[ANN] LocalBitcoins.com - a location-based bitcoin to cash marketplace

[hivewallet]

We've added LocalBitcoins and ZipZap locations to CoinMap. The code is here:
https://github.com/grabhive/coinmap

Both services have an API for searching so right now we have to query each of them. LocalBitcoins provides a list of country codes based on their user location, we use this list to get the codes. Than we convert them to the country name and query the API for each country (you can't use the country code for geolocation with Google e.g. DE == Delaware, but this is the country code for Germany - Deutschland). For ZipZap we use a list of the 50 largest cities in the USA. We are not sure if all data is mapped correctly, but the points are there.

[FullLife]

Would love to see the results within a certain radius (the same city? the same state?) sorted by PRICE, or even better, the ability to sort by PRICE AND FEEDBACK.
So I can weed out those w/o 100% positive rep and THEN decide on the right price.

The way it currently is, I dunno WHAT the criteria are.

Please don't discriminate against individuals who don't have a perfect 100% feedback.  Not having 100% feedback is not a big deal.  All it takes is one malicious individual to initiate a trade, cancel and leave negative feedback to ruin a perfect 100%.  Once this happens, it takes a LOT of transactions to get back to 100%.  What makes it even worse is that not everyone will leave you feedback.

I should know because this is exactly what happened to me.

[giszmo]

Would love to see the results within a certain radius (the same city? the same state?) sorted by PRICE, or even better, the ability to sort by PRICE AND FEEDBACK.
So I can weed out those w/o 100% positive rep and THEN decide on the right price.

The way it currently is, I dunno WHAT the criteria are.

Please don't discriminate against individuals who don't have a perfect 100% feedback.  Not having 100% feedback is not a big deal.  All it takes is one malicious individual to initiate a trade, cancel and leave negative feedback to ruin a perfect 100%.  Once this happens, it takes a LOT of transactions to get back to 100%.  What makes it even worse is that not everyone will leave you feedback.

I should know because this is exactly what happened to me.

A good price and a good rating make a trade more likely. Both have no-go-thresholds and beyond that, they should be weighted in a way.

Example:
Thresholds:
* 100km (if it has to be in person)
* 10% bad ratings
* 15% fees compared to bitstamp price

distance score = (100km - distance) / x1 km
rating score = (1 - bad_ratings / ratings) * x2
activity score = ratins * x3
fee score = price / bitstamp price * x4 [ distinguish between sell and buy]

Sort by distance score + rating score + activity score + fee score.

I would not waste one second on an offer by somebody who has 100% negative or who offers BTC at price the price of somebody with a comparable volume.
Somebody with a good price would deserve to have the negative feedback read and somebody with a good rating deserves to have slightly worse prices to be considered. Maybe he's otherwise worth it.

Setting thresholds would be an advanced option. Setting x1-4 would most likely be subject to optimization. Tolerance for different aspects might highly depend on the region and density of the bitcoin economy.

[2weiX]

Would love to see the results within a certain radius (the same city? the same state?) sorted by PRICE, or even better, the ability to sort by PRICE AND FEEDBACK.
So I can weed out those w/o 100% positive rep and THEN decide on the right price.

The way it currently is, I dunno WHAT the criteria are.

Please don't discriminate against individuals who don't have a perfect 100% feedback.  Not having 100% feedback is not a big deal.  All it takes is one malicious individual to initiate a trade, cancel and leave negative feedback to ruin a perfect 100%.  Once this happens, it takes a LOT of transactions to get back to 100%.  What makes it even worse is that not everyone will leave you feedback.

I should know because this is exactly what happened to me.

A good price and a good rating make a trade more likely. Both have no-go-thresholds and beyond that, they should be weighted in a way.

Example:
Thresholds:
* 100km (if it has to be in person)
* 10% bad ratings
* 15% fees compared to bitstamp price

distance score = (100km - distance) / x1 km
rating score = (1 - bad_ratings / ratings) * x2
activity score = ratins * x3
fee score = price / bitstamp price * x4 [ distinguish between sell and buy]

Sort by distance score + rating score + activity score + fee score.

I would not waste one second on an offer by somebody who has 100% negative or who offers BTC at price the price of somebody with a comparable volume.
Somebody with a good price would deserve to have the negative feedback read and somebody with a good rating deserves to have slightly worse prices to be considered. Maybe he's otherwise worth it.

