZ uses 7.5 komi and Fox uses 6.5 I thought there might be some instance where LZ as white thinks it's going to win to win but actually loses but it's not happened yet (also LZ black in that situation might throw away an extra point if it thought it was going to lose anyway under 7.5 komi which actually turns a 6.5 komi win into a loss). I beat one clearly human 9d fairly smoothly.

Quote:
leelazero7 [9D]: Hello! Pleased to meet you.
GCTY [9D]: Hello! Pleased to meet you.
leelazero7 [9D]: I am LeelaZero #188 on GTX 1060
GCTY [9D]: ?
leelazero7 [9D]: 我是电脑玩家 = I am a computer player
GCTY [9D]: 喔,我人机结合 = Hey, my man and machine combine
leelazero7 [9D]: 你也是电脑玩家?= Are you also a computer player?
GCTY [9D]: 看情况,有时牵狗、有时纯人、有时人机结合,多种玩法 = Look at the situation, sometimes holding dogs [dog is Chinese slang for AI] , sometimes pure people, sometimes man-machine combination, multiple gameplay


Here's the game and winrate from the game with GCTY. My LZ did think it was winning at some points but I think that's just a poor evaluation missing opp's best replies because when he played good unexpected moves like s12 instead of s13 the winrate dropped. Also my LZ didn't expected 130 but thought black could sacrifice the s15 stones and get a squeeze including capturing r11 for a little profit and thickness, but white prevented that and kept r11 alive. Move 96 L13 was an LZ move he played super quickly which I'd be really impressed with a human to find independently.



Attachment:
3rd half point loss winrate.PNG
3rd half point loss winrate.PNG [ 148.1 KiB | Viewed 440 times ]


To me playing with computer assistance, unless up-front about it (hence my username and message at game start), is cheating, but maybe it's normal on Fox (or maybe their Chinese username reveals that they use AI but I can't read it), plus so far they do admit it when asked with my broken Chinese from google translate. <insert controversial statement about cheating being more acceptable in Chinese culture>. Can anyone translate this from their user info page?

Attachment:
chinese name.PNG
chinese name.PNG [ 13.44 KiB | Viewed 440 times ]


Attachment:
jpn crop.png
jpn crop.png [ 2.68 KiB | Viewed 440 times ]


This game my LZ didn't see the s2 tesuji (just expected p3 which is the normal shape in the corner, but that usually doesn't have white already cutting and turning at s9), which is a really sharp move exploiting a shortage of liberties with the s5 cut, but once played the winrate collapsed and it never recovered. My LZ did manage to make a nice centre wall in exchange for all the death, but his reduction was excellent (and mostly as my LZ 188 expected).

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #33 Posted: Sun Nov 11, 2018 6:36 am 
Beginner

Posts: 2
Liked others: 0
Was liked: 3
I'd like to help with the Chinese information issue.
Uberdude wrote:
Attachment:
chinese name.PNG


The message is interesting. It's the second half of a Chinese ancient poetry.
You may know "绝艺(Fine Art)". The name "Fine Art" is derived from the first half of the same poetry.

About the username "石庚" seems to be a normal name. After a little search, I found a website of someone says that his nickname and Fox id is 石庚. He said he is an amateur 5 dan and a private tutor of go.


This post by sporeyang was liked by 2 people: ez4u, Uberdude
Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #34 Posted: Sun Nov 11, 2018 6:42 am 
Oza

Posts: 2216
Liked others: 14
Was liked: 3175
Quote:
To me playing with computer assistance, unless up-front about it (hence my username and message at game start), is cheating, but maybe it's normal on Fox (or maybe their Chinese username reveals that they use AI but I can't read it), plus so far they do admit it when asked with my broken Chinese from google translate. <insert controversial statement about cheating being more acceptable in Chinese culture>.


As a sidelight on the above, I read the following in connection with Game 2 of the current World Chess Championship here in London.

Quote:
...in Baguio City the two players reached the position after White's 10th move three times, with Karpov always Black. The champ played 10...Be7 twice and drew, but once played 10...Re8 with a fantastic knight sac that wasn't merely Tal-like. It was from the actual Mikhail Tal (Karpov's second)!

Carlsen said he saw a "very clear parallel" between his situation and Korchnoi's shock... In fact, the generational difference worried him even more; he explained that his predicament could be even more dire since the weaponry was surely made partly of computer silicon and not just human ingenuity.


So, Carlsen was caught unawares by a move that was apparently the result of home preparation using a computer (which has long been normal in chess, of course). Is that cheating? I don't recall hearing anyone say it is.

But if it is, what's different about that and a player using a human second (Tal here, i.e. world champion level, a bit more than just an assistant trawling through chess literature for new moves; and, as such, it seems rather more than, say, an athlete/coach relationship to me)? I don't think anyone said that was cheating either.

