Bitcoin Forum

Edit 09-05-2014: Bitstamp is increasing their trading limit to $5 on May 15th, after which the exploit as described in this post will no longer be possible.

After a response from Bitstamp, this post has been slightly edited. Bitstamp replied to my inquiry and I will discuss Bitstamp's response at the end of this post.

-----------------------------------------------------------------------------------------
I want to bring this all to your attention, because everyone trading at Bitstamp is affected negatively by a faulty implementation of the Transaction Fee Calculation on Bitstamp's side, which by definition always favors Bitstamp. At the moment this bug appears to be 'by design', but can be easily fixed.
-----------------------------------------------------------------------------------------

Bitstamp rounds their transaction fee off to the nearest cent decimal upwards. If you have a transaction of $11, the fee (0.002*13) is $0.022, which gets rounded off as $0.03. This is mentioned clearly in ther Terms of Service and I think that's okay.

The bug/faulty implementation:
The essence of this bug lies with the minimum order limit in combination with their fee rounding policy. Bitstamp's minimum order limit is $1. Because it is possible to place an order at 20% above market price, the effective minimum order limit is $0.86. A $0.86 order at 0.2% fee should be charged 0.002*0.86= $0.00172, which, due to their rounding, gets rounded off to $0.01. This is almost 6 times the actual fee. The problem is that this stacks up, leading to an actual exploit that always favors Bitstamp:

The actual exploit:
You place a sell wall at a certain price, agreeing upon 0.2% fee over that entire sell wall. A system/bot repeatedly makes many of the tiny buys into your BTC sell wall. Each of that buys is charged with 6 times the accepted fee. When this continues untill your sell order is completed, you will have effectively paid 6 times the fee over the entire sell order, instead of the 0.2% you thought.



Hypothetical Example:

When I put in a sell order of 20 bitcoins at $432.01, and some bot for some reason makes a LOT of mini tiny buys into it, you get an enormous fee percentage as a result. It is important to note that I as a seller have no control over the way somebody buys into my order! I am powerless then, and moreover I did not consent to having my order bought in increments that are so small that it disadvantages me while it sends 6 times the fee to Bitstamp. This system makes that fee 1.16% instead of 0.2%. This problem affects everyone.

I just had a 20 BTC sell order there. This bot for some reason buys into it with many tiny buys, each buy worth 0.86 USD. Every one of them get's a transaction fee of 0.01 USD because of their rounding. The proof is below. This example is hypothetical because I did not let the bot eat my entire wall that way, I pulled the order before he got that far.

In essence this 'trick' makes me pay a fee of 0.86/0.01*100% = 1.16 PERCENT. While I'm entitled to a transaction fee of 0.2%.

Everyone affected by these tiny buys/sells is charged almost 6 times the money. Let that sink in guys. A transaction fee of 1.16 percent because their system allows too small orders. They force this upon you and essentially, (by intentional design or not, it doesn't matter, you are at heavy disadvantage) take your money illegitimately due to this transaction fee bug/misimplementation.

--------------------------------------
Hypothetical calculation example:
--------------------------------------
Say I wanted to sell 20 BTC at $432.01, and it is bought completely by these mini buys:

The fee of this transaction should be 0.002*20*432.01 = $17.28
Due to the trick, actual fee is: 0.0116*20*432.01 = $100.23

Fee I agreed on when putting in sell order: $17.28
Actual fee that would be paid due to the exploit: $100.23

Bitstamp illegitimately takes 83 dollars in this scenario
---------------------------------------

As I said, I didn't let it come that far, I pulled my order before it could eat through my order that way.

Picture below is proof of the behavior.

Proposed fix: Make $5 orders the minimum for both API and Web interface. Preferebly $6 to also avoid exploits through setting orders above 20% of current market price. Then all problems are solved

http://i61.tinypic.com/98bzao.png


---------------------------------------------------

Bitstamp did reply to my inquiry about their transaction fee calculation.


