Hacker News Reader

Comment by dragontamer | Hacker News Reader

BTC is going to have issues as long as it takes a significant period of time to confirm a BTC transfer. I'm seeing reports of people trying to buy coffee, but the BTC transfer takes so long to confirm that the customer is long gone before you know whether or not the BTCs really transferred.

Nothing can stop this either, it is innate to the system. BTC confirmations will take on the average 10minutes, but can last much longer in practice. Attempts to mitigate this issue only result in rebuilding credit-cards / accounts / virtual money systems on top of BTC... which then begs the question... why not just use a Credit Card as always?

Litecoin tries to solve the problem by making the block discovery process much quicker (among other things). These "virtual coins" will need to go through several revisions before they're really useful.

View on HN

Can't there exist a company that accepts Bitcoin transactions on behalf of clients (shops) and guarantees them? Even if a transaction fails, the company coughs up, and to make ends meet they charge a small per-transaction fee.

Even without this, it's quite possible that bad bitcoin transactions will be far less costly to shops on average.

If a shop sells a coffee for $2 paid for by a bad bitcoin transaction, the shop seller has made around a 50c loss (stock mainly).

If a shop sells a coffee for $2 paid for by a later charge backed credit card transaction, the shop seller is charged $15[1], plus the 50c stock loss for a net $15.50 loss! You'd have to have over 30 bad bitcoin transactions to equal that level of loss.

Chargebacks are a money making operation by the credit card companies, at the expense of merchants. Bitcoin helps a lot.

[1] https://stripe.com/help/disputes

Why would a consumer put up with such a company? You've got the economics backwards.

Consumers want features such as reversibility of transactions, money back guarantees, and reward points. Credit Card companies offer _customers_ rights, not the business.

Businesses put up with credit card companies because it brings them additional business.

If you want to create a pro-business "transaction framework", go on ahead. (indeed, the Apple Store does this. All restrictions are strictly for pro-Apple reasons). But consumers are only willing to put up with so many pro-business storefronts.

Some consumers would rather have lower prices and happier merchants than rewards points.

I'd much rather pay for my coffee in digital cash than pay an extra 2%+ to facilitate the transaction. If the product is poor, I just won't buy it again. Same goes for lunch, etc.

Cards that return ~1% back highlight a problem; the transaction fees are too high. $0.10/bitcoin transaction [1] is also steep for minor transactions. If it were more like $0.01, tiny transactions might take off...

Chargeback/escrow services/guarantees are appropriate when the money involved isn't small in relevant units.

[1] https://en.bitcoin.it/wiki/Transaction_fees

FYI, default minimum transaction fee is 0.0001 BTC in current version, or about $0.01.

Whoops! My arithmetic error. I'd computed it, but transcribed 0.0001 to 0.001 BTC in my head last night.

Mea Culpa.

What lower prices? I pay off my credit card every month, and therefore pay nothing ever. Reward points are icing on the cake, all of these are pulled out from Merchant transaction fees.

Have you ever looked at what you're paying for in a credit card? Basically nothing. The merchant pays for everything. (occasionally, the merchant passes back the savings to you... like the occasional gas station that doesn't take credit cards, or a store that only accepts credit cards on purchases over $5)

Consumers pay damn near nothing on credit cards. That is why they use them.

Of course, the merchant fees are included in the prices.

Mtgox is aiming to do this. That way if you both have mtgox accounts its just a database update.

"Confirmed" has two definitions depending on if you apply it to credit cards or Bitcoin:

1 "Confirmed with a certainty that's good enough for values less than 10,000USD" )

Credit cards are good for this basically instantly... unless the register is "taking longer than normal today." It is on the order of seconds. Notice that PayPal doesn't accept credit card payments over 10K AND takes a large percentage. The same is true of Escrow.com, for example. Wires (SEPA in eu) take several hours, but are much more used at these amounts.

