Bitcoin Forum

EDIT: you might want to read one of my later replies first...

Whenever anyone wants to know if a gambling site is rigging the game, they run a sort of brute force test.  That's when they play a ton of hands to see if it's either letting them win for the first couple hands or making them lose more than the odds of the game suggest.  The Wizard of Ods is famous for doing that and found some semi-high profile gambling sites were rigging it so people won with their free initial buy in credits more than they should and then lowering the odds after that to below normal game levels.  He usually plays 1000 hands but I'm not that patient :P so I tested BTCFlip with 50 bets  ;D

The results were *drum roll*...

26 losses and 24 wins.  They pay a 1% bet commission on referrals and obviously keep a bit of a advantage to themselves but these numbers indicate that it's not like a 20% advantage or something.  That suggests a 2% advantage but don't think that number is set in stone because 50 is not a big enough test to dial it in precisely.  The results were:

loss
win
loss
loss
loss
loss
win
win
win
loss
win
win
loss
win
loss
loss
win
loss
win
loss
loss
loss
win
loss
win
win
loss
loss
win
loss
loss
win
win
loss
win
win
win
win
loss
loss
win
loss
loss
win
win
loss
loss
loss
win
win


so it seems that there are no conditions under which it gives an advantage or disadvantage.  They certainly weren't letting me win right at the start lol.  Just for fun, I ran a few bets that were much, much larger than their 0.01 minimum and lost 3 and won 2 which doesn't say much other there's not a 100% "they win" clause in their randomization code for high BTC amounts :P can't really comment on whether or not they up the odds at all in that case though.

By the way, I actually played 51 rounds and it dropped one of my bets.  It timed out and they said they never got it but I checked and the correct amount was sent to the correct address.  I was running all of these in firefox but I opened IE because of the view source feature to see if there were any hints or exposed code on the page and that's the one that got dropped so maybe it's not IE9 friendly or something.  It is all javascript based so who knows.

So my overall result is with no account or other apparent way of tracking a deposited bet, if anything happens to the page or their server mid-game, you're screwed, making it pretty unsafe.  Also, they don't list any information on the odds or have any guarantee that they don't change so that's pretty sketchy.  Why would anyone (other than me :P) play a game that doesn't list the odds of winning?  Other than that, from an odds point of view, it's at least somewhat close to 50-50 and appears to have no larger of an advantage than any casino game so if you want to go for it, go for it :P

EDIT: oh yeah, I should probably mention that I won 24 times and there are 24 payouts from them in my bitcoin client log so they also do pay you if you win :P  That seems to be an important part of any gambling site review lol.

26 losses and 24 wins  :)

Where on Shackleford's site does it say he ran these kinds of tests?  I don't think 1000 trials is going to provide any evidence about -- well, anything -- remotely useful for a coin flip.  That's just too little data.

a couple hundred sets of 50 flips should begin to prove this better :P

Jeeez 50 times... I had that in my sig for a while and even I never played it that much.

Got a few '1% referral' payments though even though they work out to be quite small mostly.

they were most likely mine :P

okay, I got a few reports that they do in fact possibly alter the odds for much larger bets.  So I ran a very expensive experiment and found that yeah, they use horribly unfair odds at higher amounts.  After a high enough volume to accurately measure it, it's around 80-90% they win at amounts above 0.5 BTC.  So don't bother with these assholes.

That's a pretty big accusation for providing zero evidence.

Can anyone back this up?

God damn, all these scam sites are bumming me out.  :/

what evidence do you want? the OP basically said that larger bets lose more often then win. 80-90% lose.

sites like the one being discussed just scream to me it is a scam. it is par for the course around here.
flip a coin with a bitcoin website and 50-50 that you will get screwed.

So, where are your detailed results for this like you used in your first post?

I was a little too unhappy to post the precise numbers after I got the calculation.  It turned bad real fast at low attempt numbers and I gave them the benefit of the doubt assuming it would even out.  NOPE.  And I already basically posted that they're fair.  That really pisses me off.  At all different amounts above 0.5 from the same receiving address, all but one lost in 12 attempts and then I didn't really feel like continuing cuz it got expensive.  If it was at 47% odds of winning, that's around a 2.1% chance of my results happening naturally so don't whine about the sample size.

These numbers aren't even in line with my past results of a sample size of like 5.  So originally it may have started fair, not so much anymore.  It possibly adjusts based on how much BTC they have on hand, which is a good way to not go upside down on it.  Obviously if you don't have the money to cover like a 100BTC bet or something someone could potentially make, you're going to program it to make that bet lose.  I don't think they had trouble covering a bet my size though.  But with no stated odds or any way to retrieve a reset session in your browser, it was a badly planned, hastily put together operation that doesn't even suggest an ounce of fairness.  It doesn't really suggest anything.  It's missing any info at all really!
Well hopefully their reputation is ruined enough that even if they do clean it up, nobody lets them benefit from their awful original operation.