They further explain their rounding policy, which is fine, but did not explain (yet) why the order limit is then not $5 so that the exploit as mentioned in this post cannot happen. The fact remains that an order limit of $1 in combination with the rounding off to $0.01 causes any order of $1 or smaller to be charged five times the fee.

Whether or not the low $1 order limit is intentional in order to harvest these fees or not, I will leave upon you (the community) to decide. The willingness of Bitstamp to change the order limit will show us as a community if it is intentional or not.

---------------------------------------------------

Edit 09-05-2014: Bitstamp is increasing their trading limit to $5 on May 15th, after which the exploit as described in this post will no longer be possible.

Very messed up if correct.

This is a shambles if correct, they could be stealing $100,000s a month in this way.

Have you posted this in the Bitstamp thread or on reddit? I would very much like to hear what Stamp has to say for themselves.

https://i.imgur.com/xh2bILT.png

https://www.bitstamp.net/fee_schedule/


the minimum buying order is $1.... it is really strange that someone could execute a $0.86 buy order

https://i.imgur.com/gTfGtsz.png

I see they actually intend and understand this ripoff. If they want more money, they can simply split orders up into smaller ones, brilliant.

1.16% fee in place of a promised 0.2% fee is inexcusable, and this is for a sum of 8640 USD, so not just "dust". This rounding borders on a scam.

Then its their API that isn't restricted to minimum trade of $1.
The minimum trade amount needs to be changed to $5 on both api and website, so that the charged fee always remains @ 0.2% fee of a trade.

Agree, simple solution and no side can complain. $5 is plenty small enough per order.

Yep, this.

That's a good solution, but will Bitstamp be so inclined as to give up this profit? I hope they are.

Maybe we can complain loud enough so as to pressure them.

yes I think $5 is a good minimum for an order, it would also reduce bots spamming their engine and API...

They have to if they want to be seen as a CREDIBLE exchange, as this glitch affects ALL its customers.
Its a 'rookie' mistake to even have different rules for api and website.

as if 0.5% for < $500 isn't ridiculous enough, the most honest and reliable exchange in my experience so far has to be btc-e with their 0.2% flat rate, why would you even use bitstamp in the first place  :-\

+1 lol

its stupidity tax.

Oh c'mon guys, this is math 101. This happens with so many transactions on a lot of things on a daily life. It's rounding and normal.

OP did transactions with a BOT lower than $1 as per indicated limitations. He probably was hoping that he would not pay transactions fees.

If you are doing that many transactions, just do larger transactions where rounding is in your favor. Why is he making thousands of really small transactions? He's probably just slowing down their system anyway.

I'm on with Bitstamp on this one.

Here's an example where I paid less than 0.5% because of rounding:
I bought 0.08774 BTC and only paid 0.047%

OP was the one being bought into by someone making trades lower than $1.

By allowing it Bitstamp are slowing down their own system if you think this sort of activity would slow it down.

Troll fail.

Not trolling. I'm just exposing this is how rounding works in many scenarios. Even when you pay taxes at your store.

OP was selling, this small transaction rounding problem doesn't happen if you are buying BTC, only if you are selling in small chunks.
Just try it on the exchange and you will see.

I had a 20 BTC sell order there. Someone (maybe even Bitstamp splitting up orders) bought into that with tiny orders, causing me to pay 6x the fee I am supposed to pay. I had no choice in that matter. When hitting the sell button I agreed to sell at 432.01 at a fee of 0.2% over 20BTC, not 1.16%.

It can happen to EVERYONE on any type of order whether  Buy/sell - market/ limit , as long as bots are nabbing/putting trades with < $5 trade ( or equivalent in btc)

You are trolling because you dont seem to grasp how the transaction was carried out. OP had no say in the chunks they were sold in, he didnt add them in chunks of less than a dollar.

Oh right well as long as it doesn't effect you when your buying just taking money off you when your selling.

The screenshot is in the OP  ::)

EDIT: You have less than $500 30 day volume, I dont think you trade enough to notice and certainly not enough where you would care about this so your position is not surprising.

