Topic: Icbit.se, the bucket shop. (Read 7982 times)

Thank you for the thorough explanation. I now understand the logs. It's pretty hard to get know how much you earn/lost from a particular trade if you have a few opened and keep them for two weeks or more. Addition of all variations during that time won't give you the number of the gain/loss for a single trade but the gain/loss of aggregate positions. Not to mention the trickiness of adding a dozen of 8 decimal numbers...

2013-02-02 Buy $20.00
2013-02-02 Buy $20.05

Then later that day after you bought, clearing was $20.7095, so on the two positions you were paid variation margin due to the clearing price of $20.7095.  Then on 2013-02-03 the clearing was $21.0750 and you earned even more. 

Then each day the clearing price was lower so your variation margin was negative.  Then you sold at $20.5301 which was lesss than the most recent  previous clearing price -- $20.8101.    The very last transaction in the log is the 0.5% trading fee.

So your trade resulted in half a percent trading fee when you bought (the two 0.005 BTC trading fee) plus the half a percent trading fee when you sold (the combined 0.01 BTC trading fee entry) and the sum of the margin variances.  Now remember, the variation margin is calculated as:
 VM = -(1/PriceClose - 1/PriceOpen) * S;
which is
  (-(1/PriceClose - 1/PriceOpen) * 10) * N contracts

For instance, the last variation margin had a clearing price of $20.8101 and you sold at $20.5301.  So the variation margin was:
  VM = (-(1/$20.5301 - 1/$20.8101) * 10) * 2 contracts
  VM = (-(0.0487089687 - 0.0480535893) * 10) * 2 contracts
  VM = (-(0.0006553794) * 10) * 2 contracts
  VM = (-0.0065537940) * 2 contracts
  VM = -0.01310758 BTC

(somewhere the calculation must be getting precision issues, as my rounding at 8 decimals is off by a really small amount from the 0.01310526 reported in your pic.)

Change in variation margin IS your gain/loss. It is just that it might be spread over multiple log lines.

E.g. suppose you bought at 20, then clearing price that day was 21, you get profit. Then you sell at 20.50, you get loss.

Simply sum this profit and loss and you'll get your overall profit over this trade. (Also, don't forget fees, they can totally ruin your profits, if price difference is smaller than ~1% you get overall loss.)

Thank you for your reply. I simply expected to see in the logs how much I gained/lost with closed position which is how most exchanges accounts works. When I close position I'm not interested in seeing how my variation margin was at this point. Logs lack information about closed position gain/loss.

I posted to this thread not to undermine icbit.se credibility but because it was the first one which showed after conducting the search for "icbit".

If this is inappropriate admin please move it to the correct one.

From what I can decipher from the image, it looks that the negative amounts are just the ongoing "mark to market" (variation margin). This is just how such a futures platform works. At least once a day, an effective new reference price is established for your open positions. This new price is based on the general market movement. At that point, the gains or losses with respect to the previous reference price are credited or deduced to your account. Since the BUH3 rate went from 21.44 to 20.53 your open positions incurred a loss. When the price goes up again, it will incur a gain. When you close some contracts, just the "weight" of your position is deceased. Only if you close all your positions, then you don't participate in the variation margin.

Since on a futures platform, you do not trade against the platform, but against other market participants, in case your position looses, some other participant's position (with the opposite direction) gains, and vice versa. The regular "mark to market" serves the purpose to prevent that some trader all of a sudden can't cover the losses.

hope that helps

---

Hint 1: Note, futures trading is not a bucket shop. This thread here was just opened by a competitor of ICEBIT, please take it with a grain of salt.

Hint 2: there is a dedicated thread for ICBIT in the "project development" section. When you post there, then usually "Fireball" (the author of the site) reads and answers your questions....
https://bitcointalksearch.org/topic/icbit-derivatives-market-usdbtc-futures-trading-live-50817

Hi,

does anyone know why my trades shown on the picture below have negative profit? I bought 2 contracts at 20.00$ and 20.05$, then sold at 20.5301 and logs show that I lost...

