Reputation
680
Top tag
Next privilege 750 Rep.
See votes, expandable usercard
Badges
1 7 17
Impact
~37k people reached

Mar
12
awarded  Necromancer
Mar
12
comment What different methods of pairs selection exists? (For Pairs trading)
@cf16 Thanks; not being a mathematician I suspect I am using the word correlation sloppily: I mean finding a function of the price movements of two symbols that is mean-reverting.
Mar
12
comment What different methods of pairs selection exists? (For Pairs trading)
@GoodGuyMike Your original question was about what is hot in this area, so you might be interested in the 2012 paper mentioned here: en.wikipedia.org/wiki/Cointelation
Mar
12
comment What different methods of pairs selection exists? (For Pairs trading)
@cf16 Was the "this" as a response to GoodGuyMike's comment, or directed more generally at my answer?
Mar
5
comment How to implement a long-term trade on oil?
Like Freddy I'd also question the idea that oil price is guaranteed to go up long-term. Personally I think it will, but lower-demand (eco pressures meaning alternative fuels taking hold) and more supply (new discoveries, fracking, etc.) are significant risks, so it is not a dead cert.
Mar
5
comment How to implement a long-term trade on oil?
This was my first thought (depending on the curve vs. your target price of course). The long-dated contracts exist for that purpose: people who want to lock in a lower price now.
Feb
24
comment What different methods of pairs selection exists? (For Pairs trading)
@GoodGuyMike I think 'spuriously cointegrated' means the same as what I mean by data mining noise. If you analyze 1000s of pairs that genuinely have no connection your analysis will still suggest dozens of seemingly good candidates.
Feb
22
comment What different methods of pairs selection exists? (For Pairs trading)
@GoodGuyMike Yes, news analysis can be included in automated trading, but that might not count as "easy". However your question was about pair discovery: my point was only look at symbol pairs that have something in common; anything else you discover will most likely be data mining noise.
Feb
13
revised Why is C++ still a very popular language in quantitative finance?
Updated my comments on C#
Feb
13
answered What different methods of pairs selection exists? (For Pairs trading)
Jan
30
comment A non parametric study of VaR with kernel density
@pyCthon Don't keep us in suspense, do you want to list some of them? :-)
Jan
28
comment Generate tick data from candlestick
@geektrader P.S. The website in your profile gives a "Not Found" page.
Jan
28
comment Generate tick data from candlestick
@geektrader That is misunderstanding the purpose: the 4 ticks are from a bar for a system that will then create a bar, but can only do it from ticks. OHLC vs. OLHC makes no difference (the candlestick looks the same). You can validly do backtesting at the bar-level (as your data at that level is genuine), but cannot backtest a strategy using ticks made from bar data. As I said, if you really want to backtest with ticks, add lots of random jitter, and repeat a few times. I suspect you will see variance in results, and so can prove to yourself, quantitatively, that it is a bad idea :-)
Jan
22
comment When hiring a quant, how can I protect my IP?
P.S. If you want to talk more about your project, off-line, I'd be happy to. My email is in my profile.
Jan
22
answered When hiring a quant, how can I protect my IP?
Dec
1
awarded  Popular Question
Nov
24
awarded  Notable Question
Nov
7
comment Generate tick data from candlestick
@Freddy It was a real, commercial system, being used by real traders. Not my system, I was just dealing with it. That you won't become their customer has been noted :-)
Nov
7
comment Generate tick data from candlestick
@Freddy Read the first half again. Working with a system where I needed to backfill data but we only had 1 minute bars, not ticks. The system in question didn't allow importing bars, it only allowed importing ticks.
Nov
7
comment Generate tick data from candlestick
@Freddy Just to be clear, I personally always use real tick data when I want to backtest a tick stategy, and use bar data when I want to backtest a strategy that uses bars. I think we're in full agreement that it would be foolish to do otherwise.