973200+ entries in 0.703s
mjr_: but you may need
to install some libraries for it
to work
mjr_: i put my code up earlier if anyone wanted
to look
mjr_: but i would find it very interesting
to see
them interact
mjr_: so
that way if you have a small order, i am reasonably sure i'd get filled after you for one satoshi more
mjr_: i
think i just said "if volume less
than 5000, go 1 satoshi behind, else go 1 satoshi in front
jurov: and i kept it so, i
think it discourages others from jumping in front of it
jurov: it was originally intended for random stress
testing
mjr_: and i was referring
to mpoe
jurov: nice! so far it looked like everyone is
too lazy and i'll have
to add it
to
the agent myself
mjr_: edit: i mean a browser based
trading platform
mjr_: in fact, if you made
that interface, you could actually use selenium for your
trading bot, which would be pretty cool
mjr_: which i
think would be PERFECT for a web based
trading platform
mjr_: i wanted
to play with angular.js
mjr_: which would not be fun i
think
mjr_: yeah i figured it'd be easiest
to use web so i wouldn't have
to actually create charts in a desktop app
mircea_popescu: for
the record, i
think
that guy must be posting out of about half
the btctalk accounts
jurov: well. so after it will be able
to serve files, one just
throw
there some javascript
that can do jsonrpc calls
mjr_: my app wasn't performance oriented but rather i wanted
the easy UI admin from django
mjr_: but i like
twisted a lot
mjr_: i am gonna look at
the code now
jurov: i was
thinking about adding fileserving fo mpexagent
mjr_: ie. integrated charts, depth of book,
ticker, positions, executions, real
time p&l
mjr_: which i
think would be a huge value
to
the
trading community
mjr_: i mean, IDEALLY you could
then use a lot of
those pieces
to build either a web based
trading platform or an actual desktop application
mjr_: i would have
to write
the position monitor, and i would have
to write a driver so my orders could be
translated
to mpex orders
mjr_: after
that point, it would only be creating a position monitor, and
then linking
to a driver for pympex (or something at coinbr)
mjr_: and if i could monitor
them in real
time
mjr_: i wanted
to see if
the crazy spreads i saw last night were real
mjr_: not sure about execution
time etc, but i don't
think i need
to whip out
the c++ yet, do I?
mjr_: python is easy and it is a nice way
to
try a PoC
mjr_: i mean, i mostly wrote
the spread finder cuz it was fun
mjr_: which i
think would be awesome
jurov: no.
that paragraph says "licensing
the code"
mjr_: i
think he was referring
to licensing code
to me
mjr_: so i could build
that engine
to monitor my account and
then update my positions and orders
table
mjr_: i would mostly just like an ack on order being received, and
then an execution be sent if i got filled
mjr_: would be nice i suppose, but parsing sql and parsing json is about
the same effort
mircea_popescu: The Laborers' Local 265 Pension Fund of Cincinnati and
the Plumbers and Pipefitters Local No. 572 Pension Fund of Nashville
mjr_: they are made because
they are looting returns...best performer in
their portfolio
Bugpowder: Glad nobody is big enough
to buy Vanguard.
Bugpowder: changees fee structure
to
take 40%
jurov: go
to coinbr.com/about, last paragraph
jurov: mjr_really, how much do you willing
to pay for ability
to just doing it by simple select?
mjr_: yeah, its funny cuz everyone waits
till
the end
Bugpowder: it will heat up in a week or
two
though
gribble: BTCUSD
ticker | Best bid: 20.48948, Best ask: 20.49800, Bid-ask spread: 0.00852, Last
trade: 20.49800, 24 hour volume: 32329.47027016, 24 hour low: 19.42000, 24 hour high: 20.75000, 24 hour vwap: 20.12737
mjr_: ie. sell
THOSE shares
that i got cheap, but no more
mjr_: i
think i am going
to keep
track of individual fill prices, so i can know if i should sell or not
mjr_: (if
the bot is designed well)
Bugpowder: and my bot selling 1000s contracts
to itself
Bugpowder: smickles
trading a few here and
there
mjr_: its just
that over
the rest of
the year, everyone sending market order will give me a
tiny bit of money
jurov: but if
they do everyone watches alertly :D
Bugpowder: nothing happening in options
this month
mjr_: yeah, but i mean, i can say get 5 $20 calls...at 1 satoshi over current, and
then if i ever get filled, just replace
that order and
then open up sell at 1 satoshi beneath
jurov: maybe it got better since, but still
they don't appear often in
the
ticker
jurov: mjr_i did options,
too. but you can wait like weeks
till
there's som movement
mjr_: but overall, if i can say
that in any given situation i am getting
the best possible price,
that is ok
mjr_: the volume is
the part i haven't really figured out how
to adjust for
mjr_: but, as long as
they are not killer mistakes (don't let
the machine manage more
than you can risk), you can fix it and make it stronger
mjr_: it happens
to billion dollar companies (knight capital)
jurov: i
tested my bot several months... and still occassionally does snafu
mjr_: since i am responsible
to ask constantly
mjr_: that is
the one part
that i will have
to
test very rigorously
jurov: human will usually notice
that
the order got stuck.. bot no
mjr_: i am gonna give my little guy only a little money
to play with
mjr_: yeah
that is my worry
too
jurov: that wouldn't exactly be
the problem. suppose it blows up when i'm not around
mjr_: i
think
the main way
to do
this would be with
the options...the spreads are regularly 33%-50%
mjr_: couldn't you set a limit on order number
through api?
jurov: and as for
the api, i'd really like it... but having
the orders count suddenly increase 10x is just
too big risk
mjr_: was sad
to find out no api
gribble: BTCUSD
ticker | Best bid: 20.48947, Best ask: 20.49800, Bid-ask spread: 0.00853, Last
trade: 20.49800, 24 hour volume: 32327.68690484, 24 hour low: 19.42000, 24 hour high: 20.75000, 24 hour vwap: 20.12431
mjr_: although currently i am running it with .5%
to
ttest
mjr_: as long as you are getting more
than .4% (buy and sell) you are making money...
mjr_: so i would weight it's priority
to buy and sell accordingly
jurov: you will reassssess and find out
that bid is is higher
than last sell price.. what now?
mjr_: also
the ultimate goal is not so much
to scalp and front run, but rather build up as long a position as i can for
the least money
mjr_: it doesn't seem like it is moving a
ton, but of course, if i get filled on one end, i cancel orders and reassess
mjr_: the stock is not
too much...and although over its life it has moved a lot
jurov: they are quite common and usually cause simple frontrunners
to get stuck on wrong end
mjr_: it wasn't close
to
the next higher or lower price
mjr_: but
there is still profit
there...and i set it so
that if you put a lot of volume, i go a satoshi cheaper...if it is
thin in front, i wait behind for extra satoshi
mjr_: i would obviously need
to have it use pympex
jurov: why would i, or
the bot, want
tight spread?
mjr_: or is
the bot
there
mjr_: 1.22% spread...what is
the bot doing?
gribble: BTCUSD
ticker | Best bid: 20.66002, Best ask: 20.73900, Bid-ask spread: 0.07898, Last
trade: 20.73900, 24 hour volume: 31522.03367265, 24 hour low: 19.42000, 24 hour high: 20.75000, 24 hour vwap: 20.10086
jurov: yeah
they'are amusing.. i got windows xp booting gif
the other day
Ukto_: I was debating using it
to buy some bbet hehe
jurov: i heard
they can't afford 0.09 BTC/month