Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Slack Poker Bot

github.com

Post reply on HN
Print view View original

91–100 of 100 posts

Re: Slack Poker Bot

#91

Post by jeeva »

Vexs wrote: post #33
GauntletWizard wrote: post #6

Earlier quoted context omitted.

Possibly because running a bot client is much different from running a webserver. Dialing outbound to connect to slack is much easier than figuring out or configuring your own canonical URL, opening ports, etc. On the other hand, preparing hands as a single image seems a bad way of doing this; Does the slack API only allow a single image per conversation? I agree with your preparation microservice idea - There should…

I saw it on HN a while ago- a webservice that does poker hands using a URL; you could pass it cards in the URL and it would display them.

deckofcardsapi.com, if it's the one I'm thinking of?

Re: Slack Poker Bot

#92

Post by dalore »

shultays wrote: post #90
flakmonkey wrote: post #46

Am I correct in saying that the shuffle algorithm is flawed? It looks to me like it swaps each card iteratively with a random card in the deck. This means that some cards are likely shuffled more times that others, and some sequences are then more common than others. https://github.com/CharlieHess/slack-poker-bot/blob/master/s...

Does this add a bias though? If I took first card in deck and swap it a hundred times more than others how would that make a difference?

Yes it does. http://blog.codinghorror.com/the-danger-of-naivete/

It's random but some results are overrepresented so more likely to occur. If you knew this and bet on this over time you would gain an advantage

Re: Slack Poker Bot

#93

Post by Pwntastic »

CharlieHess wrote: post #58
Pwntastic wrote: post #55

I can't seem to get it installed on windows 8.1... The lwip package in one of the other dependencies fails to build :( sad day

Does the Heroku deploy work for you?

The Heroku deploy does appear to work, but I was hoping to get it all running locally. Thanks though! It looks like there's a ticket on the lwip project on github that is similar to my issue, so hopefully they'll resolve that soon enough :)

Re: Slack Poker Bot

#94

Post by 91bananas »

vacri wrote: post #77
91bananas wrote: post #32

Earlier quoted context omitted.

Just a way better interface on it. Searchable history, multiple rooms inside rooms effectively. File sharing, lots of niceties that are probably possible in IRC but would take a lot of work. It has changed the way our team communicates, for me I am the only remote guy on a team of 6ish devs. I can pick and choose my hours, work when other guys are off, and still see all the comms that happened while I wasn't at my ma…

I love everything about Slack except the damn chat window. Oh, and the pricing. The onboarding process is buttery smooth, and there are tons of integrations. But there's so little in the way of communication density (even in 'compact mode') that I miss IRC. We have a jira integration set up and every ticket consumes five lines plus plenty of whitespace - a few of those and there's no real comms visible on the page. E…

I hear those complaints, we found them early on too. We have since separated operational noise in to other channels that ONLY contain integration related stuff, and try to keep one channel purely for actual chatter.

Re: Slack Poker Bot

#95

Post by 91bananas »

girvo wrote: post #56
aaron-lebo wrote: post #36

Earlier quoted context omitted.

> I can pick and choose my hours, work when other guys are off, and still see all the comms that happened while I wasn't at my machine. I'm sure are features IRC cannot duplicate, but all that's necessary for that is a bouncer.

Right. Think of Slack as a really well configured IRC server, except you don't have to manage it at all, and it's front-end UI is much more modern (and better); integrated file uploading, inline media, integrations with everything under the sun, etc. None of it is impossible in IRC, it's just _easier_ in Slack because it's out of the box like that. And, most importantly: I don't have to manage the damned server ;)

Yup, totally to your last line. That is it entirely, all of this stuff is probably possible with a ton of time invested into IRC.

But we have shit to build! :)

Re: Slack Poker Bot

#96

Post by aw3c2 »

girvo wrote: post #86
MichaelGG wrote: post #71

Earlier quoted context omitted.

Are there open source/open protocols for something like Slack? Rich IRC?

There are a couple of protocols that do similar things, but nearly all of them have a major fatal flaw, that is unfortunately very difficult to work around through no fault of their own: mobile push notifications, especially on iOS. An open protocol is useful, but not enough, and that's where Slack hosting and developing everything themselves has an advantage, despite open protocols and tools being what I'd normally…

Matrix has push, at least on Android.

Re: Slack Poker Bot

#97

Post by todd8 »

flakmonkey wrote: post #46

Am I correct in saying that the shuffle algorithm is flawed? It looks to me like it swaps each card iteratively with a random card in the deck. This means that some cards are likely shuffled more times that others, and some sequences are then more common than others. https://github.com/CharlieHess/slack-poker-bot/blob/master/s...

For those interested, here is the reasoning from a combinatorics perspective. There are 52! different arrangements of a deck of cards. To see this, note that in any arrangement there are 52 locations for the Ace of Spades, once that choice is made there are 51 locations for the next card to end up in and so on until the last card must go in the only open location. Hence there are 52 factorial arrangements of cards in to a deck.

Shuffling, by swapping each card with any of the 52 other spots produces 52^52 (i.e. 52 * 52 * 52 * ...) arrangements. How can this be when there are only 52! (i.e. 52 * 51 * 50 * ...) possible arrangements. The answer is than many of the arrangements generated by this shuffling technique end up with the cards in the same order. This follows from the pigeon-hole theorem since 52^52 > 52!

Furthermore, we know that 52^52 is not a multiple of 52! (to see this realize that, for example, 11 divides 52! but not 52^52). Therefore, some bad shuffle generated arrangements will occur more than others and, consequently, that bad shuffle algorithm does not produce a "fair" shuffle.

Re: Slack Poker Bot

#98

Post by tomaskafka »

vacri wrote: post #77
91bananas wrote: post #32

Earlier quoted context omitted.

Just a way better interface on it. Searchable history, multiple rooms inside rooms effectively. File sharing, lots of niceties that are probably possible in IRC but would take a lot of work. It has changed the way our team communicates, for me I am the only remote guy on a team of 6ish devs. I can pick and choose my hours, work when other guys are off, and still see all the comms that happened while I wasn't at my ma…

I love everything about Slack except the damn chat window. Oh, and the pricing. The onboarding process is buttery smooth, and there are tons of integrations. But there's so little in the way of communication density (even in 'compact mode') that I miss IRC. We have a jira integration set up and every ticket consumes five lines plus plenty of whitespace - a few of those and there's no real comms visible on the page. E…

Could you run Slack web in Chrome, and find or make a userstyle that would tighten it up?

https://userstyles.org/styles/browse/slack

Re: Slack Poker Bot

#100

Post by FireBeyond »

mlex wrote: post #40
FireBeyond wrote: post #21

Earlier quoted context omitted.

When I played poker at MSFT in WA state, another rule was that you couldn't bet truly blind. You had to at least look at your cards. To see your cards and bet was a 'game of skill', to not was 'gambling'.

Did you play at Cafe 40/42? =)

At times... I was out at RedWest.
Post reply on HN