Setting thresholds would be an advanced option. Setting x1-4 would most likely be subject to optimization. Tolerance for different aspects might highly depend on the region and density of the bitcoin economy.

That's all so very specific...
Just have the colums sortable [asc/desc]: Price / Distance / Rating
With the option [ ] hide users with 0 feedback.

easy as pie! the rest is up to the user.

[giszmo]

Would love to see the results within a certain radius (the same city? the same state?) sorted by PRICE, or even better, the ability to sort by PRICE AND FEEDBACK.
So I can weed out those w/o 100% positive rep and THEN decide on the right price.

The way it currently is, I dunno WHAT the criteria are.

Please don't discriminate against individuals who don't have a perfect 100% feedback.  Not having 100% feedback is not a big deal.  All it takes is one malicious individual to initiate a trade, cancel and leave negative feedback to ruin a perfect 100%.  Once this happens, it takes a LOT of transactions to get back to 100%.  What makes it even worse is that not everyone will leave you feedback.

I should know because this is exactly what happened to me.

A good price and a good rating make a trade more likely. Both have no-go-thresholds and beyond that, they should be weighted in a way.

Example:
Thresholds:
* 100km (if it has to be in person)
* 10% bad ratings
* 15% fees compared to bitstamp price

distance score = (100km - distance) / x1 km
rating score = (1 - bad_ratings / ratings) * x2
activity score = ratins * x3
fee score = price / bitstamp price * x4 [ distinguish between sell and buy]

Sort by distance score + rating score + activity score + fee score.

I would not waste one second on an offer by somebody who has 100% negative or who offers BTC at price the price of somebody with a comparable volume.
Somebody with a good price would deserve to have the negative feedback read and somebody with a good rating deserves to have slightly worse prices to be considered. Maybe he's otherwise worth it.

Setting thresholds would be an advanced option. Setting x1-4 would most likely be subject to optimization. Tolerance for different aspects might highly depend on the region and density of the bitcoin economy.

That's all so very specific...
Just have the colums sortable [asc/desc]: Price / Distance / Rating
With the option [ ] hide users with 0 feedback.

easy as pie! the rest is up to the user.

For the user it should be easy as pie but sorting by these columns individually with the option to even reverse the sorting is also too much. Having these options would be the poor man's solution, giving the geeks more control without much work for LB's devs.

If this was my service though I would love to apply machine learning to it. Take above parameters as the dimensions and the positively rated trades that happen as the outcome. Messages going back and forth would be yet another hidden dimension. LB should have a large training set and they can even sort by what they are more interested in, their profit, which wouldn't be too badly aligned with the customers needs, neither (with my trades though it would interfere as I don't generate profits).
Machine learning is actually quite accessible through a vast array of libraries that don't require you to actually understand much of the field but the outcome would be customer satisfaction based on past customer satisfaction.
Localbitcoins is currently searching for a backend programmer. This could be just one such thing this guy could work on

[hivewallet]

Introducing the LocalBitcoins app for Hive.

Please note: as with Hive itself, this should be handled with great care and at the moment we do not recommend it for any kind of day-to-day use.

The latest releases will be made available here:
https://github.com/grabhive/hiveapp-localbitcoins/releases

Issues can be tracked here:
https://github.com/grabhive/hiveapp-localbitcoins/issues

[giszmo]

Now that I started using wire trades, I ran into the problem that buyers are supposed to instantly get my bitcoins into escrow so using a simple formula doesn't cut it anymore if I want to protect myself against some wild swings.

I studied math but still don't feel too comfortable using the powerful formula input not knowing what could possible go wrong, so I wonder how less savvy users are supposed to use this.

I would suggest to come up with some rather tight wild swing protections.

For my online sell I currently use:

max(bitstampusd * 1.1, bitstampusd_high * 0.95 ) * USD_in_CLP

It means: take the current most accurate estimate of a fair price (bitstampusd), apply my 10% fee and display it in Chilean Pesos. In case of a falling/crashing price, never show a price lower than 95% of 24h high (bitstampusd_high).

