The links to TX1, TX2 and TX3 are in my message.

It may be hard to locate a ~350 BTC transfer to BitPay without the exact date.  The source address may not be identifiable, it may be Bitstamp like TX3, possibly in installments.

PS. Perhaps the transaction TX3 -- from a Bitstamp address to a couple dozen outputs -- may be Bistamp itself, or some intermediary, making many withdrawals for many clients, one of them being the guy who needed the 1013 BTC for the payment via BitPay.

So far, this is my best guess for the transaction that sent to BitPay the payment for the house:

(TX1) 2014-04-28 22:08:49
http://www.walletexplorer.com/txid/470dd9660d62b60ce8451c22392ba119b1c0b97d0927d6d5ef3ee0c2da043585
inputs:
1BkuugvWi7Gjmwf84rgQ94Qucohs2fSDip    ~956  BTC
1Gmn7to4T8ZgBgbnRyS83NDr1vNdTqhKoc   ~1013  BTC

outputs:
1DVkcgeJoPjzUnPv4gBggXCierQS8tJWpB (BitPay.com) ~1061 BTC
1BoykhKtyibN12Fr2cTMvEnP1EN52LzHK2               ~909 BTC

The first input of TX1 comes essentially from the output of this transaction:

(TX2) 2014-04-25 22:54:14
http://www.walletexplorer.com/txid/671b11d82f38f17a7e3c4e433518cd9f255897932c333bdcc40cfae831d84bd6

non-dust inputs:
12JiS4KLFv8DPxeshVnd6mmpSksqr8ZraM  200 BTC
148hAbyhcvpn9e6pjUYp69vSrC4MW86XyU  200 BTC
1GGXdkZ6W5tQsRMn2EpdVm3NNH4kL7FHbN  200 BTC
1MXsW4PQ3BmG2iEiP1K8ji4bnHsvGLzVnT  200 BTC
19gr6XbnyiaJDqMyKm4WMdKGb4XzvUPNeT ~156 BTC

non-dust outputs:
1BkuugvWi7Gjmwf84rgQ94Qucohs2fSDip ~956 BTC

Note the exact 200 BTC inputs.  Are those characteristic of something?

The second input of TX1 comes essentially from the output of this transaction:

(TX3) 2014-04-24 17:46:45
http://www.walletexplorer.com/txid/a6422b8787cf702bd57077041ba3ee63ddc8d372f280bf8df37676dc82b95077

only input:
1Drt3c8pSdrkyjuBiwVcSSixZwQtMZ3Tew (Bitstamp.net) 1588 BTC

outputs:
1Gmn7to4T8ZgBgbnRyS83NDr1vNdTqhKoc ~1013 BTC
16HyaukaMhbZ6idtTkM91Q13nkpDsr5Noc  ~485 BTC
1FaZztd9BrFu5AF4FfbVeW6YwQxAwBxfXQ    45 BTC
1DgH36qLGLpiJHKnaDvYv9PoQ6nZ3rPqKL    15 BTC
plus a couple dozen small outputs, including
1m9iZsENpnUYN9DY8M9CUBtZiMMDhkup2 (AgoraMarket) 0.06264615 BTC
1227BCzzzY9kBEjRUk3evo8FBggaYp52Zi (SilkRoad2) 0.408099 BTC

Thus, on 2014-04-24 17:46:45 (TX3) someone withdrew 1588 BTC (exact) from bitstamp, paid a couple dozen various small purchases, then sent ~1013 BTC to the 1Gmn7 address, ~485 BTC to 16Hyau.

On 2014-04-25 22:54:14 (TX2), someone merged four 200 BTC packets and a 156 BTC one into a ~956 BTC packet at 1Bkuug.

On 2014-04-28 22:08:49 (TX1), someone merged the 1Gmn7 and 1Bkuug packets, then split the result again into ~1061 BTC, sent to Bitpay, and ~909  BTC into another address.

If it was the same person, perhaps on 2014-04-24 he had prepared the ~1013 BTC in 1Gmn7 for the purchase, but on 2014-04-28, when he made the payment, he had to increase the amount to ~1061 BTC, so he took 47 BTC from 1Bkuug.

Makes sense?

Lots of complaints on reddit about lost/stolen bitcoins, many from blockchain.info wallets:
https://np.reddit.com/r/Bitcoin/comments/2nkias/this_is_a_list_of_rbitcoin_users_who_had_their/

Bah.  It is easier to use the Banach-Tarski Paradox

Note that it can be applied to each resulting ball to get 4, 8, 16,  solid balls, just like the original one.

