I never got a chance to respond to his allegation when he posted it in the now-locked rackback thread, here:
  https://bitcointalk.org/index.php?topic=683727.msg14433154#msg14433154

He wrote:


He's talking about a time near the start of Just-Dice where a big investor/player requested a withdrawal of 1300 BTC. That was more than was in the hot wallet so I processed the withdrawal manually from the cold wallet, but forgot to debit his account, so there was still 1300 BTC in his balance that didn't correspond to actually existing Bitcoins. Over the following hours while I was sleeping those 1300 BTC were lost to the site. At first I was going to take the loss myself, and pay the 1300 BTC out of my own pocket to make the site whole again. But a group of investors convinced me to roll back the bets that were made with the 1300 BTC that didn't correspond to real coins, so that's what I did.

The net result for investors was that the 1300 'fake' BTC was split between them, and then it was taken back away from them again. No investor lost anything as a result, and a few were sneaky enough to withdraw their windfall before I rolled back the fake bets. They got to keep their bonus coins, which I paid for out of my own pocket.

A mostly accurate story about this whole situation hit coindesk not long after it happened:
  http://www.coindesk.com/bitcoin-gambler-cheats-satoshidice-competitor-just-dice-out-of-1300-btc/

I've never had any trouble from Lutpin, Twitchy or game-protect. I advertise Just-Dice in my signature, and run the campaign myself.

I would be surprised if Twitchy was the same person as game-protect. The English on the game-protect site is pretty poor. It reads like someone who either doesn't have English as their first language:

"a player act fraudulent"
"our researches revealed"
"After you subscribed to Game Protect, please submit ..."
"Only issues arose after your payment are eligible"
"Casino account protect service"

I've read a bunch of Twitchy's posts (because he and I appear to be similarly frustrated with betcoin.ag's shady practices, so I often see him posting in the various betcoin.ag threads) and never noticed grammar errors like these.

You can save the bitaddress page source to a thumb drive and open it on an offline machine.

If you want me to see your message in a timely manner, try PMing me rather than posting in a thread I am unaware of.

I never added Lutpin to DT and don't even know who he is. It looks to me like he's in cryptodevil's trust list.

I think he means that game-protect.com is pretending to be a useful service, but actually is just a way of making money for you. I'm not familiar with your site or the services it offers, and after a quick glance at it I'm still quite unclear. I don't see how you are blackmailing anyone though, so that doesn't make much sense to me.

You might find people on the IRC channel (#clams on freenode) who can tell you about the roadmap. I'd be interested to know myself what plans are for the future.

That record-breaking output has only staked once more since then:


There's a new fastest-staking output. It staked 11 times in a little under 2 months:



And here's an updated plot of the counts of outputs of each size over time:



Edit:

Sat Sep 3 2016, an output reached 36 CLAMs:

I thought all this discussion was in response to someone wanting to know what they can expect to earn by investing in MP:

Just-Dice.com publishes daily profits. Sometimes they're positive and sometimes they're negative.

Anyone investing in a dice site ought to understand that there is a lot of variance. If showing them the facts stops them investing then good - we don't want investors who don't understand what they are investing in.

See https://just-dice.com/misc/wagered.txt for a day-by-day list of the site profits, and https://just-dice.com/images/all_time_profit.png for a chart of the site profit over time. It goes up and down, much as expected.

https://coinmarketcap.com/currencies/clams/#markets lists the size of the various CLAM exchange markets:



Almost all the volume is currently at Poloniex. They don't need any ID stuff unless you want to withdraw more than $2000 worth of coins per day - the total CLAM trading volume on Bittrex over the last 24 hours is only around half that...

I took a look at Bittrex when they first added CLAM but didn't like the site much. I can't remember exactly what my problem with it was - maybe it was just a lack of liquidity.

That's a common fallacy. Whether every player keeps playing until they bust, or stops as soon as they are ahead, the law of large numbers still applies, and the house profit still tends towards the expected house profit in the long run.

No, it's just wrong. He's saying you can divide the returns the bankroll should earn over a year by the bankroll amount to get an estimate of your expected earnings on an annual percentage basis.