Great post shedding light on a common problem. We track Bitstamp in our index and I hope to see them implement a $5 minimum trade, there is absolutely no reason not to.

http://log.bitcoin-assets.com/?date=02-10-2013&bots=true#338159

Nothing wrong with rounding but charging a fee on every chunk is a plain old SCAM. Charge the buyer not the seller! Greedy Slovenian bastards!!
Can someone make a list of exchanges that charge the same way ?

cangine ? why ?  :D

Yea you're not getting it.

Even with rounding, when you put an ask / bid in you can determine your fee.... and make decisions based on that knowledge.  Take the TOTAL amount, multiply by the percentage, then round.  The fee should be accurately calculated.  This is how it is with other exchanges.

That's not the case in this example.  When OP puts a sell order, his fee is at the mercy of buyers... which is inconsistent with how other exchanges behave.  This is definitely a flaw.

yep, that's ridiculous.

Its the Stamp Misappropriateability! 

I've posted a link to this thread on the Bitsamp Facebook page as well as tweeting to them asking why this shouldn't be fixed... and if they actually care to lose my account. I really want to hear them come back and say this is an error and we can rest assured it will be corrected so that the trading fee you think you're paying - is ACTUALLY what you are paying for the trade. Nothing is more important right now that being able to trust the exchange you use and if I get the slightest whiff that they're going to do this the wrong way- me an my trading volume are going elsewhere.

I like how the proposed fixes are all related to minimum order sizes rather than the obvious fix:

STOP ROUNDING TO THE NEAREST CENT!  You already support 8 decimals of precision on BTC.  Do it for fiat, and simply round down a user's available balance to 2 decimals when they want to pull money out.  Professional sites have done similar things for years.

Hell, even inept Karpeles at MtGox knew to do it that way.

I would like to report that on Kraken the fees are properly charged without rounding even when an order is filled by splitting it into several trades: Yesterday I had a buy order for 1 BTC on Kraken and this order was split up into 4 different trades: 0.49593 BTC, 0.1 BTC, 0.1 BTC and 0.30406 BTC.  For every one of these 4 transactions I was charged a separate 0.2% fee. The total transaction fee I was charged added up to about the same as it would have been if the trade would have been executed in one single transaction, in fact it was a little lower. So when you trade on Kraken, you will not have the inflating fee problem due to rounding reported by the OP for Bitstamp!.

Not true.  Say you place an order to sell $500 worth of BTC (or $5,000 or $5,000,000).  The fee is rounded on each trade that executes against it.  So if 1,000 $0.50 buys are made against it you don't have a single $500 sell you have 1,000 individual sales of $0.50 ea (that are all rounded up).  In any exchange you can only control what order YOU place not what orders the counterparty will place.  The size of each individual trade is based on the orders on both sides of the book.   The fee is based on the size of the trade not the order.

Thanks for that! I've been meaning to look into Kraken but this just takes what had been a, "Yeah, I should look into that"... and turned it into I'm opening a Kraken account this week. Have you traded There for long?

If you bought a gallon of milk and they split your order up into 64 ounces and rounded the price of the sales tax up each one up you would be pissed too. 

Again, it is ok to do this for small single orders.  It is not ok to split up a large and immediately executing order into many parts.  It is ONE order.  If there was not enough liquidity and part went now and part later that is different, but that is not what is happening.

I opened my account on Kraken just about a month ago. Its a small exchange but so far it looks secure and professional to me, with lots of advanced trading options (you must be verified to be able to use them). Also they have responsive tech support to answer any questions you may have.

So much Kraken spam in this thread, reason more NOT to use them!

If by spam you mean "Reasons Bitstamp is currently giving their customers (me), to look someplace else to do business", then - yes it's spam. My typical 30 day trading volume is 300k at minimum. That might not be much for some, but it adds up to a lot of fees for me. Maybe you want to be a bitsamp fanboy and hate on the competition for no good reason; but guys like me work on trust and facts. If I can't trust the place where I do business, I bounce to a place I can. That's a fact. 

