The results they post are simply too good to be true. They claim to have 10 people working for them out of a Geneva office. Just add up the costs for such an operation, in one of the most expensive cities in the world. Then consider that a 2% difference in exchange rates between different exchanges doesn't mean a 2% profit. Since you need both fiat and BTC to make the arbitrage-trade, your profit is already down to half that of the price-difference. Then you need to keep reserves of both fiat and BTC on all exchanges, while not all reserves are used every day.
The results so far have been consistent with their story and strategy. So far there few days that they return around 2%. Typically the yields are around 1% or less. For example yesterday was a .5% day which I easily beat with my scrypt mining.
I agree that their story and results match decently with the price development on the exchanges. However, this doesn't really mean much. It takes a few minutes every day to look at price charts and make up a nice story.
And while their claimed results vary in vague proportion with the volatility of the BTC price that day, the absolute value of their results are still far too high, because of several factors I've indicated before.
Often price-differences are predominantly one-way: For example, BTC-e is almost always cheaper than Bitstamp. So it is necessary to wire USD from Bitstamp to your bank account and onward to BTC-e, which can easily take over a week. So a large part of the funds is sitting inactive or is locked up in wire-transfers. The price-difference required to achieve the stated results is several times higher than the actual profit figure. And that's not counting all their operational expenses.
Some visibility into their cash/BTC management would be beneficial to showing if the operation were legit or not. However your argument too is also largely hand waving probably based on some negative experiences. While our accounts are divided into BTC/fiat, we don't have any idea how much BTC as been converted into fiat at any given time. BTC is much faster to transfer than fiat. They also talk about using minor exchanges too, but it isn't clear what those are. Too many unknowns to say anything for sure. Lack of visibility though is a valid concern.
I have no experiences, negative or positive with BTC-Arbs. My arguments are based on my own experiences with arbitrage-trading, for which I have developed a simple bot that gets me a small amount of money every now and then.
Regarding your comment on BTC conversion to fiat: This assumption is inconsistent with the reported results. As you correctly mention, BTC is much faster to move around than fiat (~1 hr versus several days to >1 week), so for maximum efficiency, an arbitrage operation would need many times the amount of actual fiat than the fiat-value of their BTC stash. Since it is highly probable that BTC-Arbs gets more BTC investments than fiat investments (arguments: 1) BTC is much easier to invest, the threshold is much lower for BTC transfers. 2) Since the site claims to deal in BTC-trading it will appeal to BTC users primarily.), BTC-Arbs should have a large short position in BTC due to BTC investments being converted to fiat to optimize the efficiency of the operation.
A long term short position in BTC is extremely risky, considering the long term price trend of BTC (which is: to the moon!). Additionally, such a short position would be reflected in the daily results, since a large drop in the BTC price should give a large spike in the profit and vice versa. On days where BTC is going up by a few % or more, negative profits are quite probable. Yet we never see a day of negative profit on BTC-Arbs. It also defeats the entire purpose of arbitrage, risk-free profits, to take such a large position with invested money. In theory, the risk can be hedged by buying BTC/USD call options, but the high volatility of BTC/USD means that the price of these options is rather high and the cost of such a hedging strategy would eliminate most if not all of the profits from the actual arbitrage trades. And even with a perfectly delta-zero hedged position, it is very unlikely to never see a day without losses.
So with the information given by the BTC-Arbs website, the conversion of investments from BTC to fiat can be excluded with extremely high levels of certainty. Which brings me back to my original conclusion that most of the invested BTC is just sitting idle, since the operation (if it were legitimate) would be heavily bottlenecked by fiat movements.
If I were to run an arbitrage operation that took investments, I would not offer the same profits for BTC as I would for fiat. Instead, the profit offered would depend on how much the specific currency is bottlenecking the operation. And to protect the interests of previous investors, I would close down investments in one of the currencies (most likely BTC) if the operation becomes severely fiat bottlenecked and additional BTC investments no longer increase profit potential and instead just dilute the profits.
Yesterday morning I created a support request, asking if btc-arbs would be willing to have a limited presence on BitcoinTalk to address some of the concerns outlined here, or if someone affiliated with the organization would be willing to meet with a trusted member of the community to discuss some of their operations. Adam responded that they would "consider" doing so in the future. This was a much less meaningful answer than I'd hoped for, and I don't think I'm willing to make any more deposits there until something gives. Just thought I'd share.
You can tell Adam that since I live very close to their "monthly office cocktail parties in Geneva", I'd be happy to witness their operation ^^