bustabit.com -- The Social Gambling Game

[statdude]

Question on the bot / scripting capabilities ----

Is it possible to make a bot based on how other players are behaving in PAST games... but even better, would it be possible to TARGET players based on their recent play and then go out right after them in the CURRENT game? (Bonus hunting)?


If so, I'd like to hire someone to build a bot. Any well known bot scripters around?

[DarkDays]

Question on the bot / scripting capabilities ----

Is it possible to make a bot based on how other players are behaving in PAST games... but even better, would it be possible to TARGET players based on their recent play and then go out right after them in the CURRENT game? (Bonus hunting)?


If so, I'd like to hire someone to build a bot. Any well known bot scripters around?

I've seen a couple of bots that did that but they are super obvious and very easy to outplay.  If you are only playing against static players (players that play BaB as if it's a dice game), you'd make money but if you are playing against thinking players you wouldn't.  It probably wouldn't be worth the money you'd spend on it.  It's notable that I haven't seen these players in awhile, so I'd assume they went busto.

[statdude]

Hmm, excellent feedback. thank you.

Would indeed make it hard to be smart enough.

[DarkDays]

Hmm, excellent feedback. thank you.

Would indeed make it hard to be smart enough.

The main problem is that it would need to be dynamic.  You need to cash out as soon as you see the top X players cash out, or asap if the top X players left typically cash out >2x, and all of those factors (except the last one) are changing every single game.

[jaysabi]

Since everyone loves charts as much as me, here's bustabit's historical wagered/profit chart:

I do love me a good chart. Couple questions though: Why does the profit start at a number above zero, but the amount wagered starts at zero? Does this chart reflect the beginning of bustabit operations, or just since you took over from the previous owner?

[dooglus]

The chart is a bit confusing, but the red line is profit and it's axis is on the right hand side.

I'm more interested in the x-axis. What did you use to create the chart, and why did it decide to space the x-axis labels out 2 months apart except for one pair where it had them overlap each other. That's really weird.

It looks like a gnuplot chart, but I've never seen gnuplot create such a weird looking grid.

[RHavar]

I'm more interested in the x-axis. What did you use to create the chart, and why did it decide to space the x-axis labels out 2 months apart except for one pair where it had them overlap each other. That's really weird.

It looks like a gnuplot chart, but I've never seen gnuplot create such a weird looking grid.

Yeah, I was wondering the same thing. I've been meaning to look into it, but a bit lazy. Here's the gnuplot file, and the raw data if you're interested:

https://gist.github.com/RHavar/31296dd5f396bb1935a098db507509c6


(It's possible there's a problem with the data, it was provided by shiba rather than me querying the master database directly. But shiba tends to do an extremely good job of keeping in sync, as it even has logic to fetch games she missed)

[yueno]

bustabit.com
 


Bustabit a social bitcoin gambling game in which you place a bet and cash out before the the bust event. All games are generated from provably fair distribution, and includes both a skill and luck element. In it's almost year of operation, it has become one of the most popular bitcoin gambling games, reliably handling  thousands of bitcoins in withdrawals, deposits and winnings

Bustabit is good for beginners because its too easy to get money there also its too easy to lost a money there. So you need strategy to not lost in a game and continue at the winning crowd.

[janggernaut]

...

Bustabit is good for beginners because its too easy to get money there also its too easy to lost a money there. So you need strategy to not lost in a game and continue at the winning crowd.

Not sure why you replied to the first post in this thread.
There is no "not lost strategy"  and "continue the winning" like you said, This is a gambling site with 99% luck and 1% skill (I bet you never tried bustabit before)

[freemanjackal]

i have seen here huge profits,wow, i wasnt even aware of this site, of course For that great profits they should have started with big amount of btc too, rigth? ill take a look to it  soon to see how do i do it, if luck is walking by my side.

[RHavar]

(It's possible there's a problem with the data, it was provided by shiba rather than me querying the master database directly. But shiba tends to do an extremely good job of keeping in sync, as it even has logic to fetch games she missed)

Now that I think about it, that x-axis glitch it's probably caused by a few days of downtime shiba had. Then when shiba did go back online, and download those days the timestamp would've been wrong, so those days statistics are being grouped in the wrong day.