LOL there are a total of 6 posts in the entire thread mentioning Kraken, including this one. I only mentioned my recent experience with this exchange since it relates to the issue with splitting up orders into several smaller trades reported by the OP, and someone earlier mentioned they were looking at Kraken for an alternative to Bitstamp.  My second post was in reply to a question of another user about my experience. I don´t know how you conclude from this that the thread is full of Kraken spam? I certainly have nothing to gain from sharing my experience about this particular exchange here, other than possibly helping other users make their own decisions.

BTW I also use Bitstamp and have not had any negative experiences with them so far myself, but do appreciate that the OP started this thread to alert us about their inflated transaction fees.

And I totally agree with ninjabenja´s last post.

And you didn't notice you were being ripped off?

Fee malleability  ;D

This issue looks like it was pre planned and calculated. It stinks.

Has it always been this way or was there a recent change ?

If I had I would have been the OP. Shit happens. I guess that somehow as "obvious" as it must have been, none of the other handful of people who use stamp noticed it either. Our bad.

Still no reply to my comment on their facebook or the tweet I sent them.

Link to tweet? I will check quality and retweet.

Ugh, rounding policies always cost the end user, which is fair enough provided that it is reasonable and can't be exploited. If something underhand is going on with the exchange in order to screw sellers then that is very concerning.

If I'm not misunderstanding this - the OP's claim to fraud is not as regards to the rounding, but that Bitstamp is intentionally breaking up large orders into micro sized ones, in order to benefit from the rounding. The later would definitely be fraud if it were proven true.

It has always been this way.
And it's been written on their website as far as I can remember.

The fix should be simple; don't allow orders to be input that generate a fee less than $0.01.

So as other have said; a $5 minimum order; which IMHO is very reasonable.  And this does not required any messing with the rounding/trading system which could be hazardous.

People, please stop being dumb and suggesting retarded fixes.
The correct fix is to learn how to fucking count with more decimal places.

Has anyone had a reply from support or Hazek that has contacted them?

Mimimum orders are not a solution. Sell/buy orders are not always done in whole but can get split up.

See this example:

• seller A and B are each selling 1 BTC
• I place my order for 1.00001 BTC
• my order is executed in 2 steps:
  
  • buy 1 BTC from seller A
  • and 0.00001 from seller B
  fee is calculated for each transaction independently and this is where error occurs - minimum order size was in place, but still
  

As other said - they should calculate fee to 5, maybe even 8, decimals – everyone (except Bitstamp) gets happy, but then again happy customers – happy CEO …

And I don't think anyone is going to matter much if there are remainder orders which are less than $5; as it's hard to abuse the system in this manner.  Whereas the current system allows to put in repeated tiny (less than $1) orders.

It doesn't matter if you occasionally get screwed for a couple of $0.01 fees the problem is when this is done repeatedly and consistently so that the amounts add up to huge fees (such as in the OP where clearly someone is placing 100s of $0.86 orders these are not remainders).


Because changing your entire decimal precision could cause unintended bugs....
This kind of change would take months of testing before live implementation for an exchange of bitstamp size.

Where as order limit could be implemented overnight with no chance of serious (money losing) bugs.

As opposed to what? Intended bugs?
Just because something is inconvenient or hard does not mean it should not be done, applying half-assed pseudo band-aids is exactly what you should not do.

When you have zero experience in a field, and no valid point to make, your best course of action is usually to refrain from talking.

well tell me how would you collect a $0.001 ?

It was explained upthread. Charge the account balance 0.001, round balances down to the penny when withdrawing.

Fees should be charged per buy or sell order, not by each split up transaction. It will not be difficult to change it. Having developed an exchange, changing fee structure should be trivial. Bitstamp do not want to make changes because they will make less money from fees. The only way to pressure them to make modifications is to stop using Bitstamp.

