Bitcoin Forum

Kill it!

The bot that's spamming <0.00 BTC trades? Yeah, very annoying. Producing loads of noise on graphs =/

I hope they keep it running... It provides some interesting data if you can decipher what it is indicating.

What is it indicating according to you?

If you compare volume weighted moving average versus simple moving average, this will cause the simple moving average to be more indicative of the current spread prices whereas the volume weighted average shows where most of the heavy trading occurred.

How does the bot spamming bogus trades worth less than a fractional penny each add value to an analysis.  It would be like saying if you turn your amplifier up so loud it adds distortion to the music you can appreciate it better. 

I have no idea why Mt. Gox doesn't implement a minimum trade size.  It wouldn't need to be large even 1 BTC would be fine.

So, you don't stop that bot, you just make it thinner, thanks!  ::)

You can ignore those trades in your analysis you you want.  It is very easy to filter out this type of "noise" since it is very ordered.  This is nothing like adding distortion to music.  It shows the available prices over time, which indicates where people are willing to buy/sell at any given time.  If the prices available are on average much higher than the volume weighted price than there are buyers buying at the low end of the volatility.  If the prices available are much lower than the volume weighted price than there are sellers selling at the high end of the volatility.  Nobody would be throwing money at the market like that if there wan't something to be gained from it.  Even it the amounts are tiny, playing the spread in reverse plus fees eats most of each trade, and these are happening every few seconds.

I understand you can ignore it but ignoring it adds nothing.  Basing your analysis on such an obviously manipulated trade seems to not provide any value either.  At best you ignore it and are not better than if the bot didn't exist.

I think you are vastly overestimating the cost of that stupid bot.  The transaction size is 0.0000001 BTC.  Lets say it trades once per second 0.0000001 * 60 * 60 * 24 * 365 * 0.6% = 0.019 BTC per year in trading fees.   How much trading losses is the bot racking up?  I don't know but lets say it is 10% of principal.  That is <1 BTC per year.

I never said there isn't something to be gained from it.  It distorts the charts and most people won't filter that out.   That misinformation has to be worth 1 BTC per year right?

I would sugget MtGox to start taking minimal transaction fee, like 0.001 BTC, regardless of the size of the transaction. Will not hurt anyone .. except that stupid bot(s).

If it is that cheap, and you want rid of it, the best way would be for 10 or so people to run the same bot.  MtGox would quickly adapt to prevent this behavior.

Someone has to be losing money on this. It buys high and sells low, + fees.

It just ate a piece of my bid... that piece was 0.00003, so my calculation (assuming every 10 seconds, 2% spread, 0.5% fee, and consistent order size)

0.00003*6*60*24*365*0.025 = 2.3652 BTC per year

lol ok!
On clark moody, these register as .0001. Whereas I can only place an order of .01 minimum.
It's spam, and it bogs down the system. My browser has *almost* as much CPU utilization as the miners core.

Sure they may be losing money but even if the spread is 10% the volume is so tiny it is a token amount. 

If you could spend <10 BTC a year and fuck up the charts so bad your opponents (and yes currency trading is a zero sum game) make mistakes that might net you 10,000 BTC in trading profits is it really losing money? :)

So what's the assessment? Is it just spamming? Or an attempt to fool bots? Ie. is the motive profit or vandalism?

Using some simple Python code, I get the following information regarding the volume of these trades (below 0.00005 BTC) in the last 24 hours:

largest trade volume: 0.000036 BTC
smallest trade volume: 0.000002 BTC
average trade volume : 0.000030 BTC
number of trades: 5515

Ie. 5515 trades were made in the last 24 hours that were below 50 µBTC. The largest of which was 36 µBTC, the smallest 2 µBTC and the average volume of all these trades were 30 µBTC. Here's the code if anyone's interesting in playing with it.

its costing my dumb bot a lot of money, lol.

the ticker data is affected by this as the small amounts test where the spread is. When im working off the ticker data and making trades to be in positions, im screwed.  my bad programming + my dumb bot + those stupid small trades =  24% loss today... needless to say, i turned my idiot bot off.

The fact so many people are upset by this makes me want it to continue.  Especially if it's eating other bots' lunch like the above poster indicates.   ;D

Why not just make your bot ignore trades under 0.01 BTC?

Ummm how come no one has noticed that this coincided with Mt Gox announcing discounting trade fees for Christmas  ;)

It just looks like whoever it is didn't notice that the discount doesn't come into effect until the 25th  ;D

The great thing about high frequency bitcoin trading is that MtGox offers an ideal playground for ideas to be tried and trading strategies tested.  See how the market reacts without losing significant amount of money in a market where there are very few rules and no enforcement.  It's great, for the person playing with their micro trading bit.

heh, I have to laugh at that.

I still don't understand the point of the 0.01BTC trade spamming.  Somebody help me out.  Why would I want to do this?

I don't really care if someone wants to waste money doing this. But I it drives me crazy because it really bogs down my browser, and especially goxlive. It also seems to take 2 minutes+ to get an order to change from "in queue" to "active", and that can make a huge difference on large movements.

Mtgoxlive is crap.  1 trade every second should be cake, let alone the 1 trade per 5 seconds we were seein when the bot was most active.  I have not noticed such long delays with orders, but then again I usually have my orders in place before large movements.  What price would it have to jump/fall to for you to trade?  Put an order in for that price and adjust further from the price if the price moves too close without spiking.

I DO *usually* have my orders in, sometimes for days ahead of time. But, occasionally, there's still that time when you want to get that quick buy/sell in.
Agreed too that goxlive can suck, but it is a helpful visual representation of the current market conditions, and much easier to just glance across then the sea of orders in the book.

It sucks, but it's what we have so far.  I use it every time I'm adjusting my positions, I just know it could be improved upon.  I'd write something myself, but I am busy working on a startup that will accept bitcoin as one of the payment methods.  Developing the merchant side of the bitcoin economy is the most fruitful endeavor right now IMHO.

Well, good luck with that! I hope to see something come from it. We ALL need it!  ;)

It will be a few months yet before you hear anything about it, but when we're ready, you'll be able to spend bitcoins to do something that millions of dollars are spent on every week in the US.  Of course, our site will be translation ready and we will have an affiliate program to help us grow outside this country.

I guess that settles the argument concerning motives :).