I don't think that's true.  By juggling the percentages you can withdraw every week and arrange to withdraw the same amount as if you withdrew only once every 4 weeks (but end up with a lower principal) or end up with the same principal (but withdraw a smaller total).

For example.  Suppose you deposit 1000 BTC and decide to withdraw 70% of your interest after letting it compound at 7% per week for 4 weeks.

You'll withdraw 217.55 BTC and leave 1093.23 BTC in the account.

Now we attempt to replicate that by withdrawing a fixed percentage of the interest each week.

If we withdraw 67.81% of the interest each week, we'll end up leaving 1093.23 BTC in the account but only withdrawing  47.46 + 48.53 + 49.63 + 50.74 = 196.38 BTC - we're 21.16 BTC worse off than leaving the interest to compound for 4 weeks.

On the other hand, if we withdraw 75.75% of the interest each week, we end up withdrawing 53.02 + 53.92 + 54.84 + 55.77 = 217.55 BTC, but end up with only 1069.65 BTC left in the account - we're 23.58 BTC worse off than in the first case where we left the interest to compound.

In summary:

   weekly withdraw         total withdrawn         total left in account
    ---------------         ---------------         ---------------------
    0% (70% at end)              217.55                   1093.23
    67.81%                       196.38                   1093.23
    70.00%                       202.26                   1086.68
    75.75%                       217.55                   1069.65


So we have to chose to either withdraw the same amount (and keep less) or keep the same amount (and withdraw less).

If you think about it, it stands to reason.  Withdrawing any amount each week is bound to earn you less than leaving the money to compound for a month before withdrawing.  That's why compound interest is 'better' than simple interest.

Edit: added 70.00% weekly withdraw line to table - thanks payb.tc.

It took me a while to work out what WPW meant.

It's this thread: https://bitcointalk.org/?topic=81542 ("Who Pays What?")

I tried talking her out of it, telling her it wasn't worth the risk.  She doesn't have a silver star account, so can't bet more than $1 per number on roulette otherwise I would have worked out a stupidly risky martingale strategy for her. 

She insisted however, put $1 on number 17 and was determined to spin until it won.  Luckily for her it came up on the 5th spin, putting her in first place and over $25 on the leaderboard, but without having bet enough to qualify.

She switched to blackjack to rack up enough bets in the 20 minutes that were remaining, and made it with minutes to spare playing basic strategy, betting a flat $2 per hand.

A "how much have I bet this weekend" indicator would be useful - she had little idea how much she had played already so I recommended she just play $100 worth of blackjack to be on the safe side.

Oh, and when she got the email confirming she had won she hit the roof:


It might be worth changing the wording in that email:


It can apparently be quite upsetting...

He's 'illpoet' on seals.

Yeah, I noticed later that I had the feature turned off anyway, and the pre-fetching I was seeing was something facebook explicitly does using javascript.

What was the trigger that was causing the double-submit on some browsers but not on others?

FYI
http://www.bitcoinfog.com/ says:
"The service takes 1%-3% (randomized for obscurity) fee on each deposit."

It is returned to you.

Here's an example of a 140 BTC bet that was refunded.

Usually the refunded amount is the same as the bet, but sometimes it's slightly more or slightly less.  I don't know why the refund isn't always the same as the bet amount.

Edit: according to my calculations, 1789 bets have been refunded.  934 of them refunded the right amount.  308 refunded slightly more than the bet amount, and 547 refunded slightly less.  The biggest overpayment was 0.00018160 BTC and the biggest underpayment was 0.00014976 BTC, so really insignificant amounts.

SatoshiDice doesn't work with that address either:

  http://blockexplorer.com/t/8eYVPqNsyA
  https://bitcointalk.org/index.php?topic=77870.msg1016122#msg1016122

Except that to qualify you have to have bet over $100 in total, and so the freerolls don't help you pass that hurdle?

Because you can put more than one bet in a single transaction, but all of them get the same 'lucky number'.  So you can bet:

* 25 BTC on lessthan 32000,
* 25 BTC on lessthan 32768,
* 25 BTC on lessthan 24000, and
* 25 BTC on lessthan 16000

in a single transaction, get lucky number 63500, and lose the whole lot at once.

I guess technically it's 4 bets, but since you only get one lucky number per transaction I was counting it as one.

Also a spaceo: "With sufficient analysis it maybe possible" (may be)

and a puncto: "0.0002 BTC miners fee" (miners')

There's a typo on http://blockchain.info/wallet/send-anonymously:

"taint analysis calculationts"

This graph is broken:

http://blockchain.info/charts/balance?address=1VayNert3x1KzbpzMGt2qdqrAThiRovi8

SatoshiDice has been a lot better in recent days at not missing any incoming transactions - that's why I've stopped posting lists of unprocessed transaction IDs.

However...

http://blockchain.info/tx-index/11502701/c9cca3a9e5a1494cee2c87c950c1283f83d091920cb6e6749b3e5930d8d69ac2 hasn't yet been processed, and is 12 hours old.  It's the only one that has been missed for days now.  I wonder what's special about it (other than that it's a bet from DeepBit's address).  Maybe the fact that the 'VayNert' address has over 353165 transactions on it triggered a bug in SatoshiDice?

Any idea what might cause trading like this?



Alternating buys and sells, steadily decreasing volume per trade.

I don't think I've seen anything quite like that before.

So long as not everyone switches at exactly the same time, the sum of the two schemes will have enough to allow people to temporarily withdraw from one and deposit into the other.

Perhaps there could be a bonus if you withdraw directly from the old fund into the new one - that way no real funds are needed at all.

But as pirate said:

I don't think he needs the funds to shut down.  All he needs to do is say that the interest in the current scheme is dropping to 1% per week, but that a new scheme is opening which pays 4%/7%, and let people switch to the new scheme if they want to.  Almost everyone will switch, then he can afford to force-withdraw the rest.

Then the bet is won, and the scheme carries on, just with a new name.

Or am I missing something?

I just noticed that my browser is 'pre-loading' the page it thinks I am going to open next.

Apparently chromium has a setting "Predict network actions to improve page load performance".

I suspect that this feature is what was causing the double-submit.

http://support.google.com/chrome/bin/answer.py?hl=en&answer=1385029