I don't really make those, as you've possibly noticed. Not like this is the first time or something.


This is nonsense. The need for trustworthy people isn't going away. They just need to be actually trustworthy, rather than whatever the cat happens to have dragged in.

how long until the "i warned you all" post?

So in the end the conclusion is that MPOE raised some accusations, but failed to prove them convincingly.

The fact that Fireball could indeed manipulate ICBIT is correct -- yes he can easily, as well as M.P. can easily manipulate MPOE.

You brought that fact to everyone's attention, which is fine.

But you failed to provide really conclusive evidence, and the way you presented your accusations leave a bad taste. You mixed several possible failure modes up, you engaged into a not related technical discussion, which counts as derailing the argument, you used some really fundamental stuff (like what PGP is or what futures are) in a way as if you didn't understand the basics of the terminology. And you repeatedly mixed unproven assumptions and clever omissions under the facts, as if your intention was to pull of a PR stunt.

And indeed. Everybody had now been warned about the situation regarding ICBIT and MPOE/MPEx likewise.


For me, the only conclusion is that what Bitcoin really needs are trading sites which are constructed in such a way that the "reputation" or the role of the "trusted owner" is replaced by a cold and unbiased trust mechanism, built in a similar way as the blockchain itself.

So in the end the conclusion to this is that "Fireball couldn't be manipulating the price because once he manipulated it in my favor", says some "customer" or other.

I guess everyone's been warned.

I was suggesting a) publishing it on your company blog (do you have one), if not the site and b) making a new thread concerning this issue with going into the accusations of the blog post in OP in detail.
I am not the one running a "company", do as you like. But don't complain about the consequences later.

Who are you, usagi?

I always discuss all issues related to the ICBIT trading platform in our thread.

However, this is just hilarious:


I reported this thread to moderator as inappropriate.

Other than the obvious point that just calling them false doesn't cut it, why exactly do you expect or imagine you're entitled to private communication? You're not in the WOT, I don't think you've ever talked in #bitcoin-assets, you don't, practically, exist (reference). We're not business partners, we're not friends, what exactly is this entitlement based on?

Until others pointed this out to me I had no idea you were in any way authorized to even talk for that project (and, formally speaking, you are not, I don't seem to see a GPG signed authorization by anyone, for all we know in two weeks someone could register Icicle as a forum handle and claim you were not authorized to speak on icbit's behalf at all).

I'm not going to go into the rest of the assertions you make, mostly because on top of being mostly false they're also quite besides the point.

That aren't false accusations until you prove them false, they are just accusations.

And MPOE only linked to the blog post which accuses you of fraud.
A professional response would be an official announcement which rigorously debunks this.... We're waiting...

MPOE - very bad move of you. You could ask questions directly, instead of putting up a thread with false accusations. Just try to offer better services - that's the best way of advertising. Making false accusations is not.

ICBIT is an open market, where everyone has equal chances and possibilities.
Everyone can participate, and unlike "bucket shops" everyone is also welcome to arbitrage (using spot market, for example), make the market (using advanced algorithms).

Lack of liquidity is an obvious problem for any exchange, but we are rapidly growing, so hopefully this problem will soon disappear. If you remember, MtGox had similar issues in the beginning of its way, because it also was and is a market open to everyone, with its own order book.

Thanks to everyone for defense, this is very appreciated.

This is true.


This is also sort-of true. For one, you can use any of the many brokers (such as coinbr.com) to trade on MPEx without paying the registration fee (you will probably face higher per-transaction fees, but I think even so they'd be less than what icbit charges).

I'd like to state in any event that there's absolutely no "political" (or however you'd call it) motivation in any of this on our side.


