Bitcoin Forum

A Thread for bid/ask analysis..

i watched this jumps on the ask-side the whole day. i think it is a bot playing with maaaany BTC or something, no manipulator.

http://img341.imageshack.us/img341/1493/jump.png

what do you think?

About 3500 or so, I've been seeing that too! Very annoying!

Hope drop below $2, then i'll buy some coins again.

I never realized the dark theme on mtgoxlive looked so awesome  :D

IMO large walls are largely irrelevant, they mainly serve as a showoff in the confidence game. The fact that they are moved so often lets me think that the Bids & Asks shouldn't be displayed in this fashion. Rather use some sort of decay parameter where orders become more relevant the longer they have been up.

Some of these walls move every minute or so and if we use this parameter they would show up as smooth curves instead of walls... So the 'manipulation' occurs when people attribute too much significance to them.

So my thesis is: If someone really wants to buy or sell at a price he'd make an order and leave it. Otherwise it shouldn't be considered important, at least only to the extent of time it has been up.
Of course in some cases were a large order has been up a long time and is suddenly hit the owner often pulls it which makes the whole thing more complicated.

This results to the conclusion that the order should be scaled by a binomial distribution curve of about a day. So one day old orders would be the most reliable ones and anything sooner or later than that takes less relevance ;)

indeed the dark theme looks awesome ;)

i support your thesis.

I like the idea of changing the display based on order longevity, but don't think that scaling is the right way to display it.  My thought would be to display orders as a stacked graph with different layers for <15 minutes, >1 hour, and >1 day longevity.

After I learned that the bid & ask walls could be completely irrelevant because of The Manipulator(s), market movement started to make more sense. It's just a learning hurdle all n00bs have to jump over, which is lame.

One of the options at Bitcoinity is to show all market depth changes for the past hour, rather than just the past few minutes.  You have to change Preferences--> linear depth scaling to make any sense of it, though.

I also had the same jumps with my wall, but I didn't do anything with it.
It is a bug in mtgoxlive.

What I think?

Uhm... I like jumpiness. Makes things look more animated. :)

Also, I don't know how to justify it, but it feels like liquidity has improved. And I don't understand why people care so much about what other people do with their orders. It's a few thousand coins jumping, wow, world's not gonna collapse.

Yes, but that kind of fails the point.

People in this forum want to derive support / resistance levels from it, and that's what the thread is about.

IMHO all bids and asks reported by mtgox should be completely disregarded or at least taken with an enormous rock of salt. Bids and asks can be automatically listed and removed in an instant by a computer program for any or no reason, and so I will not be using them to make financial decisions.

Isn't that like asking to be manipulated? Bonus points for posting what kind of orders are most effective to influence people.

I'm really not with "the manipulator" crowd, but I'm sure some people are willing to risk some money for fake orders. The depth chart hasn't proven to be very useful in the past, either.

Yes, however you cannot completely discard the information presented in the order list without compromises. Some of these orders are filled and so in order to use the information invalid information has to be discarded.

I don't really think that any subjective interpretation will help with it since that is the target of this manipulation in the first place. (What happened today is a good example, the bidwall was moving closer and closer and then the price was suddenly forcefully bought out and volume hasn't increased so the manipulation succeed)

But this is only the most obvious example, a heuristic system calculating the probability a particular order would be filled could filter out much much more.