Can Strategy Trading Work?
- Thread starter Corso482
- Start date
Person 1: A guy I had lunch with a few weeks ago. He was head of prop trading at a large investment bank until just recently. Late 30s, he also has about 30-40M in the bank and vacations 8 months out of the year. His group was profitable every single quarter for ten years and it was pure pure system trading. They would do 1000s of trades every day in exchanges all around the world. He was very critical of me for having any non-systematized aspects of my trading. "Everything should be back-tested and you should have as many edges as possible in your data: intraday as opposed to daily, as many years as possible, as many securities as possible, etc."
Person #2: Former vice-chair at a major bank. Currently runs a $2B hedge fund focusing on distressed debt which has returned about 40%/year for the past 5 years. I emailed him about my approach, which includes some system aspects. His emailed response:
" All I know about any investment model that tries to examine
very short interval abnormalities is that they tend to create rather
large abnormalities in your bank account in the long term, after raising your confidence in the short term that it's a risk free game. Never is, never will be. As long as you realize that all you've achieved are a few good insights and mapped some data looking back that may hold up for a not very long future period of time, you'll do fine"
So the question is, who is right? Well, they both have the bank accounts and the track records to be able say "I am definitely right." And they are both extremely confident in their approaches. They both had other edges: person #1 had basically no commissions so was able to capture smaller inefficiences and make 1000s of trade a day. Person #2 has tons of inside knowledge and connections at all the banks and LBO shops so has an easier time spreading the risk then normal bond traders.
So the answer is, they are both right but they had other edges. Make sure when you calculate your chances of success that you really look at your edges/strengths as well as your weaknesses. A lot of people are trading systems and a lot of people are buying stocks with low P/Es or buying stocks that breakout after the first 15 mins, etc. You have to really examine your other edges and make them work for you on a trade by trade basis.
Person #2: Former vice-chair at a major bank. Currently runs a $2B hedge fund focusing on distressed debt which has returned about 40%/year for the past 5 years. I emailed him about my approach, which includes some system aspects. His emailed response:
" All I know about any investment model that tries to examine
very short interval abnormalities is that they tend to create rather
large abnormalities in your bank account in the long term, after raising your confidence in the short term that it's a risk free game. Never is, never will be. As long as you realize that all you've achieved are a few good insights and mapped some data looking back that may hold up for a not very long future period of time, you'll do fine"
So the question is, who is right? Well, they both have the bank accounts and the track records to be able say "I am definitely right." And they are both extremely confident in their approaches. They both had other edges: person #1 had basically no commissions so was able to capture smaller inefficiences and make 1000s of trade a day. Person #2 has tons of inside knowledge and connections at all the banks and LBO shops so has an easier time spreading the risk then normal bond traders.
So the answer is, they are both right but they had other edges. Make sure when you calculate your chances of success that you really look at your edges/strengths as well as your weaknesses. A lot of people are trading systems and a lot of people are buying stocks with low P/Es or buying stocks that breakout after the first 15 mins, etc. You have to really examine your other edges and make them work for you on a trade by trade basis.
In that same book I refered to in the original message, the author said something that made me really think. Again I'm paraphrasing, but he said something like:
"Take accounting for example. It's simply a set of defined equations that you plug different numbers into. You would think that such a profession would be ideal to be completely computerized, but it's not. Human accountants are still very much needed and can account better than any computer. Trading is the same way."
This quote provides a better argument than the first. How do you guys explain that accounting cannot be done by computer, yet many claim that trading, which is even more subjective, can?
"Take accounting for example. It's simply a set of defined equations that you plug different numbers into. You would think that such a profession would be ideal to be completely computerized, but it's not. Human accountants are still very much needed and can account better than any computer. Trading is the same way."
This quote provides a better argument than the first. How do you guys explain that accounting cannot be done by computer, yet many claim that trading, which is even more subjective, can?
CORSO
Trading is all about processing/deciphering information.
For some reason your (non-computer) mind is not able to process/decipher the information that people have given you here.
Maybe we should bring in some accountants to help you
Maybe accountants are brought in to cook the books (Enron/Worldcom) another words screw up what computers are doing just fine
Your paraphrased statement is not even close to being an argument against computerized trading rather it is a sales pitch for some guys discretionary trading method - one that has no "ACCOUNTABILITY"
Inventory management software with pricing models built in does accounting just fine and can automatically reorder based on sales vs inventory. Trading software is basically no different with the exception that it gets to be written directly in to a datasource so that the info never needs to be feed in by humans/accountants
Trading is all about processing/deciphering information.
For some reason your (non-computer) mind is not able to process/decipher the information that people have given you here.
Maybe we should bring in some accountants to help you
Maybe accountants are brought in to cook the books (Enron/Worldcom) another words screw up what computers are doing just fine
Your paraphrased statement is not even close to being an argument against computerized trading rather it is a sales pitch for some guys discretionary trading method - one that has no "ACCOUNTABILITY"
Inventory management software with pricing models built in does accounting just fine and can automatically reorder based on sales vs inventory. Trading software is basically no different with the exception that it gets to be written directly in to a datasource so that the info never needs to be feed in by humans/accountants
I think they need the people to do the data entry!
For example, my Turbo Tax has been right on every year. It even catches deductions that my human brain did not! I can just imagine the # of errors and deductions I would have missed if I didn't use Turbo Tax.
I think you need a different analogy here...
I find it hard to believe it's a sales pitch considering the guy wasn't selling anything. Nevertheless, it an argument. It points out a task similar to trading and demonstrates that it cannot be computerized, implying that that the same conclusion can be reached for trading. He's drawing an analogy, and a powerful one at that, because accounting is even simpler and more easily defined than trading. However, I admit that the argument is inherently weak because it assumes that trading and accounting are the same with respect to how they can or cannot be computerized.
Yes, simple accounting can be done on the computer, but accountants won't be replaced by Quicken anytime soon.
Why don't we have computer accounting programs that simply accept variables and spit out your tax returns? Such programs would save huge corporations millions. But for some reason, human accountants are still very necessary.
Maybe computerized trading is the best thing sense sliced bread. I don't know. I'm just discussing it. Chill out.
I suppose you've never even seen Oracle Financials? You enter the data and it does the accounting. In many cases it grabs the data direct from other electronic systems. But that's besides the point. Your argument doesn't make any sense.
A human accountant + Quicken ==> produces accounting paperwork.
A human trader + software that produces an edge ==> produces profits.
And if you're arguing against 100% fully automated trading (vs. a system that produces signals and a user does the order entry), you might want to hop on one of the automated trading systems (ATS) forums out there. Also, you may want to do some research on the 100% automated arb programs out there that do millions of dollars of transactions every day arb'ing out small inefficiencies in the market.