Mathematical expectation

[xelite777]

Can expected value be calculated in trading like one can calculate the (negative) expected value of casino games such as blackjack or roulette?

For winning trading systems, that expected value is simply the average profit per trade (in dollars and cents) and can easily be calculated (total net profit divided by number of trades).

That expected value ($500 per trade per contract for instance) can also be expressed in percentage, and in that case we could say that trading system X gives us a 2% trading edge over the "house" (here the market you are trading), for example.

The expected value is only an average and just gives you an idea about the profitability of a system (so that you can compare different systems quickly), you cannot use it to predict the magnitude of future gains.

It's like counting cards in Blackjack, you know your system will give you a 1% edge per hand ($1 for each $100 wagered) but the profits can vary greatly from day to day. Of course winning trading systems with a higher expected value (edge) will have a smoother, less volatile equity curve.

In general, few mechanical trading systems can generate more than $700 average profit per trade per contract in the Futures market.

 

[plsnobull]

Of course winning trading systems with a higher expected value (edge) will have a smoother, less volatile equity curve.

In general, few mechanical trading systems can generate more than $700 average profit per trade per contract in the Futures market.

Pls no BS.

 

[xelite777]

Pls no BS.

Beg your pardon??

 

[Nighthawk]

Read a book about "black swan protection" and stop using "mathematical models" which do not deserve to be called this way. Expected value? Expect your money to blow up EVERY MINUTE because your "expected value" is not prepared for it.

I saw more than 1000 quants ranting about TA and "back testing" and saw the same quants blowing up their accounts during the financial crisis.

A true, live, real economy "expected value" is given very impressively by Ben Lichtenstein - Ben is the voice of the CME S&P futures pit. This is the video endorsed by Ben and TradersAudio.com:

http://www.youtube.com/watch?v=1mC4tu1NhUA

 

[SIUYA]

my 2 cents FWIW......
...
Van tharp and others have popularized something that confuses and gives people the wrong ideas and false hope.

Quant analysts fall into 2 types - those that unerstand that what they work on is a model, and those that think their model is the market.

technical analysis is simply an input into a quant model.

An edge is simply a mathematically applied edge that is an arbitrage - one that uses a mathematical formula to be able to show that based on a what a market given all the accurate costs and assumptions that the market must revert back to some end (or fair price)....in my mind if it cant be arbitraged around this fair value it is effectively not and edge and this fair value might be something like a futures fair price, an option value fair price and often consists of nothing more than being able to constantly buy and sell around this edge. (their edge is the buy sell margin around a fair value and is reliant on continual volume and trading and spreading around this)
This is something you can model reasonably well going forward
.... retail trades will never be able to compete for an edge in reality on this level even if they can model it.

They may however be able to compete on having a larger buffer for error, and expecting the market to return back to some idea of fair value.....but this is not an edge. (This is where Retail can thrive)

When it comes to mathematical expectancy...
to the first OP question "Can expected value be calculated in trading like one can calculate the (negative) expected value of casino games such as blackjack or roulette?"
Yes - based on actually applying an edge as I see it. ...you can model forward. That you should be able to quanitify it if you can quantify the arbitrage.(there are always lots of what ifs, but over the long term yes.)
and
No - as the market does not have the same rules - or should I say does not operate on any set of defined rules - as those games and most people are not actually able to participate in any type of arbitrage/spreading edge. Instead they operate in a world where even if they do the work it still becomes a fudge (through back testing or what ever) .......but thats better than nothing....but you cant model foward, only infer.

Because you are not applying rule that 'the market must behave like' instead you are applying rules that will profit 'if the market behaves like' a backtest is a fudge and not anything you can say must produce a positive expectancy - all it says is that if the market continues to behave in this manner, then I can expect this outcome.

 

[OddTrader]

Quote

Edge is like teenage sex: Everybody talks about it; Nobody really knows how to do it; Everyone thinks everyone else is doing it; So everyone claims having it; ...

Unquote

 

[ronblack]

Such as?

I mean seriously, how do you know that your system will make money in the long run if you do not test it first, with enough historical and/or live data?

I suppose you asking me to reveal in a public trading forum what I have found after spending thousands in real tests, not backtests. Just by saying so is already revealing enough. Your task now is to find the way yourself...

 

[ronblack]

my 2 cents FWIW......
...
Van tharp and others have popularized something that confuses and gives people the wrong ideas and false hope.

+1

You should also consider the possibility that these authors did not understand what they popularized and they diid not care aanyways eebcause their objective was to sell books and courses. Trading is a non-ergodic process and expectation is the wrong metric to use.

 

[kut2k2]

Quant analysts fall into 2 types - those that unerstand that what they work on is a model, and those that think their model is the market.

+1. You nailed them.

technical analysis is simply an input into a quant model.

No. Most TA is just ad hoc attempts to separate "significant" price moves from the random "noisy" ones but there is no explicit model involved like there is in QA.

 

[xelite777]

I suppose you asking me to reveal in a public trading forum what I have found after spending thousands in real tests, not backtests. Just by saying so is already revealing enough.

I am not asking you to reveal anything, you said that backtesting is not really necessary and that there are other alternatives and I simply asked you to be a little bit more specific, without revealing your Oh-so-secret...secrets.

Your task now is to find the way yourself...

A quote from a soap opera TV show, I suppose?