Backtesting - is it a valid concept ?

The General

Active member
199 16
Ok, rephrase. You have 3 systems each with a historical probability of 60% (here we assume when we get a winner, we win 1 pound and when we lose, we lose 1 pound). Now, as defined by classical historical probability, all 3 systems have positive expectancy.

You need a way of determining which one system will give the best chance of a winner on the next trade. How do you do this ? All 3 have the same historical success rate ?
 

TheBramble

Legendary member
8,394 1,170
The General said:
Ok, rephrase. You have 3 systems each with a historical probability of 60% (here we assume when we get a winner, we win 1 pound and when we lose, we lose 1 pound). Now, as defined by classical historical probability, all 3 systems have positive expectancy.

You need a way of determining which one system will give the best chance of a winner on the next trade. How do you do this ? All 3 have the same historical success rate ?
Whichever one has had the most recent longest string of contiguous losses.



The General said:
There is a fundamental flaw with classical probability which no-one here seems to have realised. This is of course linked to the discussion on backtesting.
You certainly know how to build tension. Come on General - what is it then? Any fundamental flaw with classical probability that the combined might of modern science has yet to discover (let alone us lowly life-forms here on t2w) has got to be a news-worthy event...!
 

The General

Active member
199 16
Champ,

Your answer is wrong I'm afraid and somewhat akin to the martingale roulette system - we have had 20 black come up so I will keep on betting on red.

In terms of what I believe is wrong with classical probability studies, if you can answer the question then you have got it. Think about it. Take it to the extremes and say that you HAD to choose between 3 systems each with a 2% chance of the next trade being a winner. If you think like that, it should become clearer.
 

mr_cassandra

Well-known member
349 36
Sounds like a question for a statistics teacher.

How would you apply this question and answer to your real world investing?

The General said:
Ok, rephrase. You have 3 systems each with a historical probability of 60% (here we assume when we get a winner, we win 1 pound and when we lose, we lose 1 pound). Now, as defined by classical historical probability, all 3 systems have positive expectancy.

You need a way of determining which one system will give the best chance of a winner on the next trade. How do you do this ? All 3 have the same historical success rate ?
 

mr_cassandra

Well-known member
349 36
mvp v21C sell short signal imminent

real-time postings at www.clearstation.com mdy, uipix boards, among others.

real-time results to date vs buy and hold posted at my web-site

I've come to find over my 4 years programming work, that only results matter, theories are just talk, until applied in real-life.

The General said:
Champ,

Your answer is wrong I'm afraid and somewhat akin to the martingale roulette system - we have had 20 black come up so I will keep on betting on red.

In terms of what I believe is wrong with classical probability studies, if you can answer the question then you have got it. Think about it. Take it to the extremes and say that you HAD to choose between 3 systems each with a 2% chance of the next trade being a winner. If you think like that, it should become clearer.
 

frugi

1
1,827 126
Ok, rephrase. You have 3 systems each with a historical probability of 60% (here we assume when we get a winner, we win 1 pound and when we lose, we lose 1 pound). Now, as defined by classical historical probability, all 3 systems have positive expectancy.

You need a way of determining which one system will give the best chance of a winner on the next trade. How do you do this ? All 3 have the same historical success rate ?

According to classical probability they all have an equal chance so it would not make any difference which one you chose. Take 3 coins, load them slightly so heads comes up 60% of the time, alwyas bet on heads, toss them at the same time and pick one at random without looking. Whichever one you picked would make no difference over the long term.

However trading systems are not strictly like tossing coins, because most of the successful ones are trend following. Thus I believe theBramble is right to find the one that has just suffered the largest number of consecutive losers, as trend becomes more and more likely as a consolidation matures. This is unlike the head/tails scenario where the chance of either is always 1 in 2 regardless of the result of the past x flips.

Therefore, if permitted, I would probably place a third of my capital in each and trade them simultaneously to smooth out the returns.
 

sidinuk

Established member
624 5
why can you only choose one system to trade? It makes more sense to trade all 3 and spread the risk
 
 
AdBlock Detected

We get it, advertisements are annoying!

But it's thanks to our sponsors that access to Trade2Win remains free for all. By viewing our ads you help us pay our bills, so please support the site and disable your AdBlocker.

I've Disabled AdBlock