I would feel much more comfortable if this was all wrapped in some form like this:
rate to use: dropdown(average, mtgox, bitstamp, ...)
fee in %: __5__
maximum price drop per day in %: __10__
maximum price drop per week in %: __20__

[justusranvier]

Now that I started using wire trades, I ran into the problem that buyers are supposed to instantly get my bitcoins into escrow so using a simple formula doesn't cut it anymore if I want to protect myself against some wild swings.

http://localbitcoins.blogspot.com/2013/06/this-is-first-in-series-of-posts.html

[Walter Rothbard]

Now that I started using wire trades, I ran into the problem that buyers are supposed to instantly get my bitcoins into escrow so using a simple formula doesn't cut it anymore if I want to protect myself against some wild swings.

http://localbitcoins.blogspot.com/2013/06/this-is-first-in-series-of-posts.html

Is there a link to the rest of that series?

[justusranvier]

Is there a link to the rest of that series?

Actually... I never wrote any more.

Probably should edit that post...

[Walter Rothbard]

Is there a link to the rest of that series?

Actually... I never wrote any more.

Probably should edit that post...

No, go write more!

[giszmo]

Now that I started using wire trades, I ran into the problem that buyers are supposed to instantly get my bitcoins into escrow so using a simple formula doesn't cut it anymore if I want to protect myself against some wild swings.

http://localbitcoins.blogspot.com/2013/06/this-is-first-in-series-of-posts.html

Thank you but that's not really my point. The thousands of users should not have to dig into these details if the developers of LB can just give them maybe exactly what you came up with. Give the users to tweak some parameters and if that's not enough, let them play but give them some protection against some too stupid price movements. Some "pause advertisement on wild swings" checkbox would maybe be the easiest to understand.

[justusranvier]

Now that I started using wire trades, I ran into the problem that buyers are supposed to instantly get my bitcoins into escrow so using a simple formula doesn't cut it anymore if I want to protect myself against some wild swings.

http://localbitcoins.blogspot.com/2013/06/this-is-first-in-series-of-posts.html

Thank you but that's not really my point. The thousands of users should not have to dig into these details if the developers of LB can just give them maybe exactly what you came up with. Give the users to tweak some parameters and if that's not enough, let them play but give them some protection against some too stupid price movements. Some "pause advertisement on wild swings" checkbox would maybe be the easiest to understand.

They  seem  pretty much overwhelmed just keeping the site operational.

Your best bet is honestly to order to code it for them.

[Walter Rothbard]

Now that I started using wire trades, I ran into the problem that buyers are supposed to instantly get my bitcoins into escrow so using a simple formula doesn't cut it anymore if I want to protect myself against some wild swings.

http://localbitcoins.blogspot.com/2013/06/this-is-first-in-series-of-posts.html

Thank you but that's not really my point. The thousands of users should not have to dig into these details if the developers of LB can just give them maybe exactly what you came up with. Give the users to tweak some parameters and if that's not enough, let them play but give them some protection against some too stupid price movements. Some "pause advertisement on wild swings" checkbox would maybe be the easiest to understand.

I get the feeling that if that many people were all using pre-defined formulas like that, the system would be easy to game.

[2weiX]

if you don't want to be gamed, set a minimum price to sell at or a max price to buy at.

Code:

min(mtgoxUSD_ask*1.05,350)

sells at mtgox+5% as long or 350$, whichever is higher.

Code:

max(btcdeEUR_bid*0.95,230)

buys 5% below bitcoin.de bid but no higher than 230€.



Edith: Just read the blogpost, really good stuff in there. Probably still helps to set a MIN and MAX integer, just in case.

[abracadabra]

Careful... just received a fishing email with an offer to buy bitcoins on localbitcoins.

Was wondering how I got an offer since I didn't have any active ads running.

The link links to llocalbitcoins.com

[deadweasel]

Careful... just received a fishing email with an offer to buy bitcoins on localbitcoins.

Was wondering how I got an offer since I didn't have any active ads running.

The link links to llocalbitcoins.com

llocalbitcoins.com is now 404

[herzmeister]

http://www.reddit.com/r/Bitcoin/comments/1qjse3/psalocalbitcoins_phishing_attempt/

[justusranvier]

Probably still helps to set a MIN and MAX integer, just in case.

Price is moving so quickly that those numbers quickly become obsolete.

[Newar]

Probably still helps to set a MIN and MAX integer, just in case.

Price is moving so quickly that those numbers quickly become obsolete.

An added 24hr average sort of "protects" against quick drops.

2weiX has recommended a reserve price earlier in this thread and that works well for me ever since. Thanks again!