Past performance good indicator for future performance?

marco

marco

Expert Member
Joined
Aug 3, 2006
Messages
2,902
Reaction score
124
Location
UK
Why do all financial websites state on the bottom "Past performance is not an indicator of it's future performance"?
I think they are just covering themselves if their stocks tank. This to me is a great method for investing and also a good method for trading as well.
What are your thoughts on this?


I don't advise this for new investors as this is a debate.
 
Personally I thought the last 20 pages of discussion about this were quite sufficient...

Also, IN BEFORE 90%+ and sharenet/sharechat.
 
They're right. Things can go wrong anytime because as you know markets are unpredictable at the best of times.
 
They have to say that, cause if they don't some people might expect past performance in the future (and might have legal issues later).
Is past performance a good indicator? I think it depends on if we talking asset managers or indexes or shares.
I would choose an asset manager based on their past performance and history, I won't pick one with an erratic past.
Shares are the hardest to pick if you ask me, therefore I just pick the index, and play a bit safer :)
 
Easy. If past performance is an an accurate predictor, what is there to predict? You would always be correct. You need to learn more about stats.
 
They likely put it there because it has statistically proven to be true since share trading and investing was invented.

Statistically it makes no sense. like saying a coin toss that comes out heads three times in a row is more likely to come out heads in the next toss. It isn't.
 
For the principle of predictably to make any sense there must be risk involved. It does not matter if you use past performance or any other criterion. Risk implies that your prediction could also be wrong or that what you predicted to happen does not happen. If someone tells me that they are able to predict future performance by using past performance, then the predictor is either a charlatan or has a superficial and dangerous grasp of statistical theory. There has not been a single example of any person able to make consistently accurate predictions before the fact. The share market is notorious for poor predictions. Every single peer reviewed research article about share predictability tells us that you cannot successfully make predictions based on past performance.
 
Lol, I thought you got 70% return per annum Marco. If that is the case you should be telling us this, not the other way around.

Also, when it comes to shares past performance has little bearing on future direction but if it does it often means a downward correction is coming as popular shares tend to get a little overbought when the hype hits. As was said earlier if you go managed funds though past performance is a good indicator, especially if it was maintained for prolonged periods.
 
Cius said:
Lol, I thought you got 70% return per annum Marco. If that is the case you should be telling us this, not the other way around.

Also, when it comes to shares past performance has little bearing on future direction but if it does it often means a downward correction is coming as popular shares tend to get a little overbought when the hype hits. As was said earlier if you go managed funds though past performance is a good indicator, especially if it was maintained for prolonged periods.
Click to expand...

That is why I'm asking. It is used as a trading strategy with up to 80% success rate. So if one combines it with fundamentals and future outlook of a company then it will work well for longer term investing with equal accuracy.
Although I have used this method for some time with success, a part of me is still unsure if it will carry on.
 
The problem is greed. For instance when a bubble arises, like we saw in property pre-2008, people think it can go forever and there is free money to be had not realizing that they are buying into seriously over priced assets. Same can happen with stocks. By the time a trend has been set on a share its probably already a little late as the smart money was already in ages ago and now the PE ratio's are above the long term average and waiting for a downward correction. Guys like Alan Gray made a lot of money by going against past performance trends and spotting companies that where unpopular but great value for money and getting into those shares before they corrected upwards. A good example was a little while ago Old Mutual had a lot of bad press about its life business and its share price tanked. Eventually the price of the Old Mutual shares was worth its Nedbank holdings essentially meaning if you bought Old Mutual you would be getting Nedbank exposure and Old mutual for free. Buying shares at R5 then was a great deal as they quickly returned back up to R10 a share. Anyways, its easy to spot in retrospect. Hard to see into the future. Hence I don't try stock pick and just buy ETF's and go for average market return at lowish risk. Its worked well so far for me with me getting about 5% above inflation growth.
 
I read about this system on a website some years ago and made notes but now I can't find the website.
Not quite Trend Following or Momentum trading. Perhaps a mixture of the two with fundamentals and future prospects thrown in.
It works with 75 to 80% accuracy and used for long term trading.
I started using it for investing due to trading "time decay" over longer periods. As from my notes, the gist of it is:

Look for liquid stocks that have a good uptrend history.
Good fundamentals to keep the uptrend going.
SL at -8% to ride out any dips. Sellers might dry up quickly and a few buyers will chase it back up even quicker due to no resistance on the backtrack.
Don't sell at new highs as all in profit now and happy traders will not become sellers.
As long as fundamentals are good, stay in.
Set SL to -8% trailing once in profit over 10%.
 
marco said:
That is why I'm asking. It is used as a trading strategy with up to 80% success rate.
Click to expand...

Yet another in a long line of bullschit marco stats. Traders do not trade on past performance and we have discussed this hundreds of times with you on this (and other) forums before. You're simply using this as a tool to start talking about yourself again...
 
I'm weary to venture into these types of threads because sometimes "financial advice" is given, which may be in contravention of FAIS.
 
Marco, please provide peer reviewed research to support your claim that this 'strategy' has a 75%-80% accuracy rate. Also what do you mean by accuracy rate?
 
DJ... said:
Yet another in a long line of bullschit marco stats. Traders do not trade on past performance and we have discussed this hundreds of times with you on this (and other) forums before. You're simply using this as a tool to start talking about yourself again...
Click to expand...

No I am absolutely not. This method is actually starting to worry me. Have a look at CML.
 
marco said:
No I am absolutely not. This method is actually starting to worry me. Have a look at CML.
Click to expand...

Marco, in another thread you predict CML will do 123% by year end, how do you predict that?
And now you seem worried?

I'm not gunning for you, just that you seem confident one minute and worried the next.
 
Last edited:
marco said:
No I am absolutely not. This method is actually starting to worry me. Have a look at CML.
Click to expand...

What happened to long term investing? There are many dips like this over the last three years so why is this one different?
 
This method of decision making is ultimately completely useless and is in fact destructive to your financial and mental wellbeing. Most, if not all statistical methods attempt to strip out fluctuations in an effort to create a smoother effect and in so doing create a seemingly smoother foundation from which to make a prediction of future performance. This effectively hides extremes. Here is the kicker: All techniques hide the fact that there is no actual performance or predicted performance that is symmetrical. Closer examination will show that performance is assymetrical with the left hand tail much more prominent than the right hand tail. This means that the effect of negative performance will be far greater than the equivalent effect of positive performance. If, like most investors, you are fragile to negative performance, a single major negative event will wipe you out. We don't see single major positive events that affect the market as a whole, but major negative events are seen.
You can provide for this quite easily but that is for another time if interested.
 
Disagree, jump-diffusion processes can, reasonably effectively, compensate for the "black swan" events and lead to a non-parametric, non-symmetric distribution. In any case, almost no useful models assume symmetry of asset prices or, where they do, it is compensated for after the fact e.g. volatility smiles/skews for option pricing (yes, yes, the assumption in black-scholes is that the log returns are symmetrical and not the asset price, before someone jumps in).

I think the general issue is that people do not understand that a model is not a predictor of future performance, models simply give a best guess for the average (or some other aggregate measure) going forward. The value of models is that they can help measure what might happen going forward as well as how likely an outcome is.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X