Let’t talk pair trading and arbitrage video playback

Links from the groups discussion

http://www.financial-spread-betting.com/wp-content/uploads/2011/04/pairs-trading.gif

Standard deviation 2

we use a time-based exit

the spread is an Ornstein-Uhlenbeck process

so you can compute the first passage time

once that time has elapsed, you know the trade has broken down

yeah we exit when the spread goes near the mean

usually 0.5 standard deviations away

Roughly

yeah, we use GPUs to findd the parameters and pairs to trade

opencl

https://www.quantstart.com/articles/Backtesting-An-Intraday-Mean-Reversion-Pairs-Strategy-Between-SPY-And-IWM

http://www.surinotes.com/

the power of GPUs 🙂

we have tick data

it’s all historical patterns

we’re a pure quant fund – no fundamentals

well, we use 1M for testing

but we use bid and ask from tick

Do you use convolution methods to search for patterns?

yeah we use mid for entries

no we don’t, we literally just test every parameter combination

You can use stuff like shift invarient dictionary learning, basis search, any sort of multiscale wavelet will also give you a pattern dictionary.

post it on your blog if you remember 🙂

ATR EPS Market Cap

• Trendline on the spread between closing to long/short
• SMA

EXIT/Close-> Beta / SMA trendline TA/stats distribution/IV

you can use bollinger bands to do the same as what I’m doing

Sorry, Volume weighted price average for entry/exit

Jurik’s JMA is pretty good

https://www.quantstart.com/articles/Backtesting-An-Intraday-Mean-Reversion-Pairs-Strategy-Between-SPY-And-IWM

Join my FREE newsletter to learn more baout automatd trading like this

 NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!