We've been trying to get real futures going for a while (as pointed out in the original article linked from the first post). There's no plans to introduce BTC futures nor were there ever any (nor do BTC futures make sense, it'd have to be dollar futures, as explained in the same original article linked in the original post). There were hash-based contracts at some point, that ended with significant losses for the marketmaker.


This makes any further discussion impossible. If you really think that I don't see what can be said further, we'll just have to agree to disagree.

MPex has an options exchange, in direct competition with ICBIT's futures market.  Oh, and they benefit greatly from being without competitors, it actually costs 30 BTC just to sign up on MPex!

I don't say this is MPOE's motives, but it sure looks suspicious.

It all depends on how much liquidity "the manipulator" had, obviously it could be way more than "daily liquidity", whatever it means.


In which case you get amount equivalent to initial margin, which isn't bad.

But, yes, due to a lack of transparency a web site can actually do whatever it wants to, in theory. (This is why I'm advocating trading via blockchain. Not just advocating, I'm actually working on this.)

But back to our topic, just because exchange could be dishonest, it doesn't mean that it was.

As I already mentioned, I talked to Fireball before he implemented ICBIT, and I know he had big plans. And I'm sure he is a very good programmer.

I'm sure he put at least $10,000 worth of his time into this project, likely more. This only makes sense if he wanted it to become THE Bitcoin derivatives exchange.

Does it make sense to spend that much effort to run a scammy low-volume bucket shop? Absolutely not.

(BTW making it a scammy bucket shop might be even harder than to make it a true futures exchange since it would require a lot of ad-hoc doohickeys, like a functionality to block a specific trader right before clearing time. It also requires much more daily attention.)

So, again, Fireball could do it, but I'm 99% sure he didn't.


MP's article was just a piece of evidence. OK, we can consider it. Perhaps ICBIT should be more transparent. I want to see full trade log, full information about what happens during clearing time and so on. I don't understand why this information isn't available.

But now you claim that you know that ICBIT is fraudulent with certainty. This is simply bullshit, it just screams cheap propaganda. Do you want to start MPex futures exchanges or something like that?

Only two things imply that exchange operator might be behind it: "changing logs" (which might be simple a mistake on trader's side) and "locking MP's broker's account".

But if you think about it, locking MP's broker's account doesn't make that much sense since MP's broker could as well put a bid wall beforehand. Posting bids in last seconds really gives you nothing. So it could be simply a technical glitch.

Of course, it is possible that exchange operator is doing this, but it might be just some trader trying to manipulate the market, thinking that he has more resources than others. As Stephen have noted, you don't really need that much money to manipulate such a thin market.

As for economic sense, see here, educate yourself: https://bitcointalksearch.org/topic/m.1340759 (I've outlined this situation back in June 2011.)

I you wonder why manipulator perished not long after MP got involved, it might be simply because MP added more liquidity on long side, thus depleting manipulator's reserves. So it is not a coincidence, there is a rational explanation.

And, by the way, I think your forum signature is wrong: "THE Bitcoin Stock Exchange" will be based on Bitcoin protocol itself: colored coins to represent securities, multi-party transactions for trade, perhaps even orders will be advertised via blockchain if we figure out how to solve bloat issues. Web site you're pointing to is definitely not the definitive solution, it is just a web site ffs. Use of PGP to authorize orders is cute, but (if I'm not mistaken) people need to send their bitcoins to this web site in order to trade, and so they have to trust some random person's web site. That's not how Bitcoin works. You can as well add LR USD payment option, it can work exactly the same way. So there is definitely nothing Bitcoin-specific about MPex, aside from Bitcoin being the only payment system it accepts (which is more like a deficiency, it simply makes sense to accept as many currencies as possible if you aren't utilizing any Bitcoin-specific features anyway). This isn't related to ICBIT discussion, of course, just noticed your signature...

None of these are inconsistent with "my intended message". In fact, they support it quite well.


This is inconsistent with the theory Fireball is right: a 1000% of daily liquidity failed to do anything.