It may be that those who think its cheating (in either case) have left chess in disgust and so we never hear their views on chess forums. But there its seems accepted as normal.

I'd be disinclined to believe that cheating is regarded as more acceptable by Chinese people in general, but (as with Karpov + Tal) I'd be inclined to believe that the state (any state, or people believing they are acting in the interests of the state) would be willing in certain cases to, shall we say, bend the rules as far as they will go. Go pros maybe should expect in future to "live in interesting times," to subvert the Chinese saying - but maybe they even want to, like chess players.

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #35 Posted: Sun Nov 11, 2018 11:18 am 
Lives in gote

Posts: 328
Liked others: 0
Was liked: 77
Rank: KGS 2k
GD Posts: 100
KGS: Tryss
@John : he's not talking about using a computer to prepare before the game, but using a computer during the game.

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #36 Posted: Sun Nov 11, 2018 12:17 pm 
Oza

Posts: 2216
Liked others: 14
Was liked: 3175
Quote:
@John : he's not talking about using a computer to prepare before the game, but using a computer during the game.


If you mean Carlsen by "he" (as opposed to uberdude), I can't agree, unless there's part of the press conference I missed. It's obvious this shocking move (which was known to Carlsen, and Caruana's side would know that) was resurrected after computer research, which is why Carlsen felt terrified.

If you mean uberdude, yes but it's only a matter of time before computer preparation becomes common-place in go (it's happening already but is not reliable enough yet to be common-place). And once that happens, I'd expect online cheating during games to become the norm (I gather that in chess you can assume your opponent has either prepared something on the computer or is using one live, so the only defence is seen to be to use a computer during the game yourself. Then the question is: is self-preservation a heinous or an excusable form of cheating?)

It's a good topic for an ethics course.

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #37 Posted: Sun Nov 11, 2018 2:27 pm 
Lives in sente

Posts: 1261
Liked others: 568
Was liked: 368
Rank: AGA 4k KGS 2k
GD Posts: 61
KGS: dfan
As a chess player who has been used to this stuff for 20 years it's a little odd to me to see the use of computers during games and the use of computers between games be compared as lying on some sort of continuum. In chess, the rules have always been:

  • Using a computer during a game = 100% cheating, unless it's a correspondence server where it is explicitly allowed. (And yes, this sort of cheating is indeed rampant online and continually being actively combated by the servers.)
  • Using a computer between games to analyze openings and come up with new ideas = 100% acceptable.

Quote:
In chess you can assume your opponent has either prepared something on the computer or is using one live, so the only defence is seen to be to use a computer during the game yourself.

In chess you assume your opponent has prepared something with the aid of a computer (at the very least to double-check their own ideas), and your defense is to play well (and look at likely lines in advance).

Chess is not generally so fragile a game that one unexpected opening move coming out of nowhere decides the course of the game.

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #38 Posted: Sun Nov 11, 2018 8:43 pm 
Lives in gote

Posts: 546
Location: Indiana
Liked others: 92
Was liked: 153
I started playing postal correspondence chess in the 1960s, mainly owing to lack of strong players in my region. When computer services, like AOL, Genie, etc., arrived I played there. Finally, after the internet became available, I joined the Internation E-mail Chess Club, where I reached a rating of 2296 before I quit more than a dozen years ago. Why did I quit, you ask? It felt more and more as though I was playing against computers, based on the "style" of play -- and perhaps a modicum of paranoia. For me, at least, computers ruined correspondence chess. Their prevalence also contributes to my reluctance to take up correspondence Go. But then I need an excuse to get out of the house after being married more than 25 years... ,-)

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #39 Posted: Mon Nov 12, 2018 12:18 am 
Beginner

Posts: 12
Liked others: 4
Was liked: 0
Rank: 15k
Uberdude wrote:
I just lost 3 games in a row by half a point using #188 on my 1060 GPU (pretty sure at least 2 of those were using bots, many moves matched LZ's, plus game chat). As LZ uses 7.5 komi and Fox uses 6.5 I thought there might be some instance where LZ as white thinks it's going to win to win but actually loses but it's not happened yet (also LZ black in that situation might throw away an extra point if it thought it was going to lose anyway under 7.5 komi which actually turns a 6.5 komi win into a loss).

You might try alreadydone's fork of leela that uses a hack to approximate dynamic komi. LZ is trained on two komi values: 7.5 and -7.5 (for the opponent), so alreadydone's fork just interpolates between and extrapolates from those two values. However, it relies on the winrate increasing monotonically with komi, which doesn't happen in all LZ networks. In other words, some LZ networks are better than others for this purpose.

From the readme:
Quote:
How would I use the engine to play games with 6.5 komi as white (e.g. on Fox) or other values?

