Topic: Mtgox high frequency trading. (Read 4646 times)

I want to do one. But there is a certain issue. If the price swings one way or the other more predominately, you will run out of one or the other liquid currency in your account. You need to perform some stochastic calculations and go away from the median price by calculate percentages and calculated quantities such that you don't need to mess with your liquidity too much. The bitcoin part is instant, but to add/remove USD you might need a few days where you can't trade.

And do you realize that NANEX has an exhaustive analysis of what happened to the NBBO (National Best Bid/Offer) in those minutes? I think you'd have less of a cavalier attitude if you read their in-depth research. The HFT algos pulled out of the market, instead of staying in and making bids/offers like real trading specialists do on the floor with the order book.

I don't call that 'providing liquidity', when their first action was to power down and run like hell.

There have been daily patterns in other stocks that mirror the flash-crash, things trading at 12$ going to pennies and back. That isn't a healthy market, it is a feedback loop of chaos that will one day converge in a very real way on the indexes. Hope you don't have any equities or mutual funds, because they will all get smashed to bits.

I've traded markets for a good while, you would find it in your best interest to educate yourself on what is happening to them.

You do realize the flash crash only lasted about 4-5 minutes right? The market exactly did its job.

High Frequency Trading is a festering crap-pile. It is market cancer. May flash crash? Oh, that's right, the HFT boys pulled the plug, instead of 'providing liquidity'. Anyone who believes they do either works for the squid, or has their own co-located box getting exchange 'rebate' kickbacks.

HFT can 'eff up the equity markets all they want, we don't need them here.

It doesn't have to be complex or complicated. Can be something like: If bitcoins drop below 10$/btc then buy 3 coins. Place sell order at 15$/btc for 3 coins. If sell order successful then place buy order for 10$/btc for 3 coins, rinse and repeat. If the market is volatile between 8-17$/btc bouncing up and down then this could be profitable. If you buy and the price of btc drops to 7-8 then the software just stands idle until the coins are sold at 15$. You could also make the buy/sell orders for 1$ spreads or whatever amount you want. The idea is to buy low and sell high. If that fails then its basically buy and hold long term until price recovery.

Of course there is always the risk of loss in such a situation, I am just thinking of this as a, "for fun" project. Maybe writing a simple script, dropping $500 into the mtgox account. Configure the script and let it run for a couple days. If I lose the money, then it will be fun. If I make money then great.


I don't understand what you mean by this. You would write the trade bot so it pulls btc price data from mtgox and then write triggers based on that price. The triggers would send buy/sell orders. Once one of those buy/sell orders is filled then it would place a different order. All the orders will be placed in such a fashion that you benefit from the order being executed.

I was thinking of making a javascript trade bot.  The problem is sometimes I just can't react fast enough only a computer can.

I've made a positive and not insignificant (percent-wise) profit day-trading ~10 times per day on volatile market days. I'm not sure that HF trading would net me any more though. My MtGox client could theoretically support it, but I'm really not the person to write an AI for trading...

Goldman sachs of the bitcoin community? Hardly. HFT has many benefits that may not be apparent, for example it may also increase liquidity. As bitcoin gains in popularity people are going to want to have instant access to USD. Take for example, a bitcoin based debit card. I can see a point where the debit card is backed by bitcoins. When you want to make a purchase, the debit card company goes to mtgox (or other exchange) and instantly converts your bitcoins to currency to process the transaction. HF traders may have the cash in their account to fund these types of transactions without forcing the markets to move drastically.

Also, yes, I totally agree that more merchants should start accepting bitcoins. However, why wait for merchants to adopt it? Its better to have consumers adopt it (in my opinion) by using user friendly services like bitcoin debit cards where the account is held in bitcoins then converted to USD when a transaction is processed. This allows consumers to use their bitcoin account as their primary savings/checking account - that could have HUGE potential to the bitcoin community.

MtGox is slow and has high fees (relative to the volume that is moved by HFT) for it to be worth the effort right now. Tradehill has been staying within ~$1 or so of MtGox, which is fine.

This would be great. Please make one. (seriously)

A high frequency trader that does Mt. Gox <-> Trade Hill arbitrage would be more beneficial to the Bitcoin community.

Here comes Goldman Sachs to the Bitcoin Community , is that what we really need?

Trade exchanges have often had this problem when they 1st start, but once the number of products and services grows the value of the dollar finally begins to settle. If you want to determine the value offer it to the market place for cash and see what offers you receive.

To stabalize the price you need more product that can be purchased using it. Once you are able to purchase gadgets such as iphone 4, TV and Clothes, Food , real estate etc the price will stabalize. Bitcoin needs to backed by goods and services to stop the volatility

To achieve this we need to encourage more businesses to get involved.

The value of any currency is realitive to its purchasing power.

Anyone interested in doing something like this? Basically writing a software that does analysis of the trade charts and places buy/sell orders in an attempt to make a profit. Something like this:

Buy order for 13$/btc
Execute at 13$/btc
Sell order for 14$/btc
Execute at 14$/btc  (profit realized of ~1$/btc Less cost)

Then the software makes another similar bet and waits. The software never sells for a loss. Of course the risk is that you could buy for x price and the price drops far below your buy point and never reaches your sell order. In that case you might have to wait a while for the price to reach your sell price.

Right now it seems there is good opportunity since the markets are very volatile.

Also this type of automated trading could help benefit the bitcoin exchanges because this type of trading will help stabilize the price of bitcoins. During periods of high bitcoin demand (price rising) the automated software will start dumping btc which should slow down the price rise. During periods of lots of bitcoin selling, the software will buy up btc which will slow down the price drops.

The only thing that will move the markets are huge buy/sell orders.

Thoughts?