Think more, I already gave the correct solution.
What you're suggesting is strictly equivalent to charging the fee at the transaction level, that is because you need to be able to handle partial fills properly.

I have received a response from Bitstamp. I request that you all re-read my starting post, as I changed it to better reflect the situation as it is.

Bitstamp did reply to my inquiry about their transaction fee calculation.


They further explain their rounding policy, which is fine, but did not explain (yet) why the order limit is then not $5 so that the exploit as mentioned in this post cannot happen. The fact remains that an order limit of $1 in combination with the rounding off to $0.01 causes any order of $1 or smaller to be charged five times the fee.

Whether or not the low $1 order limit is intentional in order to harvest these fees or not, I will leave upon you (the community) to decide. The willingness of Bitstamp to change the order limit will show us as a community if it is intentional or not.

You should point them in the direction of davouts solution as well which is probably fairer.

Yes we know fiat currencies has 2 decimal places but a system can store more than two. Just use more decimal places and round on withdrawal.

Has Bitstamp commented this thing at all? There are multiple ways to fix this... and being quiet makes it worse.

See https://bitcointalk.org/index.php?topic=597647.msg6590228#msg6590228

Actually there happens to be a single correct way to fix it.
For pretty much the same reason there is one single correct way to count.

Their answer is simply magical.

My point is that the order limit can be place immediately; while the more permanent solution of higher precision decimal rounding can be tested and implemented over a longer period of time.

So your solution would be what?

a) Leaves the system running as is (stealing fees) for months while safely developing and test the change in rounding.

b) Rush out a change in decimal precision without adequate testing and hope it has no bugs?

Try to discuss the point rather than semantics and ad hominem

This exchange is straight up scamming it's customers.

https://twitter.com/NinjaBenjamin/status/463721726493589504

Your display of braindamage does not constitute a point in any way.



+1

Transaction Fee Malleability?


Allegation that DGEX.com also does this: https://nextcoin.org/index.php/topic,4705.0.html

Personally, I ditched Bitstamp for Bitfinex because stamp suck.

That hasn't been my experience. Deposits at BTC-e were credited days or even weeks late. Mysterious three figure fees approaching 1% per deposit were deducted and customer service queries went unanswered. In fact I moved from BTC-e to Stamp for exactly these reasons and so far I have been very happy with them.  

Fixing it by applying the fee to the order instead of by transaction is also valid. The way this would be done is to reserve the full fee (based on order quantity) when an order is opened, but then credit back only the unfilled portion of the fee when the order is closed (canceled or filled). Round the first up, round the second down (except in the special case of zero filled). They get the generated rounding breakage they seem to want, but once per order.

It is indeed a much less scammy behaviour than what's currently happening.
That doesn't really work for small orders though, and is actually harder from a technical point of view, than simply increasing the precision of the stored amounts.

I think that'd be a good blog post subject, technical and accounting issues related to rounding and the best ways to solve them.

Not sure this will satisfy davout but;  ;)


Bitstamp have updated their minimum order amount to $5:


https://www.bitstamp.net/article/bitstamp-minimum-trade-changing-to-5/

it wont, because no matter what Bitstamp would do he wouldn't be satisfied..... the reason is simple when you know who he is  :)

Hence the  ;)

I'm slightly amused by the implication that the fact I operate another exchange somehow makes the hard figures, and the shittyness of the workaround any different :)

This is ridiculous: there would be no issue at all if they let you set embedded fees, e.g. I buy 0.0023BTC for 1$ gross price, I get 0.0023/1.002 = 0.00229541BTC net amount and the difference = 4.59µBTC is deducted as fees

Writing here to voice my support for the '$5 trade minimum' fix.

On the plus side: fees can be expected to be near the advertised fee percentage from now on.

Also on the plus side: Bitstamp reacted fast to customer complaints and implemented a fix.

Also also on the plus side: Automated high frequency/low volume traders will hate it.

On the negative side, it seems to bring out the trolls. Oh well, can't have everything.


tl;dr To some it might look like a hack, but it's a solid hack, so no complaints from me.