That method may not work with gold, because of those gritty atoms; but since bitcoin is a purely mathematical precious metal, perhaps it can be used to pull a double-spend.

Well, at this point he is just one of the 7 billion suspects -- all equally likely, IMHO. 

So,  do you people think this could be the deposit to BitPay?

The date of sale could mean the date when BitPay deposited the USD in the owners account, or the date when they filed the form.

This one is about the right amount, two days before:

2014-04-28 22:08:49 +1060.776 BTC
http://www.walletexplorer.com/txid/470dd9660d62b60ce8451c22392ba119b1c0b97d0927d6d5ef3ee0c2da043585

This is another big one, one day before:

2014-04-29 18:26:19 +6575.5696 BTC
http://www.walletexplorer.com/txid/3ebeec72e64a869799dd62cee89b42a96452efb3f52f09aa2bbcd94753e9c79e

The price paid for the house should be about 1000 BTC, correct?

Well, some large deposits before the posted sale date:

2014-04-10 12:08:40 +325.8392 BTC
http://www.walletexplorer.com/txid/aa775ac6b01dc26b56dd614ad18cf6e3d62028d489d5c4afcc6f7e63385f09de

2014-04-07 13:42:39 +6879.6572 BTC
http://www.walletexplorer.com/txid/99be40389113434ccabd587a6d4e780379e60b27fe458e3bf955a93a9c5829b4

He may have made many deposits smaller than ~200 BTC.

We do not know the date, hence we do not know the amount in BTC (unless it was reported in some other article).

Did he use BitPay? The article is not clear on that, but since the previous owners were not very confident, presumably they wanted dollars. 

The authorities could easily get the amount, and perhaps the address, by subpoening the previous owners and/or BitPay.  Or they could ask Josh again.

Note that the article says that his two cars were paid with bitcoin, too.

#ASKFTC
#ASKIRS

If you un-ignore me just this once, here is what happened last time:

[ Click on image for a full-resolution version.  If the image is truncated, blame the forum (non-)admins... ]

(A) 06-12 Official announcement of ~30'000 BTC auction by USMS
(B) 06-27 (Fri) Auction carried out.
(C) 06-30 (Mon) USMS announces that all lots were taken by a single bidder.

(D) 11-17 USMS announces new 50'000 BTC auction.

Note that there was no drop after he announcement this time.  I do not think that history will repeat itself.

Now you can ignore me again.

That 2010 bug was a trivial coding mistake.  If one issued a transaction request with two very large outputs, the sum would overflow the number of bits allocated for the total, yielding a negative value; and the software then would not realize that the total outputs exceeded (by a huge margin) the amount available in the address.   That bug was quickly fixed at the time, but required a rewind of the blockchain to remove the absurd transaction.

The bug that Mark blamed for the MtGOX loss was more complicated. The protocol itself was not broken, but allowed certain cosmetic changes to a trasaction request that would not invalidate its signature.  By manipulating those cosmetic details, the hacker could cause the MtGOX software to think that a client's withdrawal request did not go through, when in fact it had.  The software would then credit back the coins to the internal client account, and try to withdraw again.  This was possible only because the MtGOX software was carelessly coded and did not ignore changes in those those cosmetic details.

I do not think that this explanation makes sense.  Perhaps some hacker tried this attack, Mark noticed it, and  thought of blaming the entire loss on it, so that people would not look for the real cause.  Or perhaps he noticed the weak spot in the code, but, instead of fixing, it he decided to steal all the coins and then blame it on this bug.  Or someone else noticed the weak spot, and convinced or forced Mark to do that.

Should we infer that you could top up a half-full glass with milkshake, but not with wiskey and vodka?

You think no one will be interested in my PneumoultramicroscopicsilicovolcanoconiosisCoin (XPUMSIVCC) then?

How can we tell, if you don't post a photo?

It depends; which half would you want to top up?

You realize that there is a fantastic get-rich opportunity there?  Those lizard-shaped pasta dumplings -- can be fried or served in soup, like wontons, or al sugo like ravioli...

By the way, is it known which blockchain address those bitcoins came out of?

So, when someone says "totally empty", does that mean 0% of 100%, or 100% of 0%, or what?

It seems that he paid with bitcoin, so he would not have needed access to the BFL bank accounts:

World Now, 2014-05-20
Bitcoin used to buy $500,000 Kansas home
http://raycomgroup.worldnow.com/story/25568078/bitcoin-used-to-buy-500000-olathe-home