Suppose the bankroll is 2000 BTC and the expected earnings are 1000 BTC per year. Using QS's formula that gives 1000/2000 = 0.5 as his annual percentage expected earnings whereas in reality earning 1000 from 2000 is more like 50%. I think he forgot a factor of 100 somewhere.

In that case I should get it right.

Change:

  if the output size is going to be more than twice 'splitsize'

to

  if the output size is going to be at least twice 'splitsize'

If you have a 99 that stakes, you will get 25,25,25,25 (not 25,25,50, which would be the result of how I previously explained it).

In unrelated news, the JD staking wallet just made its first 34 CLAM output (starting from 100 -> 25/26 -> 27 -> 28 -> ...)

When staking, if the output size is going to be more than twice 'splitsize' then it will add an output of size splitsize, and repeat. So if you have an output of size 100 which stakes (making 101) and splitsize=25, you will get 4 outputs (25, 25, 25, and 26).

If the output size isn't more than twice 'splitsize', it will look for other small outputs to combine with the one that is staking, aiming to make the output as big as it can, but no bigger than 'combinelimit'. So if you have a bunch of 10 CLAM outputs and one of them stakes and combinelimit=50, it will combine 4 of the 10's together and you'll end up with an output of size 41. (Combining 5 would be too much, giving 51). Since you're staking, you don't pay a transaction fee, so it's a good way of keeping the 'dust' outputs in your wallet under control without having to pay a fee to do so.

I like 25/50 because the JD staking wallet has so many CLAMs in it. Making the outputs any smaller means there are just too many of them. The client has to run through all the outputs every 16 seconds to check whether they can stake or not. If you don't have a lot of CLAMs you could safely use smaller numbers. Maybe 10 and 20.

The reason we split the outputs up is because when they stake they have to wait 500 blocks before they can start trying to stake again. I figured I was OK with 1% of the wallet's value sitting idle waiting to mature at any time. That means I want each output to take approximately 50k blocks to stake, so that the 500 blocks it takes to mature is 1% of its life. If 1100 CLAMs stake once per day (and we have 1440 blocks per day) then to stake every 50k blocks the output would need to be 1100 * 1440 / 50k = 31.68 CLAMs. So having them range from 25 to 50 is about in the right range.

Edit: here's an updated chart of the number of outputs of each size in the JD staking wallet:



The vast majority of the outputs are still 25 and 26 CLAMs, so we should be seeing less than 1% of the wallet sitting idle.

Right now, the balance is 1,454,986 of which 13,801 is idle - a little under 1%.

119,756btc

https://www.reddit.com/r/Bitcoin/comments/4vupa6/p2shinfo_shows_movement_out_of_multisig_wallets/d61oe33

Well, kind of. Usually when people talk about "hardware wallets" they mean some kind of device which can't be hacked due to the code it runs being very limited in scope, and uneditable without physically pressing some buttons on the device. Android on the other hand is a general purpose operating system and probably has many vulnerabilities and possibly even deliberate backdoors.


If the 16 included the fee, yes, but probably it doesn't. In which case 5+5+1+5 isn't quite enough because you need to fund the tx fee as well.

Yes I have.

1) breadwallet has the same shortcoming - you can only have one receiving address at a time

2) the PIN is only 4 digits, not 6, so it's worse in this respect

3) breadwallet connects to random full nodes; I cannot tell it which nodes to connect to

In addition, breadwallet shows my balance in 'b', which it defines as a millionth of a bitcoin. The developers refuse to allow us to see amounts in BTC. I'll be sticking with Mycelium for now, despite its shortcomings.

Except it has already been going on for over a year, and new people fall for the scam every day.

Is it a web wallet? Do you stake the coins on deposit? Do you share the stake rewards with the depositors? Do you take a fee?

To make the outputs I did something like:

    $ clamd sendtoaddress <address-goes-here> '{"amount":100,"count":50}'

And to auto split on stake, I put this in my clam.conf:

    splitsize=25
    combinelimit=50