+1

no need to raise the limit with embedded fees. At current rate you could even lower it down to 10 cents and still have a negligible rounding within a satoshi

Agreed - issue raised, exchange responded with a reasonable fix.  Job done.   

We have had far worse concerns ignored by other exchanges in the past, let's face it.

I still do not agree with this partial fix of Bitstamp. If I place one order I want to be charged for just one order, regardless if it is bought into by three different orders. I decided to place one order so should be charged for just one. Bitstamp is just a scammer exchange with a bad reputation.

They are still going to charge much higher than the posted fees in some cases. The minimum is $5, that's great. But what about $5.01? The fee should be $0.01002 but instead you will be charged $0.02. Okay the breakage seems capped at double now, but that's still a lot.

Their base fee is .5%.
If you place a $5 order you should be charged $0.025, but instead you will be charged $0.03.

In other words, even with their "fix", you'd get charged 20% more than advertised.

I see folks cheering because their answer to the situation was fast...
It's as if the people here would be happy to be served a plate of steaming shit in a restaurant, instead of actual food, as long as the waiter's fast.

Except that the subject of the thread says 0.2% which leads to the (almost) double fee on a $5.01 trade

so lets close our accounts in Bitstamp and start using your shitty exchange right ? because you've got it all figured out....

It doesn't matter whether he has an exchange thats worse or better he's demonstrating that there is still an issue. Rather than attacking him why don't you raise some valid counter points?

Agree.

It is good to see that Bitstamp has done some improvement to reduce it though. Hopefully they will increase precision eventually.

For now I am happy with the fix as it solves the exploit I mentioned in the OP, however, better rounding precision would be preferred as it is still possible to pay double.

Effectively and through proper planning it is now only possible to have people with 0.2% trading fee pay 0.4% if an order is completely bought by a streak of $5.01 orders.
This is a direct result of their trading fee policy which is mentioned in their ToS. I still do think this needs to be fixed, but we need a lot more community effort to get this done as this is mentioned in their ToS and we agreed to that by trading there.

The main issue I wanted to raise here is that their minimum order limit was not in agreement with their own minimum trading fee. This allowed for a disproportionate rounding error at the customer's expense, and was a (hopefully unintended) exploit to their fee rounding policy. I am happy to see that this particular exploit will be fixed and that their minimum order size has now been attuned to the lower end of their minimum trading fee.

- The exploit was illegitimate and bordering on illegal, which is why it was imperative that it had to be fixed.
- The trading fee rounding is mentioned in their ToS.

This is the main difference I want to stress.

I do however believe that a fair trading fee rounding would be much preferred

because it is really retarded, this simply imply that he has no fucking clue about what he is talking nor know how to solve a simple math issue, because the same thing happens everyday in your daily life and almost every where, when you go to the gas station they round up the price and I don't see anyone complaining and crying there, and when you go to the grocery store to buy a kilogram of potato the same happens because it is simply never an exact kilogram, yet I don't see people crying there as well.

and by the way, when you trade a fucking $200.01 and pay $0.01 more as fee, of course a $0.01 will change your life right ? and if you do that 100 time a month this would sum up to $1 in total, this will get you broke!!! but of course it is not about the $0.01 it is about how can we live with the fact that there is a successful and efficient business like Bitstamp, No we have to fuck them hard to have peace (fucking jealousy)...

Only partially agreed. Not all 'ad hominem' arguments are fallacies.

If a Visa exec writes about the dangers of Bitcoin, we'd at least suspect him to be biased. If the owner of a competing exchange vocally slams Bitstamp, it at least invites additional scrutiny.


I'll revisit this topic iff an actual user reports another large enough divergence between advertised and actually paid fees, say 2.2 instead of 2%, and more than just a dollar. Until then, I consider this a decent fix.

Clearly the one with "no fucking clue about what he is talking [about]" is you. Your comparison is invalid. Read the OP to find out why. Davout is accused of being a scammer by others on this forum, so I'm not here to defend him, but on this issue he's right.