Obviously Fireball would claim "lack of liquidity" is the problem: he has a vested interest into maybe tricking some idiot to go to war with this mysterious "manipulator" and lose his shirt fighting the very site.


This is perfectly consistent with the theory Fireball is a scammer: a few hours after his scam was blown wide open (as by the original post link) his site suddenly goes into "maintenance". This is just a coincidence, right? Wrong. This is trying to cover his tracks.


I took no screenshot.


You either naively or purposefully fail to notice that the site is allowed to CLAIM it's selling the winner's position because of lack of liquidity on the loser side of the deal.

What happens in practice is that the site puts out bait bids. If someone bites and the rate goes against them they lose their investment. If it goes against the site they can just arbitrarily close the position in a few points.

Again, this is so much a bucket shop fraudulent establishment it screams. By the book, bucket shop.


The only thing "it is possible" to do is to donate relatively small sums of money to the dishonest people behind the site. Plain and simple.

You are, either innocently or not, completely misrepresenting what happened. Let's go through the steps.

1. Icbit.se lists a future which massively diverges from the spot in the wrong direction.

2. Every day right before settlement (~30 seconds) time "someone" sells ~50 contracts.

3. MP buys 200 BTC's worth of contracts.

4. Suddenly, that day the sell is not the regular 50 contracts, but ~400.

5. Next day, broker expects the daily sell volume to be back to the usual 50 contracts, on the assumption that it tracks total volume somehow, and next day volume is back to the usual nothing. This turns out to be false, about 1k contracts are sold in spite of it not making economic sense.

6. Same time, broker tries to sell after that "someone". His order is ignored. Ten seconds to settlement time he tries three more times. Order is ignored again. Ten seconds after settlement time, the ~400 put on the book by "someone", which have the magical property of not being sellable into, disappear from the book.

7. This is publicly announced, as per the original post.

8. The site management has a nervous breakdown, sells BTC all the way to 9.5. This fails to wipe MP's position (which was indisputably the intended consequence).

9. Two hours later, after having done a little research/settling down/having a glass of water, site management pushes the price to 11, in the process retroactively changing trader's logs.

10. To try and get out of this they lock MP's broker's account, on the theory that they will be able to argue that the 30 second pre-settlement locking of the book so as to manipulate the price discussed in 6 will be then played down as "something wrong with one account".

This story does not wash. Icbit is a scammy piece of shit. If the operator were here to admit his wrongdoing it would be just plain scammy. The fact that even once caught the operator is trying to destroy the evidence/play the lie game makes it Bitcoin-level scammy.

That's all.

As I said exchange is allowed to close winning guy's position too (if there is not enough liquidity), and so winning and losing guys will simply cancel each other: var. margin amount will be moved from losing guy to winning guy.

So it IS closed circuit. And that means that there IS counter-party risk, i.e. you cannot depend on futures contract being honored.

Of course, exchange might simply choose to put its own money on the table instead of forced liquidation of winning side's position. But it is not required to do that according to contract.


Yep, people need to realize that currently it's possible to wreak havoc with relatively small sums of money.

I edited my post, I was trying to make the point how there could be exposure to the site when there is forced margin selling (and thus not a "closed circuit" as it was claimed", and that argument was easier made from a simple example of just two participants.)



"shitload" is relative.  I don't know the actual trading volume but it appears a single participant with just a 100 or so BTC balance using 1:10 leverage could have been responsible for most of -- if not all of, the bizarre trading.

MPOE, what you are doing here is utterly dishonest.

You, like myself (and several other readers can certainly wittness this) where actively watching what happened last evening. You were reading (and posting to) the forums as everyone else. Yet you deliberately pick some "evidence" and leave out a lot of important contextual information that doesn't fit with your intended message.

So to put all of this back into perspective, I'll give a summary of the events as I observed them.
Please feel free to correct / add further information.