With bitcoins, this is true once 90% of the nodes have a transaction in their memory pools which ALSO happens to be on the order of seconds (you can watch propagation of transactions on blockchain.info)

2 "Truly irreversible")

This level of confirmation is true of credit cards after, usually, 180 or 270 days. Whereas with Bitcoin this is true after 4 or 5 confirmations (~40-50 minutes).

Bitcoin's speed gets a bad rep because it is held up to definition #2 and credit cards are only held to a standard of definition #1. This whole non-sense about litecoin solving ANY problems is just that: nonsense. The some members bitcoin community clamored for bitcoin to decrease the confirmation time, but the core devs knew it wasn't necessary. In fact, Charlie Lee, in an effort to satiate these requests created an altcoin, Litecoin, to confirm faster, so that the discussion of bitcoin could regain its track.

TL;DR: Bitcoin is very much confirmed within seconds, just like credit cards. Litecoin's faster confirmation time contributes very little (unneeded) benefit.

What you don't get is #2 is a feature of credit cards. Consumers don't like irreversible spends.

Obviously, if you're a merchant, you want to reach the point of irreversibility sooner. But consumers want 30-day money back guarantees. The credit card companies provide that guarantee if stores do not (and will retract the money from the business's account).

Consumers do not have privileges with BTC, once spent, the money is gone. There is no 3-day window to contest the transaction, there isn't any 30-day money back guarantee.

I agree with you that BTC is more merchant oriented than consumer oriented, but #2 causes serious bugs: credit card transaction limits and additional fees to combat fraud (like those aforementioned). These fees function as an inefficient tax on both merchants and consumers.

Credit card transactions can be reversed weeks/months after they are made. This is necessary to combat fraud, and is part of the reason why fees around 2% exist. In contrast, bitcoin transactions are infeasible to reverse after a few hours.

That said, anything on the order of minutes is obviously too long to wait for certain types of transactions. Litecoin, with a target of 2.5 minutes between confirmations, does nothing to solve this. Bitcoin/P2P cryptocurrencies will need a solution for this problem if they are to be used for in person sales at retailers.

> Credit card transactions can be reversed weeks/months after they are made

Exactly.

The instant approval from the credit card company is a close analogue to a signed transaction. Both prove that the person has the funds available and that they have the authority to authorize the transaction.

The 40-minute confirmation delay is roughly equivalent to the 6 months or so that it takes before credit card transactions become irreversible. At approximately one block mined every 10 minutes, after 40 minutes there would a 4-block-long chain including the transaction. The idea that someone with the resources necessary to discover 5 blocks before everyone else in the network could discover one block would bring them to bear to steal a cup of coffee is insane.

If I'm selling convenience goods - say, under $100 - I'm fine with zero-confirmation transactions. If we're talking $100-$1000, I'd wait until the network confirms it once. If I'm selling you a house, then I'll buy you lunch while we wait an hour or so for 6 confirmations.

As a merchant, I would certainly prefer to suffer the loss of a few small transactions than pay 2 to 3 percent of every transaction to a payment processor. I imagine losses due to double-spends would be comparable to those of accepting counterfeit currency--miniscule, due to the high-risk of perpertrating fraud for such a small reward.

Grocery shopping is regularly over $100. A single network confirmation typically takes 10 minutes, but can take as long as an hour. (Block 152218 took 1 hour 40 minutes). Of course, there are plenty of examples where a blockchain was calculated in seconds... but such is the random nature of waiting for blockchain confirmations.

A quick back-of-the-envelope calculation says that 30% of the time, blockchains will take 20 or more minutes... and 5% of the time, blockchains will take 30 or more minutes. Are your customers willing to wait 30 minutes 5% of the time while doing their groceries?

Hell no. Waiting for even a _single_ blockchain takes too damn long, even if you reserve it for substantial amounts of money like $100+ transactions.