Are you serious?  You clearly have no experience with statistics and gambling.  Learn what it is you are doing and how to audit code and what constitutes an acceptable test.  "ZOMG i got 11 tails and only 1 heads, im clearly being cheated" reeks of ignorance and inexperience with gambling.  The site may or may not be a scam, but your post doesn't sway the argument one way or the other.

I'm a professional poker player by the way.

FYP

You didn't mention how my math is wrong.  Okay, you run a longer test at high bets.  I'm sure an experiment by someone with the username luv2drnkbr would be really respected.  And don't let the fact that you really seem like the site's owner from your posts deter you.

Hi Thanks for the info i like to gamble for fun. So your info was great . Thanks Warren

I'm sure something like bitcoinstars.com would be fun.  Flip is more like...

http://i129.photobucket.com/albums/p227/wizzerd911/a92100e5-ee4e-4835-8612-e587752a8a55.gif


Watch what happens when this lady tries to use BTCFlip.

http://i129.photobucket.com/albums/p227/wizzerd911/e08e7b3896e85b7346713883f5799dde.gif

has anyone run a test on doubletrouble.bitcoinbet.com?

You can view the source of the page with Firefox too (Control-U is the keyboard shortcut, but it's somewhere in the menus too), am i missing somthing? Why did you need to use IE at all?

wow, one that actually places the odds on the page.  And a support number!  And....words....on the page!  Lots of them!  Now we've got something.  I might look into it but I'm kinda broke atm :P all non-sent to the exchange for holding money in my wallet is gone unfortunately after that awesome losing spree at flip.  BTCGuild has been having all kinds of random stupid issues (like being "unlucky" consistently for like 2 weeks and sending out incorrect idle notifications) but I finally have decent BTC coming in now from my backup pool.  Like 25% more per day!  Screw BTCGuild.  They're the best but they have some kind of scam or severe problem going on at the moment.

So I'll have to wait a bit to run a test but I'll post it in another thread.  You know, I'd have plenty of money if someone were to donate some to me for all my high risk research  ;D lol.  But since that's not gonna happen, wait a day or two and I'll have some results.  I'm shooting for another 50 round test.

if it's shown to be bad with large bets and good with small bets, why not just take it for what it is, and play with small amounts?

i won 0.24 btc so why not :D

The correct answer to that seems to be because in a sample size of 1 round (which is what i had the money for) I won on double trouble lol.  So 49% (not yet confirmed) vs mystery odds that appear to be unfair...go with double trouble.  vy the way, the commission is 1% for flip so obviously their odds are set above 1%.  But like I said, I haven't tested it completely yet.  They do things opposite too.  Your payment is confirmed first for a long time then you do the determination then instantly get paid.  Flip does it like it takes your word up front, generates a result (unfairly), then doesn't pay you until the initial bet was confirmed.  That seems exploitable on the surface btw but I'm not sure if you can make a forged payment real at will after the fact.

And btw, you don't need 100,000 attempts at a simplistic odds thing.  You'd need 100,000 samples for blackjack or video poker but not a two outcome system.  I think 100 is even enough to say hmm that was within a few of 50/50 or wow I just won or lost 90 times.  And since i'm poor, I'm doing 50.

Is btcflip.com still running normally?

I sent 2btc in order to play a game on btcflip.com at around 05:50 AM Feb/26th, 2012, GMT, but the game didn't start at all.

The address I sent 2btc is 1QKXYztB2zmVixTTvd89E2Bgh2LM7Uo3cD

Then I wrote a mail to admin@btcflip.com for help, no one reply the mail.

At 13:06 PM Feb/26th, 2012 GMT, I wrote another mail to admin@btcflip.com in case they didn't receive my previous mail, and still nothing replied.

Almost 12 hours has passed, now I wonder if this website is still runing normally? Is there anyone else  ever encountered a similar problem as I did?

So, according to the forum rules, I'm now stuck with having this in my signature, unless I want to have no banner at all.

Has anyone tested mysitethatmayormaynotreturnyourbitcoins.com??

jeesus people...

I just tested it. It appears down from here.

 :'(

You know what? With 100 fair coin flips you have 5.7% chance of getting a 10% wrong result (<40% or >60%), so NO. 100 tests don't tell you anything about being "within a few percent".

Please use statistics properly if you are trying to estimate fairness.

The chance of getting 11 out of 12 heads with a fair coin is much worse though: 3 out of a thousand. Still far from impossible though.

Exactly, so if they were tipping it more than a variance of 5.7% in their direction, there's a 90% chance I would have noticed it.  Tada.  So we can safely say it's not a 75/25 for example lol.

+1