• there were market manipulations going on at clearing time (20:00 UTC) since some weeks now
• there was a long discussion in the ICBIT thread, with several proposals how to prevent such manipulations
• Fireball was aware of that discussion, participated, and indicated that ICBIT will take some measures.
• but Fireball made also clear that he sees a lack of liquidity as the underlying problem
• several people tried to cross the actions of the manipulator(s) more or less successfully now since several days.
• today, Fireball announced that the clearing will be delayed by half an hour or so, because of "maintenance work on the current session"
• shortly after 20:00, there was clearly some DB/maintenance work done, since all the bids / asks were empty, and then they were back unaltered just a minute later.
• then, the manipulator tried for the second time to sell into the bids to shift the market down. He cleared away all bids down to about $10
• then probably someone bought at the lowest ask (10.64), so the "current rate" jumped up back.
• next, the manipulator bought away every bid which was reachable within today's trading range
• again, other market participators bought at 10.64, letting the current rate jump back
• then the manipulator placed a really huge Ask order at 9.70. Shortly thereafter, he placed a smaller, Ask order (60 if I recall correct) at the very lower bound of the possible trade range
• this was commented in the forum
• shortly thereafter, other participants, including myself, started to buy away the lowest Ask
• then the manipulator sprayed the spread with small orders, to conceal where the large order is placed
• the other participants bought away the small camouflage orders.
• this is the time when you took your 1st screen shot
• next, other participants started to buy into the huge blocker Ask order at 9.70, reducing it significantly
• at this point, suddenly the walls at 9.70 and at 10.64 were removed
• the market quickly re-formed around 10.64 and then started moving up
• there was some pushing up and down around 11, until someone placed a 2000 BUZ bid wall at $11
• some sold into that, then the market re-formed above and basically settled down
• somewhere around this time, the closure and mark-to-market happened (I can't reliably recall the precise time point)
• this is when you took your second snapshot
• immediately thereafter, Fireball posted to the forum and indicated, that he forcibly closed some dangerously large short positions

What I see here is that somebody sold a shitload of BTC, filling all available bids.

He then spammed order book with many small asks going down to 9.6. (People say that this is ICBIT's deficiency: number of rows it shows in order book is limited so such spam can hide true market depth.)

9.5 and 9.1 orders you see on this screenshot are mine, BTW.

Apparently "manipulator" tried to squeeze the market right before settling time, but it was delayed this time. And apparently he was out of reserves since market was able to recover up to ~10.80 before settling.

It then shot up to 11.21 in the new session because some of manipulator's positions had to be liquidated and he no longer was able to post ask walls.

So if anything this shows that manipulator and exchange is not the same person, as apparently manipulator didn't get the news that settlement is delayed.

Pretty much the most stupid defense, this "call it FUD" bullshit.

Unsurprising, as MPOE-PR is a competitor

At this point, I would like to point out two little tidbits of evidence:



A little later,



So basically, a bucket shop which artificially showed lower-than-reasonable prices for weeks was called out on the practice. The first reaction was to lower the price even more (down to 9!), the second reaction was to bring them back in line with reality, all this while retroactively changing logs shown to people.

Leaving aside the obvious (ie, that this is why the website model does not work for an exchange, this "you did too!" "we did not!" bs is pointless), you want me to believe that prices just naturally happened to move hours after someone blew the whistle, right?

I find that very hard to do.

Wait, where does it come from? As I understand, clearing price is based on spot price only on settlement day, and before that it is based on futures price.

And to get futures price to move from $11 to $12 you need other market participants.


Have you read this? https://icbit.se/margincall


"Worst case scenario" is why I abandoned the idea of making a traditional futures exchange, it's kinda too messy for my taste.

Just post them here http://www.reddit.com/r/douchebagfoundation/

FWIW I was discussing this kind of a problem (manipulation/margin calls) with Fireball back in summer 2011, before ICBIT was developed, when we considered making a futures exchange together.