There is the concept of green addresses (https://en.bitcoin.it/wiki/Green_address) to solve this. Basically if customers use a wallet that the merchant trusts not to commit a double spend then they can safely accept a zero confirmation transaction. MtGox and Bridgewalker (https://www.bridgewalkerapp.com) support this so if you trust them and your customer uses them then you can accept the transaction immediately.

Of course for something like a cup of coffee or even a purchase for $80 worth of clothes I think most companies would be fine with accepting a zero confirmation transaction and taking the risk as it will be far less of a cost then shoplifting and will be less than the 2% they currently pay on every credit card transaction. If you are buying a $1000 laptop then maybe they make you hang out while they explain the features and wait for a single confirmation.

And now you've got "Credit Cards". Merchants won't accept your business unless you've got a trusted "green address", and you won't shop with merchants who don't support your "green address".

Again, attempts to solve this problem lead to a situation similar to credit cards... except worse (inherent to the extreme volatility and instability currently experienced in the BTC market).

Also, lol Mt. Gox. That company that you can't switch to USD every couple of weeks because they're not doing everything they can to keep their bank accounts open. Mt. Gox's instability is one of the worst things happening to the BTC market right now. If you give away your BTCs and store them on someone else's server, you're doing BTC wrong. Its no longer a decentralized system, but instead a new centralized system of currency.

Green addresses, specifically, are not a great solution because they thoroughly disrupt Bitcoin's privacy and frangibility model by making that source identification highly public.

But fortunately you can do something which is almost exactly equal: Have the trusted party just hand a buyer a signature of the transaction ID, which they hand to their vendor.

No one but the involved parties see the data... and it allows other models like ones where you retain control of your own funds and the signing party just agrees to take the risk the you cheat (and the cost of tracking you down if you do).

You can just wait until its in the memory pool of some trusted nodes. Double spend risk on coffee isn't worth waiting for a block.

What about a $100+ grocery transaction? Are people willing to wait around 10 minutes in front of the cash register to wait for a single blockchain confirmation?

There isn't any real risk of double spend attacks. The effort it takes to make a double spend happen requires far more money than that (and no, I'm not taking about a 51% attack), and is not reliable in any way.

A transaction is broadcasted to most of the network, and certainly to the main pools within seconds. So once you see a transaction in the pool of trusted miners, you can assume it's okay. The only "real" possibility of it being a double spend is if they appear on the network at the same time, but is visible when the second transaction uses money that the previous transaction have spend. At that point it's equivalent to the credit card failing from an interruption in the connection - you just try again (the oldest win).

If it is a coffee, just accept it at the first block, that should only be a few minutes. For the rare occurance of an orphaned block, it is still less of a loss than the normal credit card processing fees for a typical day at such a place.

First block may take anywhere between 10 minutes to an hour. No one knows when the next block will be discovered.

Are you just going to hold up the line while you wait for the next block? Or are you going to just accept zero-block transactions?

The Bitcoin protocol is designed to make it average 10 minutes between each block, so on average the wait is 5 minutes. And if I was Starbucks selling a coffee, I probably would be willing to accept zero-block transactions, especially if the user has one of those rewards cards. Just re-charge them next time around. Only very rarely will they need to eat the cost of a $4 coffee, in exchange for not paying 1%-2% on every single purchase.

> Attempts to mitigate this issue only result in rebuilding credit-cards / accounts / virtual money systems on top of BTC... which then begs the question... why not just use a Credit Card as always?

Because you can have the benefits of both worlds.

Any modern store would have an API that allows you to schedule a pickup through a mobile app and use Bitcoins as long as you schedule it 30 minutes to an hour before you arrive. What is this mentality Q2-2013?

I do not think you sufficiently understand Bitcoin or double spend attacks work.

You should read this section of the Bitcoin FAQ: https://en.bitcoin.it/wiki/FAQ#Do_you_have_to_wait_until_my_...

Reply on news.ycombinator.com