I am not an accountant or a lawyer, but I'll share my thoughts...

A gambling win of bitcoins has two tax-related effects:
1. The income realized in that moment (in USD terms of the current market value)
2. The current market value becomes the cost basis for those coins

A gambling loss of bitcoins may also have two tax-related effects:
1. The loss realized in the moment (in USD terms of the current market value) which can be deducted against gambling wins
2. A possible effect on cost basis (which I will now explain below)

Buying and selling/spending bitcoins is pretty straightforward from a tax perspective.  Cost basis is set when bitcoins are bought (or won gambling as mentioned above) and a capital gain (or loss) is realized when they are sold or spent.  FIFO (first in, first out) is the typical method used, so that if you sell/spend a portion of your bitcoins, the earliest bitcoins are the ones you are using and thus it is their cost basis which should be used for calculating capital gains.

However, where it gets tricky is with gambling losses.  If you have 100 bitcoins, 50 that you bought for $400/coin and 50 that you bought for $1000/coin, and you then gamble and lose 10 bitcoins, following the same FIFO rule, the 10 that you lose are the first 10 that you bought, which in this case would be coins with a cost basis of $400/coin.  This doesn't change the loss realized in the moment in USD terms to be deducted against gambling winnings that year, but it does have a potentially large effect on the cost basis of the remaining coins.  Following FIFO for the loss, you would be left with 40 coins at $400/ea cost basis and 50 coins at $1000/ea cost basis.

If, however, for some reason, you are supposed to have act as if you "lost" 10 bitcoins that are the higher $1000/ea cost basis, you would be left with 50 coins at $400/ea and 40 coins at $1000/ea. 

Thanks for the explanation, fairlay.

For the record, I definitely wasn't accusing you of dishonest practices.  I was just saying that a case of a losing bet that is canceled due to being past the closing time is much more illustrative of a fair policy than a winning bet that is canceled.  I haven't had a single wager canceled myself.

Similarly, after some time using a book a user will somewhat frequently encounter its "odds have changed" notice when trying to place wagers.  It pretty quickly becomes apparent if a book is only showing that notice when the odds move in the user's favor.  I suppose it's arguable that it is not dishonest for a book to do that, but that is blatantly unfair in my mind.  Bitbook, which in its heyday was a top BTC book, did this and did not respond when it was pointed out publicly.

Canceling a bet that was placed after the game started seems like a reasonable thing to do.  To me, the BIG question is, would that bet have been canceled had it been a loser and not a winner?

I just want to say that cashouts have been much, much faster lately.  Thanks for improving in that area, Fairlay team.

I am the one who posted the American odds userscript.  Please PM me if you would like me to help debug and fix the problem.

Can we please have an update?

My pending cashout and pending deposit have both been resolved.  Thank you.

I have a deposit in the same state right now.  It doesn't show unless I click "Check for missing transactions", then it shows up as confirmed at the bottom but is gone if I refresh and is not credited to my account.

I am the one who wrote the userscript and posted it on your forum.  Instead of having to open up developer tools and pasting the code, anyone can install Tampermonkey extension in Chrome and the @include statement at the top of the script will make the code execute only on fairlay.

Notes:
1. The script inverts the odds on the Lay side to show what you will pay if you take that side.
2. The script currently only executes on the event listing pages; once you click a specific event they will disappear.  This is good enough for me, but I might enhance the script in the future when I'm bored if fairlay doesn't implement their own American odds option in the meantime.

If Gox fixes BTC withdrawals, there will also be a ton of buying pressure from all of the people who converted to fiat (which drove the price this low in the first place) wanting to convert back to BTC so they can transfer them out, perhaps to sell for a profit, if it is indeed profitable for them.

Essentially, Coinbase is currently operating like a sportsbook does; trying to get equal action on both sides and collecting their vig.

I assume they gradually raised their quoted prices (importantly both buy and sell prices) until the sell demand started to overtake the buy demand.  Combined with a spread of almost $20 between buy and sell prices, they should still be making plenty of money without exposing themselves to very much risk.

There is nothing preventing arbitrage from happening.  In fact, it almost certainly is happening right now.  It will only last until they wire more money in tomorrow and the price works its way back to the normal market price.  It does hint toward significant US buying demand right now, which suggests an impending rise in the morning.

Sorry if I missed it, but was the fact that Coinbase ran out of fiat on bitstamp talked about in here?  Extremely bullish...

http://www.reddit.com/r/CoinBase/comments/1xek7t/why_is_coinbase_buy_price_so_much_higher_than/

Withdrawals have been very slow lately, as if they are all requiring manual review.  I'm not sure what their issue is currently, but I don't think they are running away with our coins.  I'm sure there will be an explanation soon enough.

justusranvier brought up a great point in that USD purchasing power has been following an exponential decay model.  Doesn't that suggest the growth should be super exponential, at least for some phase?

https://bitcointalk.org/index.php?topic=178336.msg4210004#msg4210004

Let me start off by saying I think you're doing a great job running bitbook and your customer support seems very good.

This odds change situation is very bothersome.  To me, it is a dishonest business practice.  I understand you trying to justify it as avoiding the inconvenience of rejecting a bet and forcing a refresh, but I think I speak for the majority when I say I'd rather refresh and get better odds.  Taking an edge when it is in your favor and rejecting it when it is not, unbeknownst to the user, isn't much different from secretly stealing funds.

You seem like you care a lot about customer service and you are running the site with financial responsibility, which is great to see.  In this industry, gaining the trust and confidence of your customers is paramount.  I hope you will be open-minded on this issue and see that 100% transparency and a fair set of rules are extremely important.  This is not meant to be a personal attack or bitbook bashing.  I just want to bring this issue to light further because I feel it is so important.

I don't think there's anything you can call these "high-risk" transaction cancellations anything other than blatantly dishonest business.  The company is getting an unfair advantage during 3-4 day swings in price where they can arbitrarily decide to cancel the ones that work out in the consumer's favor, while keeping every transaction that is "high-risk" to the consumer.  Yes, they are growing and seem like the most reputable source to buy BTC in the US, but to me, behavior like this says a lot and raises a major red flag.  I'd really like to see this issue get more exposure in order to correct this bad business practice.

It appears bitbook is doing something similar with odds changes.  It's really a shame to see some of these bitcoin businesses trying to unknowingly squeeze extra money out of their customers rather than simply taking what is theirs for the honestly great service that they provide.