You may be tempted to use --target-komi 6.5, but due to inaccuracy of winrates under komi values other than +/-7.5, it's safer to set --target-komi 3, for example.

Top
 Profile  
 
Offline
 Post subject: Re: LeelaZero adventures on Fox
Post #40 Posted: Mon Nov 12, 2018 2:09 am 
Tengen
User avatar

Posts: 4487
Location: Chatteris, UK
Liked others: 1566
Was liked: 641
Rank: Nebulous
GD Posts: 918
KGS: topazg
John Fairbairn wrote:
If you mean uberdude, yes but it's only a matter of time before computer preparation becomes common-place in go (it's happening already but is not reliable enough yet to be common-place). And once that happens, I'd expect online cheating during games to become the norm (I gather that in chess you can assume your opponent has either prepared something on the computer or is using one live, so the only defence is seen to be to use a computer during the game yourself. Then the question is: is self-preservation a heinous or an excusable form of cheating?)

It's a good topic for an ethics course.


This is an interesting issue. Having played OTB and correspondence chess a fair bit, although never at a very high level, there are some parallels and there are some differences between the situations. In OTB chess, it's been generally considered that any form of assistance of any kind during the game by a book, computer, or anything else, is unacceptable. I understand that when chess title matches went overnight though it was considered acceptable to discuss the position with seconds (I suppose if it's a world championship, the theory is that the second shouldn't be a fundamentally stronger player anyway), but that was pretty much the extent of acceptable external influence. Outside the game I've never heard of complaints from using chess opening databases or using engines to test new ideas on existing theory. This is sort of considered to be "how everyone does it", and almost all the opening novelties these days come from players (or their seconds) coming up with a novel idea with practical appeal, and testing it out on engines until they find lines they're happy with, which then subsequently get memorised. There is certainly no acceptability for using engines in the game itself once the players sit down.

In correspondence chess, it was a bit looser. It was equally strongly frowned upon to use engines in game (and most correspondence sites will perma-ban a player caught doing it - a lot explicitly disallow their use even if advertised specifically as an engine player). The only difference is people would expect opening databases to be used quite extensively during the game, which gives a lot of very high level grand master games to analyse and draw from .... but doesn't give the insights of a "player" several hundred ELO stronger as the lines are playing out. I had a good experience with a certain Italian correspondence grand master in the early 2000's who was very keen on playing computer engines in correspondence chess. He was very successful at finding their weak points in analysis and creating positions they'd misplay, and seemed to thrive on the challenge. Of course, chess engines now are about 1000 ELO stronger than they were then, and I suspect it's not something that's really possible now, but other than him most other players seemed to find the idea of playing engines either just uninteresting or outright dirty.

In Go, I find it interesting that there was no objection (other than it making the player weaker) to a certain teammate using a joseki book in the children tournaments in Hikaru No Go. The equivalent in chess would definitely be disallowed. The whole idea of bots being simply superior to humans at Go is such a novel issue that I suspect the ethics side of it hasn't caught up with reality yet. I have no idea if there are cultural tendencies towards finding it "acceptable if explicitly declared", acceptable full stop (outside of issues where clearly someone is sandbagging by playing as themselves normally and then turning the bot on now and again just to destroy their opponent), or unacceptable. I would assume tournaments and serious play will have to be cautious of the possibility (maybe even to the point where devices to record the game on are banned, as it will be relatively easy to set up a private web server to suggest moves to you instead of just being innocently entering your game as you play), but otherwise I suspect the only people encountering the issue will be at the very, very top end of play.

ADDENDUM: Actually, one thing that interested me in the discussion between Andrew and the 9d who clearly admitted to using a bot. There is a side discipline in chess that has seen very little attention called Centaur chess (actually, now called Advanced chess I believe). It's quite possible with Alpha Zero and other neural network developments in chess that such a setup will cease to exist as AI becomes simply far too strong for humans to even help with, but the idea of centaur chess was a hybrid team, where the strong players would deliberately guide the engine into lines where it would thrive. People would develop opening books purely for their choice of engine to get positions that it was extremely strong in against other engine players in complicated tactical situations, and often had a strong titled player to help pick choices that a computer engine was less likely to misplay. For a while there were some very interesting hybrid tournaments held, but as of about 2-3 years ago (and pre alpha zero) the popularity was declining as pure engine play was becoming so strong that it was beating even GM/engine teams.

I do somewhat wonder if the comments in the 9d game imply that at the moment there is some real possibility for professional strength players and engines to form an interesting hybrid team. I would be very interested Andrew, if you did a best of 21 or something against that guy, whether the match result would be close or not.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 49 posts ]  Go to page Previous  1, 2, 3  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 10 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group