Topic: MtGox API: zero volume depth updates (Read 1212 times)
Multiple connections with synchronization layer to the same websocket (websocket.mtgox.com) does not help. It seems server drops all connections simultaneously. Synchronization can help with two connections to different websockets probably: websocket.mtgox.com and socketio.mtgox.com
Hi!
I am software engineer and designer, developing some free bitcoin-oriented visualisation. And i really cannot get in with Mt.Gox API(s) too. Initially i want to write my own topic, but when i saw this one, decided to join in. I've got the same problem (zero volume packet) plus even more curious data sometimes:
Depth channel of streaming API v2:
Total Volume. It's seems semi-legit for sometimes, but sometimes it has a negative value. wtf?
If zero volume diff may looks like a redundant artefact, negative total volume is just impossible.
What is the best-practice to listen price- and timestamp-accurate orders history?
By the way, issue about the session expiring. Sometimes server drops the connection every several minutes. How can i avoid it? Authenticated access may improve it? Multiple connections with synchronization layer? (Oh shi...)
Just show the right way, please. I googled all of these questions back and forth and the only helpful link that i've got is this thread. Docs are missy and messy.
I am software engineer and designer, developing some free bitcoin-oriented visualisation. And i really cannot get in with Mt.Gox API(s) too. Initially i want to write my own topic, but when i saw this one, decided to join in. I've got the same problem (zero volume packet) plus even more curious data sometimes:
Depth channel of streaming API v2:
Total Volume. It's seems semi-legit for sometimes, but sometimes it has a negative value. wtf?
If zero volume diff may looks like a redundant artefact, negative total volume is just impossible.
What is the best-practice to listen price- and timestamp-accurate orders history?
By the way, issue about the session expiring. Sometimes server drops the connection every several minutes. How can i avoid it? Authenticated access may improve it? Multiple connections with synchronization layer? (Oh shi...)
Just show the right way, please. I googled all of these questions back and forth and the only helpful link that i've got is this thread. Docs are missy and messy.
I will repost in a more appropriate thread.
A very good question! I'd like to know, too... 
[Posted here because of unbelievably CENSORED posting restriction]
Does anyone know what the deal is with those streaming API depth updates with volume_int set to zero? The total volume still changes, breaking cumulative checks. It appears that about 5% of all depth updates are of this form, with some clustering with respect to the price.
Does anyone know what the deal is with those streaming API depth updates with volume_int set to zero? The total volume still changes, breaking cumulative checks. It appears that about 5% of all depth updates are of this form, with some clustering with respect to the price.
Jump to: