Google Web Toolkit + Texas Holdem Poker = gpokr.com

August 17, 2006 9:57 PM

Google Web Toolkit + Texas Holdem Poker = gpokr.com. I should probably be embarrassed about how much time I've spent in the last few weeks playing poker online for pretend money. As the site operator mentions in his development blog, it seems to be the small things that make the site so sticky: elegant ajax design, players' rankings displayed and updated right next to their names at the table, a slew of player statistics presented on the main ranking page, even more statistics and graphs on each user page. (Oh, and out of 5000 or so current players, I seem to be #1).
posted by nobody (34 comments total)

Doesn't seem to work with Safari.
posted by Blazecock Pileon at 10:02 PM on August 17, 2006

You should not be ashamed.
posted by crunchyk9 at 10:03 PM on August 17, 2006

Clearly, someone needs to name a table.
posted by cortex at 10:08 PM on August 17, 2006

OMG self link!
posted by delmoi at 10:12 PM on August 17, 2006

also, you need a better icon.
posted by delmoi at 10:13 PM on August 17, 2006

I'm glad I fired up Firefox. Fun stuff.
posted by Blazecock Pileon at 10:23 PM on August 17, 2006

Bragger.
posted by bob sarabia at 10:25 PM on August 17, 2006

The guy went from running the site on his home machine, to a 7 dollar a month godaddy acct, to being features on techcrunch. Crazy stuff.

I bet he gets at least a few job offers out of it.

Still, GWT seems like a bad idea. An additional layer of abstraction over an already high level language (javascript) in an arguably less productive, less powerful language (java).
posted by rsanheim at 10:26 PM on August 17, 2006

i've always hated people who are poker players, but i secretly long to be one.
posted by wumpus at 10:27 PM on August 17, 2006

An additional layer of abstraction over an already high level language (javascript) in an arguably less productive, less powerful language (java).

JavaScript is a right PITA, since the dev tools are largely lacking.

It's easy to blow your foot off in JS, and it is something of a stunted language, development-wise.

I wrote a Hearts client (that replicated the L&F of Windows' hearts.exe) in DHTML for S&G last year. chez fun but I think I'll stick with safer languages.

(though Apple's Dashcode does look interesting)
posted by Heywood Mogroot at 10:39 PM on August 17, 2006

also, you need a better icon.

I'll have you know that my icon is a delectably subtle low-res version of the site's default player icon. I mean, since I'm a bragger, I might as well keep at at.
posted by nobody at 10:41 PM on August 17, 2006

Heywood: No, its actually a great little language and well suited to web development. For awhile browsers screwed things up with their own crappy bugs and incompatible add-ons. With IE 7 and FF 1.5, JS + a library like Dojo or Prototype is a very productive environment for getting stuff done. This is not javascript of 99, gone are the days are cobbling together "DHTML" scripts to do rollovers and pop ups.

See also: this and this
posted by rsanheim at 10:50 PM on August 17, 2006

Still, GWT seems like a bad idea. An additional layer of abstraction over an already high level language (javascript) in an arguably less productive, less powerful language (java).

Javascript is a scripting language, and writing large applications in DHTML and Javascript is not fun. A high level wrapper is just what's needed. If it makes programming easier, it's better. If you want performance, write a client app.
posted by delmoi at 11:11 PM on August 17, 2006

Awful lot of fish there. I went fro $1500 to $4000 on my first hand after watching bad play after bad play.

You'll be #2 in about a week, nobody. ;-)
posted by solid-one-love at 11:12 PM on August 17, 2006

I'll have you know that my icon is a delectably subtle low-res version of the site's default player icon. I mean, since I'm a bragger, I might as well keep at at.

I did notice that after poking around the site a bit. It looks like a lot of people riff on that standard icon, so I withdraw my criticism.
posted by delmoi at 11:13 PM on August 17, 2006

posted by Milkman Dan at 11:20 PM on August 17, 2006

Delmoi: not sure if what you are trying to say here...did you know that there are a lot of large apps written in scripting languages, and they do just fine?