Basically I've outlined a situation where a person with big enough pockets can effectively pwn the market getting profit out of nothing. Moreover this situation is very likely on bitcoin exchange: there are people with many bitcoins who do not mind risky endeavors. On the other hand, futures market is rather thin. Also, participants are pseudo-anonymous, so it's feasible to make an alt-account which would lose more money than it had.

So, well, I didn't want to deal with this mess (margin calls, forced liquidation, pissed off customers) and I designed a different kind of contracts ("capped forwards") which kinda serve the same purpose as futures, but do not suffer from same issues. I've also designed some non-traditional kind of futures.

But Fireball made it clear that he wants futures exchange similar to "traditional" ones, and apparently this margin call issue is a norm.

So problems were expected (it would in fact surprise me if nobody tried to squeeze the market), and there is no evidence that exchange operator is responsible for that. This simply make no sense in long term.

As for Mircea Popescu's accusations, they are largely baseless.

1. As I understand, volume displayed by ICBIT is a total for a security, and don't forget that it's leveraged.

2. Cash-settled futures are fine, it is a legitimate financial instrument. Yes, don't have direct effect on spot market, but that's OK.

3. There is no evidence that it was exchange operator who tried to squeeze investors.

4. I don't think that you should rely on being able to send order in last seconds, it could have been just a glitch. It is a web-based trading platform ffs.

So it's just FUD.

(While we are at it, colored coins would allow to organize delivery in USDcoins, for example.)

So let's say the site has a total of two traders,

Trader A deposits 1 BTC and uses some of the margin available to put an order to buy 7 BUZ2 at $11.
Trader B deposits 1 BTC and uses some of the margin available to put an order to sell 7 BUZ2 at $11.

The trade is matched.

Then the next day say the clearing price rises to $12.   []Edit: Which wouldn't happen with just two traders, but just for my example, assume it happens.]
Trader A sees variation margin adjustment of 0.53030303 BTC on A's long position.
Trader B sees variation margin adjustment of -0.53030303 BTC on B's short position.

Trader B no longer has sufficient margin maintenance (under 0.5 BTC holding 7 BUZ2 exceeds the 1:10 minimum), and sees a forced liquidation of the 7 BUZ2 contracts that were sold short.

As we see, there isn't all that much liquidity.  With that forced liquidation, does ICBIT take the hit if the Trader B's BTC balance is insufficient as the result of the liquidation?   So this "closed" system is only closed when there's no forced liquidation.  We just saw that type of liquidation occur today, so we know it happens but we can't know if ICBIT's reserves took a hit or how much they can absorb.


Well, the variation margin adjustments occur daily, so if the spread narrows between the BUZ2 clearing price and spot  (which is now happening after some accounts that were too short BUZ2 were forcefully liquidated), it isn't like traders need to wait until Dec 15th to cash out.

To put that a bit more into perspective.
ICBIT offers futures contracts. http://en.wikipedia.org/wiki/Futures_contract
So they did never claim that users would actually trade through to the spot market. Beacuse of that, the accusation of "bucket shop" is somewhat of a stretch.

Basically, someone buying 1 "BUZ12" contract there agrees to sell $10 at 15.Dec.2012
And the counterparty picking up this bid and thus entering to sell 1 BUZ12 agrees to buy $10 at 15.Dec at the agreed on rate.
Every evening at 20:00 UTC, the platform "marks-to-market" every open contract. That is, it takes the "current" rate within this futures market and adjusts the gain/loss of the two counterparties -- kindof like an advance payment for the expected settlement at 15.12.2012

What makes this setup interesting for the trader is that you can engage into positions with roughly 1:10 leverage.
And contrary to a construction like Bitcoinica (Contract-for-Difference), the financial construction of the platform is way less risky, since it is a closed circuit (and never claimed to be anything else). So basically people's BTC deposits are sitting there and will be re-distributed among the traders on closure. The site doesn't need ever to touch USD, and it doesn't need a trading connection to any stock market. The site gets the trading fees and could possibly engage into short-term lending of BTC. And, of course, they could just run with the coinz.