It should be pretty easy to confirm by checking the raw data for missing days, but I can't think of a 1-line bash command for that. Anyway, it's not a big deal as the end graph results should all be correct, just a few days lumped together.

[dooglus]

Here's the gnuplot file, and the raw data if you're interested:

https://gist.github.com/RHavar/31296dd5f396bb1935a098db507509c6

That works well for me with "gnuplot 4.6 patchlevel 6":



Edit: how about trying this instead of using 60 days per 2 months:

set xtics 365.25/6 * 24*60*60 # an average 2 months in seconds

[RHavar]

That works well for me with "gnuplot 4.6 patchlevel 6":

Mine was generated with:

Code:

$ gnuplot --version
gnuplot 5.0 patchlevel 5

[libertynow]

If anybody is looking to gamble vicariously through my bonus bot, check this out:
https://nxtforum.org/index.php?topic=13275.msg229488#msg229488

Bot is EDGE_NXTAE.  https://www.bustabit.com/user/EDGE_NXTAE

All questions welcome.

Thanks for your interest,
LibertyNow

[dooglus]

Mine was generated with:

Code:

$ gnuplot --version
gnuplot 5.0 patchlevel 5

What if you change the xtics to this?

set xtics 365.25/6 * 24*60*60 # an average 2 months in seconds

[Lilmon]

I was looking for something like this. Thank you!

[fiscorcle]

If anybody is looking to gamble vicariously through my bonus bot, check this out:
https://nxtforum.org/index.php?topic=13275.msg229488#msg229488

Bot is EDGE_NXTAE.  https://www.bustabit.com/user/EDGE_NXTAE

All questions welcome.

Thanks for your interest,
LibertyNow

https://www.bustabit.com/user/EDGE_NXTAE

Not looking too good right now, is it?

[libertynow]

If anybody is looking to gamble vicariously through my bonus bot, check this out:
https://nxtforum.org/index.php?topic=13275.msg229488#msg229488

Bot is EDGE_NXTAE.  https://www.bustabit.com/user/EDGE_NXTAE

All questions welcome.

Thanks for your interest,
LibertyNow

https://www.bustabit.com/user/EDGE_NXTAE

Not looking too good right now, is it?

Actually, I'm very pleased.  You can see the bet amount right now is 6,132.  That means EV is positive and a whopping .32%/bet.  The variance swings wildly, but if EV is positive the peaks and troughs will move up. Yesterday, it was over 300k.  Just watch and you'll see. 

I also take BTC investments directly.  Just msg me with your NXT account and how much your sending. I'll reply with the BTC address to send to and what your NXT/BTC price lock is.

Thanks,
LibertyNow

[dooglus]

What if you change the xtics to this?

set xtics 365.25/6 * 24*60*60 # an average 2 months in seconds

  Yeah, that fixes it. Why is that? I don't get it

I've not looked into what's going on in detail, but if you're asking for xtics every 60 days, and it is rounding down to the start of the month then eventually you're going to get two consecutive months showing up: March 1st + 60 days = April 30th, which rounds down to April 1st. The problem is that 60 days per 2 months means only 360 days per year, so you're off by 5 or 6 days each year. Eventually that drift will show up as a 1 month gap in the x-axis labels. Working off 365.25 days per year instead of 360 days per year results in a smaller error, and less likelyhood of a visible error.

Edit: I've not been able to reproduce the observed results exactly, but this is pretty close:

$ echo $(for i in {0..16}; do echo "select date_add(date('2014-07-13'), interval 60*$i day)" | mysql -N; done | cut -d- -f1,2)
2014-07 2014-09 2014-11 2015-01 2015-03 2015-05 2015-07 2015-09 2015-11 2016-01 2016-03 2016-05 2016-07 2016-08 2016-10 2016-12 2017-02

[dooglus]

In hindsight, it's pretty obvious. But damn, once again I'm impressed.

What's weird is I've always used this:

    set xtics "1371427200", 86400*28

for my monthly axis labels and never seen anything other than 1 month gaps. Maybe because I'm using single months, and having two dates 'round down' to the start of the same month wouldn't be a problem.