The $5 fix is better than nothing, but not the most honest fix. It suggests that the original problem was a deliberate method by Bitstamp to increase fee revenue.

No, you get to keep your bitcents there.
Your dysfunctional brain would be much better served by a dysfunctional engine, neither of them displays the simple ability to add and multiply numbers correctly.

ok Mr.smart ass, explain to me how is this different, explain it like I am 5 years old.


so your brain and logic is the functional one ? amazing...

If it was actually explained like you were five years old it might already get too technical for you.
See, the prerequisite to explaining stuff in such a way, is that the person you explain stuff to is assumed to have mental capabilities that are equal, or exceed, those of an actual five years old.



Start with additions, when you're confident about them, move on to multiplications, you'll see, it's easy, a multiplication is just a repeated addition after all.

congrats, you just lost me here, and welcome to my ignore list.....another fact, bitstamp still makes millions when you are crying as a baby here, this didnt change for sure, see my limited retarded brain still understand that  :D

So is BitStamp going to refund the excess fees they mischarged customers?

Still waiting for the Fix ! Bots having fun meanwhile  ::)

Time             Price         Volume      USD           
21:00:17   452.58000   0.0022   1.0100
21:00:17   452.59000   0.0022   1.0100
21:00:17   452.60000   0.0022   1.0100
21:00:16   452.61000   0.2972   134.5055
21:00:07   452.61000   0.0022   0.9981
21:00:05   452.61000   0.0022   0.9981
21:00:05   452.61000   0.0022   0.9981
21:00:00   453.47000   0.0022   1.0100
21:00:00   453.48000   0.0022   1.0100
20:59:51   453.49000   0.0022   1.0100
20:59:35   453.50000   0.0022   1.0000
20:59:32   453.50000   0.0022   1.0000
20:59:31   453.50000   0.0022   0.9999
20:59:28   451.51000   0.5533   249.8211
20:59:23   453.50000   0.5060   229.4500
20:59:22   453.50000   0.0022   0.9999
20:59:18   451.51000   0.4508   203.5385
20:59:16   451.76000   0.2570   116.1023
20:59:13   453.51000   0.0022   1.0100
20:59:13   453.52000   0.0022   1.0100
20:59:12   453.53000   0.0022   1.0000
20:59:07   453.53000   0.0022   1.0000

I never had a gas station round UP on me... Damn right i would complain, it's just not fair.

Adding more decimals is the only fair solution for both parties.

Mischarged is going to far. It was according to their posted (but dumb) policy.

They can certainly refund if they want to do that as a customer relations move, but I don't see any mischarging here.

Either this, or round fairly, because on average it would even out.



+1, it's not like it wasn't clearly laid out.

I think this could likely be gamed unless rounding were done randomly (and then it is difficult to audit) and it would not average out.

This is wrong because there is automated trading that does a high volume of trades in a short period of time. Even if you don't do this yourself, if your orders are traded against by algorithms you may get many, many executions. There is no reason to assume 100 trades in a month is a lot. You could get 100 trades in a minute.

I am aware of bots and the way they work but 100 trade a minute !!! why ? it is not like you see major price moves over the course of a minute...it is just plain retard to do that then complain about paying more fees, and again, rounding prices up is all around your daily life, you just don't feel it or know about it.

if I had an exchange I would stick to the $0.01 as well, there is nothing less then a penny so complaining about it is just plain stupid, the only thing that could work is taking fees in Bitcoin but I don't see how you can work this out, because at some point you have to sell the coins and I also don't know how can this be reported for taxes...   

It may be retarded but the problem is you don't have to be the one doing it. You can put in one big order and bots can trade against your order in a 100 (or 1000) little pieces, costing you up to an extra 0.01 in fees each time. This may add up to a lot.

sure, that is the case of the OP, and now 5$ minimum will limit this problem, unless a retarded person set his bot to execute orders of $5.01 to pay more fees !!! except if the subject here is not only the fees but it is an accusation to Bitstamp of trading against people's orders to make more money from the fees, is this what you want to say ?

