95% ASX100 success rate

[cynic]

A 95% winning average should deliver positive expectancy, i would reckon one would have to be pretty stupid to lose money with a 95% winning average, especially if the time frame was less than 3 months.

Any strike rate below 100%, can easily deliver negative expectancy!

Some of my past failures had greater than 97% strike rates!

 

[tech/a]

Any strike rate below 100%, can easily deliver negative expectancy!

Some of my past failures had greater than 97% strike rates!

Then I take my hat off to you.
That is a remarkable feat!

 

[cynic]

Then I take my hat off to you.
That is a remarkable feat!

It's a lot easier to achieve than it sounds!

My more successful strategies had a significantly lower strike rate, which is hardly surprising to those whom understand its irrelevance to positive expectancy.

 

[tech/a]

Still 3% losses to out weigh 97% wins
Good job!!!

 

[cynic]

Still 3% losses to out weigh 97% wins
Good job!!!

Cut your winners fast and let the losers run! It's so easy, even an expert could do it!

 

[skyQuake]

A 95% winning average should deliver positive expectancy, i would reckon one would have to be pretty stupid to lose money with a 95% winning average, especially if the time frame was less than 3 months.

I think it was implied the system would be martingale

 

[RobertoHood]

3 months on a week by week basis is 12 periods... 12!

I can flip 12 heads in a row. Across sectors doesn't mean too much as there is a lot of correlation. eg a breakdown system that spits out oil stock shorts day after day for the past 6 months.

Why not push your system back a few years and see how it fares?
Good luck

The funny thing about back testing is if you backtest too far , it smoothes out any edges that may have appeared and then gone, they get lost in the "noise" and are ignored due to being covered up by the overwhelming general randomness over extended periods.

As for your coin flip example, its more like like you flipping a coin 12 times with 100 different coins (1200 coin flips in total), and me being able to say with 95% certainty, which ones will be tails some of the time.

Sure you could possibly flip coins 12 times in a row , it is possible but unlikely , whats likely though , is the longer you flip a coin the more the head:tails ratio draws closer and closer to .5

 

[skyQuake]

The funny thing about back testing is if you backtest too far , it smoothes out any edges that may have appeared and then gone, they get lost in the "noise" and are ignored due to being covered up by the overwhelming general randomness over extended periods.

Then that's probably not an edge. Ideally you'd have something that you can backtest and curve fit to your heart's content, THEN backtest out of sample.

As for your coin flip example, its more like like you flipping a coin 12 times with 100 different coins (1200 coin flips in total), and me being able to say with 95% certainty, which ones will be tails some of the time.

Key difference here is that in the coin flip example there is no correlation between the different coins.
Whereas say CBA/NAB/ANZ/WBC 'flip' very similar.

 

[minwa]

This will be a good thread if you provide timely EXIT signals, to validate the 95%.

Anyway the picks are lagging the index for Monday, with average gain of .44% while the XAO100 gained .60%.

2 ended down for the day while 4 up - at what point is the unrealised loss enough to be taken and dumped into the 5% losers ? And how would it affect the success rate should it later recover back into profit ?

 

[RobertoHood]

This will be a good thread if you provide timely EXIT signals, to validate the 95%.

Anyway the picks are lagging the index for Monday, with average gain of .44% while the XAO100 gained .60%.

2 ended down for the day while 4 up - at what point is the unrealised loss enough to be taken and dumped into the 5% losers ? And how would it affect the success rate should it later recover back into profit ?

Minwa my friend , thank you for your reply ... but you are slightly incorrect.

At the moment only one stock is down , which is S32 , CTX , is up on my books it opened @ $32.73 and closed at $32.81 ending the day with a .244% gain. The calculation i believe you re working from is Fridays close compared to Mondays close which is a drop from $33.14 to $32.81. (-1.00%) , My systems works from BUYING Mondays open and selling at Fridays close.

So here is actually how the stock picks sit so far if you bought the OPEN yesterday, which can easily be achieved with any decent broker (unless with commsec where you can not put in a market order from Monday mornings)

CTY: +.244%
LLC: +2.75$
NAB: +1.277%
OSH: +1.464%
REC: +.429%
S32: -.885%

 

[RobertoHood]

Minwa my friend , thank you for your reply ... but you are slightly incorrect.

At the moment only one stock is down , which is S32 , CTX , is up on my books it opened @ $32.73 and closed at $32.81 ending the day with a .244% gain. The calculation i believe you re working from is Fridays close compared to Mondays close which is a drop from $33.14 to $32.81. (-1.00%) , My systems works from BUYING Mondays open and selling at Fridays close.

So here is actually how the stock picks sit so far if you bought the OPEN yesterday, which can easily be achieved with any decent broker (unless with commsec where you can not put in a market order from Monday mornings)

CTY: +.244%
LLC: +2.75$
NAB: +1.277%
OSH: +1.464%
REC: +.429%
S32: -.885%

So we are actually ahead of the market. All this is irrelevant though , the close on friday will tell the true story.. although some decent profits could be taken from the table if one wished.

To quote Nathaniel Rothschild - " Treat the stock market like a cold shower, quick in quick out"

 

[systematic]

To quote Nathaniel Rothschild - " Treat the stock market like a cold shower, quick in quick out"

I'm glad he said shower!

RobertoHood, could you clarify the exit you are using...as what you said above could be read two ways:

"Purchases are to be made at the open monday morning and sold prior to close Friday (14/08/15) "

I read that for its more obvious (to me) meaning: "just before the close on Friday afternoon but now I realise that this could also be covered with, "anytime before Friday's close.

Which is it, please?

And...best of luck with your experiment.

 

[RobertoHood]

I'm glad he said shower!

