Normal
So I will attach 3 of the papers that are associated with the book 'Evidence Based Technical Analysis'.Having read the papers, then in a highly simplified definition: what is being sought is a statistically significant edge. To differentiate out of the mass of data a causation as opposed to random luck.Obviously I haven't even received the book yet, never mind read it, so that definition may become more nuanced moving forward.From Mr Chipp's post abovr:Much of what i code is testing to exploit price imbalances , this is where a lot of ' gold ' exists . Mostly based around logical patterns created by events .So these two sentences drive the following thoughts:(i) Arbitrage is 1 instrument with 2 different prices existing at the same time. Sell 1 buy 1 and lock in the profit. Buy or produce physical gold against a futures contract for gold. This was the basis of the futures markets.(ii) Risk Arbitrage 1: An example: An off the run 10yr bond and the newly issued 10yr bond. now technically, these are 2 different instruments, with slightly different durations. However both will pay par at maturation, or $10,000 face value. The price differential is usually measured in basis points. The risk is the differential in time to maturity.Obviously, this is the trade that (amongst others) that blew up LTCM due to taking massive leverage to increase the dollar returnsI would classify this as Quant.(ii) Risk Arbitrage 2: where Company A is going to buy Company B and tenders an offer for outstanding shares above the market price. Buy shares of B and wait for completion. The risk is that it never closes.So returning to the quote above:Price imbalance: yes in our 2 examples.Event: yes in our 2 examples.Now on a purely technical analysis type of scenario we have some examples:(i) Dogs of the Dow,(ii) The January effect,(iii) Closing imbalances between SPX Futures and SPY ETFThis I would classify as the more fertile ground for what we are discussing re. backtesting and systematic trading.Again we do have:(i) Price imbalances,(ii) EventReturning to the LTCM example: the issues were that because of the 2 different instruments involved, prices were not locked together, ie as one went down, the other went down and the massive leverage involved meant that as the spread opened the losses magnified past the risk capital held.Hedge Funds are still using variants of the LTCM example: The Yen Carry trade is simply a variation of the above, which went into meltdown last Monday.My boss (technically) had his hands full last Sunday night/Monday morning as a cadre of SPX traders in the firm were over-leveraged and just about blew themselves up. Lucky escape for them. They lost a ton of money but are still alive...just.The reason I mention it is as a firm, many traders follow a simplified market quant system. I post it most weeks. I personally don't, but apparently it has been backtested and is profitable.So I'll post them here just in case anyone wants to test them:[ATTACH=full]182351[/ATTACH][ATTACH=full]182347[/ATTACH][ATTACH=full]182348[/ATTACH][ATTACH=full]182349[/ATTACH]So nice and simple.This is the 'edge'.Which 'should' be in excess of random luck.jog onduc
So I will attach 3 of the papers that are associated with the book 'Evidence Based Technical Analysis'.
Having read the papers, then in a highly simplified definition: what is being sought is a statistically significant edge. To differentiate out of the mass of data a causation as opposed to random luck.
Obviously I haven't even received the book yet, never mind read it, so that definition may become more nuanced moving forward.
From Mr Chipp's post abovr:
Much of what i code is testing to exploit price imbalances , this is where a lot of ' gold ' exists . Mostly based around logical patterns created by events .
So these two sentences drive the following thoughts:
(i) Arbitrage is 1 instrument with 2 different prices existing at the same time. Sell 1 buy 1 and lock in the profit. Buy or produce physical gold against a futures contract for gold. This was the basis of the futures markets.
(ii) Risk Arbitrage 1: An example: An off the run 10yr bond and the newly issued 10yr bond. now technically, these are 2 different instruments, with slightly different durations. However both will pay par at maturation, or $10,000 face value. The price differential is usually measured in basis points. The risk is the differential in time to maturity.
Obviously, this is the trade that (amongst others) that blew up LTCM due to taking massive leverage to increase the dollar returns
I would classify this as Quant.
(ii) Risk Arbitrage 2: where Company A is going to buy Company B and tenders an offer for outstanding shares above the market price. Buy shares of B and wait for completion. The risk is that it never closes.
So returning to the quote above:
Price imbalance: yes in our 2 examples.
Event: yes in our 2 examples.
Now on a purely technical analysis type of scenario we have some examples:
(i) Dogs of the Dow,
(ii) The January effect,
(iii) Closing imbalances between SPX Futures and SPY ETF
This I would classify as the more fertile ground for what we are discussing re. backtesting and systematic trading.
Again we do have:
(i) Price imbalances,
(ii) Event
Returning to the LTCM example: the issues were that because of the 2 different instruments involved, prices were not locked together, ie as one went down, the other went down and the massive leverage involved meant that as the spread opened the losses magnified past the risk capital held.
Hedge Funds are still using variants of the LTCM example: The Yen Carry trade is simply a variation of the above, which went into meltdown last Monday.
My boss (technically) had his hands full last Sunday night/Monday morning as a cadre of SPX traders in the firm were over-leveraged and just about blew themselves up. Lucky escape for them. They lost a ton of money but are still alive...just.
The reason I mention it is as a firm, many traders follow a simplified market quant system. I post it most weeks. I personally don't, but apparently it has been backtested and is profitable.
So I'll post them here just in case anyone wants to test them:
[ATTACH=full]182351[/ATTACH][ATTACH=full]182347[/ATTACH][ATTACH=full]182348[/ATTACH][ATTACH=full]182349[/ATTACH]
So nice and simple.
This is the 'edge'.
Which 'should' be in excess of random luck.
jog on
duc
Hello and welcome to Aussie Stock Forums!
To gain full access you must register. Registration is free and takes only a few seconds to complete.
Already a member? Log in here.