Also, a "high level wrapper" is exactly what I was talking about when I mentioned the javascript frameworks of Dojo or Prototype - they build on the language instead of trying to pretend its not there. If people want to write their apps thru a Java to Javascript framework, thats fine. If you talk to most experienced web developers who know both java and javascript (and many other languages), they will tell you they prefer JS over GWT or Echo2 every time.
posted by rsanheim at 11:25 PM on August 17, 2006

Delmoi: not sure if what you are trying to say here...did you know that there are a lot of large apps written in scripting languages, and they do just fine?

Yeah, but not Javascript. Writing java code is so much easier then writing Javascript. Plus, if you write the Javascript manually you also need to write the server side stuff yourself in another programming language, and come up with how you want to do communication and so forth. If you have a wrapper, or something that generates Javascript for you then you don't need to worry about figuring out the client server communication for you, writing a web app becomes like writing a client app.

If you talk to most experienced web developers who know both java and javascript (and many other languages), they will tell you they prefer JS over GWT or Echo2 every time.

Sure, but obviously if you talk to experienced client app writers, they'll tell you they prefer writing web apps like they write their client apps (I would imagine).

Obviously people are going to stick with what they're used too. I'd rather use something like GWT then try to wrangle javascript
posted by delmoi at 11:30 PM on August 17, 2006

I got on and within an hour was $20,000 up, and ranked 180th or so.


SO MANY BAD POKER PLAYERS.
posted by stenseng at 12:16 AM on August 18, 2006

Speaking of, I just lost it all because I didn't know which hand beats which other hands.
posted by beatrice at 12:24 AM on August 18, 2006

I can't get it to work, dammit.
posted by The Card Cheat at 1:06 AM on August 18, 2006

gopokr? I hardly knowr.
posted by pax digita at 4:50 AM on August 18, 2006

It won't work properly on Opera, and on Firefox, the chat window just extends further and further below the bottom of the screen.
posted by alloneword at 7:40 AM on August 18, 2006

SO MANY BAD POKER PLAYERS.

/Waves hello, knocking over the unsuited three and nine I'm expecting the river to redeem
posted by Alvy Ampersand at 8:14 AM on August 18, 2006

Three nine offsuit? Fuck.

*drops two seven, stomps off*
posted by cortex at 8:24 AM on August 18, 2006

Unsuited, offsuited, whatever...

I raise- no, I mean fold, fold!
Can I check?
posted by Alvy Ampersand at 8:45 AM on August 18, 2006

Lessons to play pokr bettr:

1) Never fold. Raise! You came to play, not to fold!

2) Never raise with a good hand, you'll scare off the fish. Just call calmly to the river.

3) Your flush draw pays off one third of the time. This means you just have to win three times the money off a guy to make it a winning propositiong. Therefore, stay in for all flush draws and go all-in when you make it. (Go all-in sometimes when you don't. Bluffs are awesome!)

4) Haxxor the xml and make all aces come into your hand. You're leet, right?

5) When you get beaten, LEAVE THE SITE IMMEDIATELY. Obviously the site is cheating to give your money away, and you'd be better off on another site that isn't cheating. (NB: All sites cheat. Also, make sure to let all the players at your table know that you were cheated before you leave. They'll thank you for that valuable information.)
posted by Imperfect at 9:46 AM on August 18, 2006

Good gracious, it's actually quite good. Got some nasty sharp edges left over from a rough'n'ready beta, but this is something of an embarrassing open-mouthed moment.

Still, why are there so few artists working with these new buzz sites? I tire of sky blue.
Not so much that I'd stop playing poker, but I tire.

Actually, the empty chipstack stops me.
posted by NinjaTadpole at 10:38 AM on August 18, 2006

Some funny stuff there, Imperfect.

In all seriousness, if you're a duffer, the following tips will improve your poker game dramatically:

0. Play tight. Tighter than that. Really tight. If you're seeing the flop 25% of the time in a nine player game, you're probably not tight enough.

1. Out of the blinds, play only the following hands: pocket pairs, hands where both cards are greater than 9, suited Ace combos and suited connectors higher than 4-5. All other hands are marginal at best. Skilled players can get away with playing hands like K6 suited or Q9 offsuit, but you can't. Don't. Fold all other hands. I know you want to play that A8 offsuit. Don't.

