BitBet is quite the lulzmine

Tuesday, 23 April, Year 5 d.Tr. | Author: Mircea Popescu

Consider the "Bitcoin at $420 on or before 4/20" bet which recently closed.

When this bet started (early on April 10th) people were pretty excited about Bitcoin "exploding", "going to a new level" etc scl. Reality soon diverged from their expectations, as you'd expect. Early enthusiasm - odds actually were ever so slightlyi in favour of Bitcoin breaching 420 - soon melted away, and by the end the odds were 10:1 against.

Let's now review the comments users left during all this excitement.

Yah mon. Bitcoin gonna be blazin on 420

BTC is smokin hot right now. Value will be pretty high on 4/20.

Shit shit, I'm going to party so hard on 4/20 after winning this bet.


Totally. Moving on.

I wonder if the 26BTC better realizes he'll lose money on the win.

The only 26 bet is of course

15-04-13 04:46 No 42`222 26.00000000 1Bm8f 27.28887849 182y8

Turns out he didn't lose money on the win. Moving on.

Now everyone below 7214 is gonna lose money. If the weighted 'no' reaches 9'978'000, and no one throws money away on a 'yes' bet, every better will come-out behind. I should go buy BITBET stock.

Stop betting, you won't make your btc back!

The buying BitBet stock is a great idea, of course, but otherwise this is all a little bizzare. Out of a total of 81 bets, exactly 4 made less than what they bet. To wit : a 0.23 BTC bet made 0.22982524, a 0.15 BTC bet made 0.14972581 BTC, a 0.0002 BTC bet made 0.00019919 and finally a 1 BTC bet made 0.99499186 BTC. This comes to a total of 1.3747421 BTC returned on a total of 1.3802 BTC bet, so the loss is 0.0054579 BTC (0.39%). The 1.3802 bets are just under 1% of the total bets made.

Its the dumbasses of the world that make the rest of us wealthy...

Rational Better
Actually, it's the dumbasses that make us all poor. As the payback cutoff window is narrowed by people making sure-loss bets, the number of actual winners is reduced. Some crazy bastard lost two whole BTCs just to knock three other people out the payback window.

This is even more bizzare. How would people making sure-loss bets reduce the number of actual winners ? The winners don't get their winnings exactly out of thin air, do they ?!


Well, here's the lesson: Don't bother betting if the wager accepts new entrants until the last minute. Only bets which close before certainty has been reached can prevent fools from crowding-out winners. FYI, fully 70% of the "yes" betters will lose money on the win.

You didn't think it can get any more bizzare, did you ? Well... of course it can. This is after all Bitcoin.

For one, Yes didn't win. For the other, had Yes won the odds were so greatly against it that the winnings would have been in the 10x range. What sort of nonsense is this 70% ?! Anyone's guess. Apparently everyone on the Internet has a percentage or two at the ready for any circumstance.

Yes, and this is a stupid method of calculating things. The winning side should never lose money period, not sure why bitbet implemented such stupid calculation. Just take a 2% fee from the losing side, instead of taking 1% fee from both side. How dumb is that.

It's fair and therefore stupid. Internet logic.

The simple truth is, the only way it can work is if all the winners are repaid, in total, before the winnings are split. Then the remaining amount (Pool less winner's bets)should be apportioned according to the weighting scheme. If a winner can lose money, if discourages betting. ARE YOU LISTENING BITBET?!? Like this: The A pool bets one way, the B pool bets the other. Suppose A wins. Each better in A receives 100% of their bet back. The remaining amount (which is equal to B's pool) then has 1% subtracted as fees for BITBET. The remaining amount (Total less A less 1% of B) is then distributed to the A pool betters. Then there's a tremendous incentive to place and create bets, which equals greater fees for BITBET.

Yes, we're listening. This shit's better than Comedy Central.

This is the best way the bets can work. If you were guaranteed to get back 100% of your winning bet no matter your bet weight, then people would come in and dump 1000 BTC near closing. Though their bet would be weighted poorly, the mass of the bet would make up for it. This way people are discouraged from making a stupid bet twice. After one stupid bet like this, they will learn.

This is the idea, pretty much.

obviously bitbet takes 1% from all returning money because its works out better for them taking 2% from the losing bets wouldnt be close to this in most cases

I'm not sure this judgement is correct. Short odds win most of the time but long odds probably win often enough to bring the figure somewhere in range. Just a guess on my part, I've not run the numbers (mostly because I don't think we have enough bets to get usable results yet).

