Topic: A quick calculation on Bitstamp's loss due to the hack in relation to their fees (Read 1490 times)

Stop hunting. Possible, but not on Bitstamp: doesn't have margin accounts >_<

here is another way in which exchanges (or exchange operators) might be generating revenue:

http://www.reddit.com/r/BitcoinMarkets/comments/2s0hpd/z/cnm249o

i suppose they also generate significant income from arbitrage, but i would not be able to estimate how much, despite availability of market volume & depth data.

anyway, clearly stamp and pantera prefer to take the hit than to lose the brand / clients, so a loss of USD$5M must be manageable.

Good idea for a startup

I'm even a bit more pessimistic (and cynical) here: Assuming for a moment the owners and outside investors don't give two fucks about Bitcoin (not saying they do, just asking "what if they would"), and assuming they'd try to calculate if they should let the boat sink and hand the loss over to the customers or not, I was trying to determine if it was in their best interest to not damage the Bitstamp brand to the point where it becomes unsalvageable. Upwards of 10 million USD as a yearly revenue they can reasonably expect in the future if the brand stays strong makes me think it could be in their interest to take the loss, but the real answer of course depends on what their net profit is, which I have no way of deducing.

EDIT: Guess the tl;dr is, you're looking at it in terms of them being able to take the loss in absolute terms (being early adopters and all), I'm looking at it from the perspective of opportunity cost.

Number of coins returned to their owners by regular police to date: zero
Number of coins taken from their owners by regular police to date: over 175k

So I'd say, very slim chances here Or we need a special bitcoin police, bounty-based, which would get percentage of returns.

Good summary.
Bitstamp started in 2011 when coins were cheap. So they may well have been early investors and held more than 19k on their own account. So this hit is painful but not fatal to the business.

The hacker might get identified in the future, and regular policing chases him down. So the 19k is not a guaranteed loss, yet.

That's my impression as well. I tried to get some highest and lowest estimates for their revenue, just to get an idea how much the loss of the hack is in relation to that revenue estimate.

Possible objection is of course: if their running cost are high enough to eat up most of the fee revenue, the costs of the hack are indeed insurmountable.

If, on the other hand, their expenses are comparably low (does anyone know how many people work for Bitstamp?), then it is likely in the best interest of Bitstamp owners (and investors) to swallow the loss to keep the brand alive, betting that further income will be worth taking that loss.

Great estimates OP. So Bitstamp will be running healthy again in 12 to 18 months and start making profits again. That is good to hear. A confidence boost for a rather fragile market at the moment.

BitcoinCharts keeps a record of the 365d volume, both in USD and in BTC: http://bitcoincharts.com/markets/bitstampUSD_trades.html

Bitstamps current 365d volume is ~$2.4B, which should be quite close to their 2014 result, since 2015 has barely started (though volume measured in USD was higher in the first 10 days of 2014 than 2015). Bitstamp takes all fees in USD (checked on transaction history page, both buys and sells have USD deducted for fees). So the minimum amount earned in fees last year is 0.4% (if all trades were done in the lowest fee tier) of that $2.4B, or $9.6M. The maximum is 1% (if all trades were done in the highest fee tier) of $2.4B or $24M. In reality, I expect that the average fee is somewhere between 0.3% and 0.4% for each side of the trade (so 0.6%-0.8% income for Bitstamp).

Of course, one has to deduct operating costs from this income. But while taking a $5M hit is by no means a small loss to them, according to these numbers it should not be insurmountable.

Strictly speaking yeah, guess I read too much into "...in the worst case they lost about 20 month's worth of fees in the hack, and in the best case, 3.5 months worth of fees."
I assumed he was implying that the "hack" could be covered by 3.5 to 20 months of operation (which is not the case if operating expenses & other expenditures occured).

oda never said he was talking about profits, only revenue (all of which may be assumed to come from fees).

d) they had no operating expenses.

My bad, I thought fees were actually 0.4%-1% on either side. ( I never used Bitstamp so..)

Care to explain why quadruple?

An arbitrary executed order is of the form 'sell x btc at price y' on one side, and of the form 'buy x btc at price y' on the other. It will incur between 0.2% and 0.5% per side, so between 0.4% and 1% per trade as batou correctly pointed out. Fees are identical for buying and selling side. So that's twice the amount I calculated in my OP (my excuse: it was late, I was tired, and I was young and needed the money :/). What am I missing?

It's quadruple of that - total, double of that in BTC

That and they receive bitcoins and dollars half-half, so the they have to buy them with dollars scenario may be possible but not likely.

Completely forgot about that, you're right. So, double everything up, or cut the 'time to recover' estimates by half.

This, indeed. You always have a selling and a buying party.

Nonetheless nice thread. Thanks for providing the calculations.

They take fees from both parties, 0.4% is minimum, 1% max.

The basic idea: summing up over the past 365 days of trading volume multiplied by (H+L)/2 we should get a decent approximation of the yearly trading volume on Bitstamp in USD.

That value peaked in November this year at around 3.5 billion USD. The lowest value during the past 365 days was at the beginning of 2014, around 1.5 billion USD (lower price, and smaller market share of BS during 2013).

We happen to know Bitstamp's fee structure (between 0.2% and 0.5%) but we don't know exactly which share of the volume commands which fee. Assuming that a lot of smaller trades make up a bulk of the volume, we'd get closer to 0.5%, assuming (more likely, imo) that large trades and volume coming from payment processors linked to BS make up the bulk, it'll be closer to 0.2% (in fact, the really big players might even get better fees).

That said, let's estimate the range of fees Bitstamp could expect to rake in within a year in the past:

max: 3.5 billion at 0.5% ~= 17.5 million USD in fees
min: 1.5 billion at 0.2% ~= 3 million USD in fees

So by my approximation, in the worst case they lost about 20 month's worth of fees in the hack, and in the best case, 3.5 months worth of fees.

Note that the above is resting on a few more important assumptions:
a) 19k coins is indeed what they lost, not more, not less
b) they "cash out" fees, i.e. my calculation of using USD volume corresponds to how they run their business i.e. they don't store profits in coins (only)
c) they weren't insured against a hack (no idea if that's possible anyway).