Yeah, you're right, completely forgot about this. It can totally be gamed, just sell your coins in small chunks to avoid the fee altogether, guess we're back at "the correct answer is to count properly with higher precision"

Theoretically, even counting at an arbitrarily large, but finite, precision can be gamed this way. Even with 8 digits of precision, buying satoshis one by one at a 100 Currency/BTC spot and a .5% commission rate would lead to 5E-11 BTC fees being rounded to 0 (given a 1E-8 precision).

Even with a minimum trade amount you'd have some small rounding artefacts on fees due to small partially filled order remainders being executed against, so you have to take an extra step.

So, to summarize, the most correct course of action is to
 - Count with an arbitrary precision, preferably very small,
 - Impose a minimum order size that is a very large multiple of the precision
 - Remove bids from the book when their remaining volume is smaller than (precision /(2*fee))
 - Remove asks from the book when their remaining volume is smaller than (execution_price * precision / (2 * fee))

Numerical example for the bid side
 - Spot is 500 USD/BTC
 - Precision is 10^-8
 - Fee : .5% (0.005)

https://i.imgur.com/Jx7Ki4q.gif

(The reasoning is similar for asks but I can't be fucked right now)

Who knew so much fun could come from working with finite-precision numbers :D

i actually don't think a 5$ minimum fixes the rounding issue.
strictly speaking: only having orders that are multiples 5$ would fix it.

i think Bitstamp needs to consider adding more decimal places, as has been suggested

Exactly. You can never reach 0.5% fee when the rounding is always UP.

wow that is incredibly dishonest and you know as well as i do, anyone going over logs would have seen that and been, to a buddy of course, "man you wont believe this but ive figured out a way to more fee's".

meh

Just for the record you retarded forumites, I had a discussion with Nejc about this problem in 2012 (was said to be worked on).

I put out a warning even on this forum in 2013 but noone gives a shit. (You can look it up, history if now banned username 'kakobrekla')

You deserve even worse tbh. Delete your wallets, at least you wont make the scammers richer.

And as for the new fee structure, it is still possible for their trading engine to take 1 satoshi from your order and charge you 0.01 usd for it. Calc the fee amount on this one.

Enjoy.

Oh yeah, side story, I also K-Lined Nejc from freenode in 2012 because of spamming. Such ethics.

What an unprofessional exchange, it turns out. Money greedy bastards. Rounding for your favour is a criminal act. Book keepers have went to jail for that in the past. Although I have a verified account in Bitstamp, I hope to never use their services again. Asking dumb questions about my money was one thing, but stealing in the given way destroys all trust I had for them.

Hello Bitcointalk community,

I have been around as a visitor for a while and registered only today. I learned a lot from this community about Bitcoin, so thanks a lot! I hope it's not against any rules to post in a year old thread as a noob (who is getting familiar with trading only since a short while.)

I just stepped into this "fee trap" at Bitstamp and like to ask if someone has experience with  trading on Bitstamp and could explain why on earth would a bot want to split my sell order of 0.1 BTC into five small buys ? The buyer also has to pay fees for each buy, and due to the rounding of fees pays more than what's listed in the fee schedule.

BTC   USD   Price   FEE
-0.020559   5.03   244.9   $0.02
-0.020559   5.03   244.9   $0.02
-0.020559   5.03   244.9   $0.02
-0.020559   5.03   244.9   $0.02
-0.017764   4.35   244.9   $0.02


In this case I paid $0.10 of fees instead of $0.07 (0.25%).
any other explanation than what mmitech suggested earlier in this thread ?

Does it make any difference on how orders get split if one places an instant or limit order ? I think I read in this thread it doesn't, so there's no other way than keeping your fingers crossed that an order won't be split too often to keep the fees down ?

thanks for any hints
netpeak

Well this is quite surprising....In a way. Then again, it really shouldn't...