2. Pocket pairs under 88 cannot stand a preflop raise. A pair of threes is absolutely worthless if a 3 doesn't come down on the flop and very valuable if it does. You'll hit your trips something like 6% of the time and take a pile of money. That's worth paying the blind, but not a preflop raise. And when I say raise, I mean raise it 2 to 5 times. If the blinds are at 50 and you have KK, bet somewhere between 100 and 250.

2. Raise AA, AK, pocket pairs over 88 preflop in an unraised pot. Sometimes AQ. Again, 2 to 5 times the previous bet. So, you're in late position with AK. The big blind is 50. A guy in middle position raises to 150. You should raise to somewhere between 300 and 750.

3. Re-raise AA, AK, KK preflop.

4. Hands like KQ, KJ, K10, AJ, A10, QJ and Q10 are trap hands. You have to play them very carefully. You can lose a lot of money with them.

5. All-ins are usually coinflips. Try to avoid going all-in or calling an all-in unless you have a monster hand (QQ, KK, AK, AA) or have a stack many times larger than your opponent. If you've got 7,000 chips and he goes all-in with 1,000, call his all-in with anything, because no hand is a 7:1 dog.

6. Your suited connectors will not hit very often. When they do, you will make a fortune.

7. Call the small blind with any two cards. If you have a trash hand, fold it if the big blind raises. Otherwise, in the long run, it's worth seeing the flop.

8. The only times you should chase are: when you have four to a flush; and when you have top pair and an open ended straight draw. Skilled players can get away with chasing with other hands. You can't. Yeah, you might hit a Queen on the river to make top pair and take down an unexpected pot, but in the long run, you'll lose a fortune and a stupid play like that is blood in the water. You become a target. You gonna get bust, because now the table knows that you're an idiot.

9. It's worth leading to the pot when the players behind you check, even if you have a nothing hand. You will take the pot more often that not. It pays off in the long run.

10. You have to gamble to win. But you have to minimize your risk to avoid losing.

11. That guy who raises every hand will fold to your re-raise when you have the cards to make one. And then he'll stop for a while. Until then, let him steal the blinds.

And once you've stopped the bleeding, bone up on positional play, and learn why A10 is a valuable hand in late position but a dog in early position.
posted by solid-one-love at 11:26 AM on August 18, 2006 [2 favorites]

Due to numbering, my tips for raising were shifted slightly. it should read:

2. Pocket pairs under 88 cannot stand a preflop raise. A pair of threes is absolutely worthless if a 3 doesn't come down on the flop and very valuable if it does. You'll hit your trips something like 6% of the time and take a pile of money. That's worth paying the blind, but not a preflop raise.

3. Raise AA, AK, pocket pairs over 88 preflop in an unraised pot. Sometimes AQ. And when I say raise, I mean raise it 2 to 5 times. If the blinds are at 50 and you have KK, bet somewhere between 100 and 250.

4. Re-raise AA, AK, KK preflop. Again, 2 to 5 times the previous bet. So, you're in late position with AK. The big blind is 50. A guy in middle position raises to 150. You should raise to somewhere between 300 and 750.

(and the rest should be renumbered.)
posted by solid-one-love at 11:33 AM on August 18, 2006 [1 favorite]

If you've got 7,000 chips and he goes all-in with 1,000, call his all-in with anything, because no hand is a 7:1 dog.

Awesome post, I'm just not sure I understand the reasoning here. If the pot is 1000 after his raise, and it costs you 1000 to call, you better make sure your hand is at least a coin flip to win (because pot odds = 1:1). Or else you'd end up losing money in the long run, wouldn't you?
posted by Nquire at 6:27 PM on August 18, 2006

Against any two cards, no card is worse than 29% to win, or something like that. Online, people will make deserate all-ins, so in the long run, if you are far in the lead, youwill make money because knocking a player out will increase your odds of a piece of the pot.
posted by solid-one-love at 6:39 PM on August 18, 2006

Better explanation: Against any random two cards, even a 7-2 offsuit has a 29% chance to win. Sure, its chances suck against AA, but you until he flips 'em, he could have anything. Online, with a short stack, what he has is as likely to be J4o as AKx.

Always pressure the short stacks.
posted by solid-one-love at 6:41 PM on August 18, 2006

I'm having wierd freeze-ups with Firefox on OSX. Any suggestions?
posted by TheCoug at 11:15 AM on August 19, 2006