So yes -- what we really need is these kind of sites, but implemented with multisig+contracts+colored coins.

can't confirm that. Orders and even trades work up until the very moment the clearing happens.
Since about the time of the last bitcoin mini crash, some traders seem to be sitting on an unfortunate short position and are desperately trying to manipulate the closure rate down.

Anyway, during the last days, on numerous instances, I was able to get a trade through only 5 secs before closure and thus managed to cross the manipulator's scheme. I saw other traders do the same thing. Moreover, since it's now known that such manipulatinons take place, you can watch people stage up bids and asks about one hour before the closure and clean up right afterwards. Since the manipulator regularily buys away the whole bid side (or ask side in the case of the oil futures), placing some additional orders just in expectation of this happenging can be quite lucrative.


All these observations don't really dismiss the suspicion, but make it look less likely.
Well -- of course it could also be someone from the platform, but this looks like a quite risky and complicated scheme for a scam.


PS: not affiliated with ICBIT in any way, just trading there

What just happened doesn't exactly match this theory.

The manipulation occurred almost on the hour, but clearing is delayed due to maintenance, so whoever placed the huge sell order sold a lot of futures cheaply, and was unable to maintain the low price.  Too bad for him.    And too bad for the conspiration theory (although the site theoretically could be a scam, I think this thread paints a bleak picture based on too little evidence.)

Still, never place more bitcoins on a site than you can afford to loose - even if it is honest it could be hacked....

And the site operator is anonymous, that is never a good sign....

Well no, it's a site that makes it look like you could get cheap BTC except it simply takes your deposit under the guise of "price has moved against you" (and against the market).


Maybe, but it would be a significant investment of someone's time. This is not an easy thing to do.


Orders work in general, as smickles pointed out, within a second or less. However, right before settlement time (say about a minute) the site partially closes the book: it's closed for customer orders (which simply get ignored) but the site places its own orders, to manipulate the settlement price. These orders are then canceled within a minute or so of the settlement time.


At the rate BTC/USD is going, default may come as early as 12th of December, when the contract ends and they will have to actually settle to market.

Interesting, historically it takes 2-3 months from the first whistle blowing to the official fuckup.

Lets be honest: It's been too quiet for a while in that regard and we need to sweat out those pointless endeavors at some point anyway.

Icbit.se appeared as a ticker first on the Deepbit.net a few weeks after Tradehill folded. Tycho might reconsider advertising their services, as it could be considered an endorsement of their credibility.

When you say it refuses orders, are you saying new orders placed right before clearing (like within a minute or so) are refused, or are you saying during other times that orders can't be entered as well?

Several times well before the daily clearing (e.g., more than ten minutes) I've placed bids at various levels and the ones that were at a price above the eventual clearing price did get filled, so I've not seen where my bids were not processed.  

As far as those trades before the bell occurring, it definitely has been noticed:
 - https://bitcointalksearch.org/topic/m.1325425

So it's basically a site, that tries to seem as if it would actually do trades, but in fact just have your BTC laying around, probably waiting to accumulate enough and then run?

Or did I get it totally wrong?

Thanks for the warning though. As more and more scam sites show up lately (http://www.bitcointalks.com & many more), maybe it's time to start collecting known scam sites on a sticky thread?

Find the whole story here: Icbit.se, the bucket shop.

tl;dr: icbit btc future contract is settled at 22:00 each day. The site puts in fake orders at time of settlement, refuses customer orders at that time and enforces an arbitrarily small price (> 1 dollar under spot) to blow out margins. The site also reports fake volume.

You can either comment there and get answers from Mircea Popescu himself or comment here and I'll do my best to answer.