The reason for the fee structure however is not some arcane income optimisation. The reason for the fee structure is much more stolid and unclever than that : BitBet provides a service, for which service it charges a fee. Both sides, the winners as well as the losers have taken advantage of the service BitBet provides. Consequently both sides are billed for said service. No restaurant charges only the patrons that got food poisoning from eating there, for instance, giving everyone else a free meal.

Sure, people are used to marketing gimmickry to the point that they expect it by now and act strangely when offered a square deal, pining instead for "loser only" fees, line betting and all the rest of the scammy nonsense robbing them blind irl. This is quite irrelevant as far as we're concerned : they'll just have to get re-used to normalcy and come to sanity, much like Plato's cave dwellers will have to get used to the light of day.

With a 1% fee on bets, BITBET will make 1.41 BTC on this bet. If BITBET took 15% of the losing side, they'd make just over 1.5 BTC off this bet, and every winner would get, at least, 100% payback. In fact, 15% of the losing side (as a fee) is a better arrangement for everyone involved. The winners will, like I am, gripe about losing money. The losers don't expect anything back. If I were the BITBET owner, I'd take a fee from the losing side, such that the fee always exceeds 1% of the total pot size. More money = better.

The proportional payout is a great idea, but allowing late-comers to reduce early-entrants winnings to less than the 1% fee discourages early better too. Everyone below 78'' will bet less than 100% of their bet back. That's silly. If all fees were taken from the losing pool before the winnings were proportionately distributed, every winner would get >100% back, and BITBET could make more than it will now in fees.


  1. 1.09 BTC on YES to 0.96 BTC on NO. []
Category: Meta psihoza
Comments feed : RSS 2.0. Leave your own comment below, or send a trackback.

3 Responses

  1. Aside from lulz, let's explore some (although extreme) scenario where later bets cause winners to get less than their bet.
    Let's start with 1:1 btc bets at weights 90000. Then someone bets 200BTC with weight 85000 and ends up winning. Payouts would be:

    total winning weight: 17590000
    losing weight: 90000

    0.99 + 0.99*1*90000/17590000 = 0.995065 BTC (0.4935% loss)

    198+198*1*85000/17590000 = 198.95679 BTC (0.5216% loss)

    0.5% potential loss for such unlikely case is not really something to scream bloody murder about, imho.

  2. Oh and there's an error in the faq, it says: 99% of his original bet sum (1% goes to plus 99% of his bet times the total bet by the losing side
    multiplied by
    *the winning side total weight*
    and divided by
    *his bet's total weight.*

    and it should be: 99% of his original bet sum (1% goes to plus 99% of his bet times the total bet by the losing side
    multiplied by
    *his bet's total weight*
    and divided by
    *winning side total weight.*

  3. Mircea Popescu`s avatar
    Mircea Popescu 
    Wednesday, 24 April 2013

    @jurov Your example (1:1 at 90`000, 201:1 at 85`000) proposes that no further bets are made. While it is indeed possible that in that scenario described the later bettor spends ~1 BTC of his own money to produce a ~0.005 loss to the other people on his side, nevertheless I would say it is more likely that his large bet attracts further betting on the opposite side (from people who judge the odds to be better than 201:1) and thus on the balance of probabilities I'd say he's likely to create a profit, rather than a loss for his own side, this because out of every 1 BTC bet against the winning side, the 200 BTC guy and the 1 BTC guys split the proceeds not exactly by capital contributed, but slightly in favour of the earlier 1 BTC guys (because of weight).

    To give numeric clothing to these theoretical considerations :

    1. #1 Bet 1 BTC on loser side at 90k weight. Total loser weight = 90k.
    2. #2 Bet 1 BTC on winner side at 90k weight. Total winner weight = 90k.
    3. #3 Bet 200 BTC on winner side at 85k weight. Total winner weight = 17.09mn.

    At this point the #2 are entitled to 90/17090 = 0.5266238% of the loser side in spite of contributing 1/201 = 0.4975124% of the capital on the winning side. The difference of 0.0291114% means that should someone make a further bet on the loser side the #2 people will get slightly more than what their risked capital entitles them.

    Since BitBet doesn't allow extremely lopsided bets ("The moon will come out tonight" etc) it is unlikely that a 200 BTC bet on a previously 1:1 proposition will attract absolutely no further betting from anyone, and so I'd say on the balance of probabilities the earlier betters benefit from later heavy betting. I would go as far as to say that seed betting (ie, placing some bets early on on many propositions) is a legitimate business model if properly administered, liable to generate reasonable (for a BTC investment) monthly profits and ROI.

Add your cents! »
    If this is your first comment, it will wait to be approved. This usually takes a few hours. Subsequent comments are not delayed.