RobertoHood, could you clarify the exit you are using...as what you said above could be read two ways:

"Purchases are to be made at the open monday morning and sold prior to close Friday (14/08/15) "

I read that for its more obvious (to me) meaning: "just before the close on Friday afternoon but now I realise that this could also be covered with, "anytime before Friday's close.

Which is it, please?

And...best of luck with your experiment.

Ok Let me clarify.

The system predicts which stocks will be up for the week from Mondays OPEN to Fridays CLOSE.

At this stage , 95% percent of the stocks which i get BUY signals for are UP on the close of FRIDAY from MONDAYS OPEN.

Where the buyer who purchases the stocks decides to lock in profits during the week is up to them.

But the system has a 95 percent accuracy rate so far of selecting stocks which will end the week up from mondays open.

Here is an example of where someone may lock in profits.

Lets say you purchase WOW shares on the open of MONDAY because the system says BUY (meaning there is a 95% chance it will close the week up from the open), now lets say WOW is up 3.2% by wednesday , the buyer might have a rule where they wish to take 3% profits (or whatever % they may have) and decide to sell to lock in profits.

Now WOW could continue to rise and close the week up 5% or it could fall and close the week up .01% , or it could be one of the 5% stocks that close below the open on Monday. Its all at the buyers descretion and their personal risk management rules.

Hope this clears up your question ☺

 

[systematic]

Still 3% losses to out weigh 97% wins
Good job!!!

tech/a, it's not that difficult to imagine.

Take someone with a forex system. They (think) they've designed a good, "take quick profits" system.
Average 20 trades a day, 5 days a week - all positions closed for the weekend. Cost is 0.5 pip round trip.
They are aiming for a quick profit of 3 pips per trade.

They put in a "catastrophe" stop loss of 100 pips...to guard against being caught in a true market shift away from their position.

Initial testing showed promise. 3 months (or whatever) of testing showed that this 100 pip stop loss was hardly or never hit. End of week (closing out positions) is where we get the 3 losses from, and initial testing showed an average losing position of 50 pips. Great. The budding system trader is already dreaming of how they are going to be able to live on a steady 91 pip per week profit.* And even if the system makes only half of that...fine!

*97 trades make the 3 pip profit target = 291 pips. 0.5 pip round trip = 50 pips in costs for the week. Now we have 241 pips. The 3 losses that we close out on Friday average 50 pips per loss in testing. We have our 91 pip profit. At $25 per pip and trading 10 months per year...a 6 figure income awaits.



Now...is it difficult to imagine that our enthusiastic system trader finds out in the real world, over 12 months of trading...that the average loss of those 3 losing trades was in fact 85 pips, not 50 pips? Perhaps the 100 pip loss was hit just a few more times than was expected? Still a 97% strike rate. But those 3 losers only have to have an average loss of 80.3 pips to cause the system to lose money. Our now not-so-enthusiastic system trader loses $15,400 for the year. Side note: The trader still has an edge, in a cost free trading world. So their trading idea in and of itself is not a terrible idea.

 

[RobertoHood]

Still 3% losses to out weigh 97% wins
Good job!!!

This could happen but you have to be Homer Simpson e.g some who risks 100% of their capital on each and every trade with no stoploss. "Risk of ruin" comes to mind

 

[systematic]

Ok Let me clarify.

The system predicts which stocks will be up for the week from Mondays OPEN to Fridays CLOSE...

...Hope this clears up your question ☺

Totally. Thank you, it was the more obvious meaning. I thought it best to get this 100% clear to everyone now, rather than ambiguity ruin things later on.

Also, from your original post - there is a profit of 1.62% on the account based on, "buy Monday's open, sell Friday's close."
I take that from your comment, "With percentage gains averaging 1.62% per company per week"

So there is a positive edge from trading this on a simple Monday open to Friday close basis (assuming costs including slippage are contained within that).

I think some have slightly accused you of an ambiguity that is not there...just needed a bit more clarifying. You were simply saying that someone else might be able to do better than waiting until Friday close.

 

[systematic]

I was thinking about ideas for anyone who wanted to track the selections, and realised I needed to ask this first.

RobertoHood, without giving away your system...would you describe it as being based on trend/strength/continuation etc? Or would you describe it as mean reversion/short-term weakness etc?

Reason I ask is that it might help those who want to track it with different sell criteria.

For example, if it's based on current weakness, someone might like to try the, "first profitable close" exit (nicked from Rob Hanna, Cesar Alvarez, Nick Radge).

But if it's trend following / continuation in nature, that is not going to be a good exit to use! Here, you'd obviously want to do something to cut the losers and let the profits run.

I'm assuming, based on the strike rate, that you would describe it as weakness or reversion based.

 

[rnr]

Of the 23 buy signals i got last week, only 1 was not successful. Resulting in a -2.31% loss if you held until the close on friday, but profits were there during the midweek.
Robertohood

Ok Let me clarify.

The system predicts which stocks will be up for the week from Mondays OPEN to Fridays CLOSE.

Hope this clears up your question ☺

Below are the summarised results for the ASX100 for the week ending 7th August (last week):-

12 made gains from 0.31% to 6.82%
2 made no change
86 made losses from 0.14% to 18.96%

How do you arrive at the conclusion that 22 were profitable for that week?

 

[RobertoHood]

Still 3% losses to out weigh 97% wins
Good job!!!

Below are the summarised results for the ASX100 for the week ending 7th August (last week):-

12 made gains from 0.31% to 6.82%
2 made no change
86 made losses from 0.14% to 18.96%

How do you arrive at the conclusion that 22 were profitable for that week?

Sorry , let me reprhase of the 23 buy signals i got over the 12 WEEK PERIOD i tested , 22/23 closed the Friday positive.

My apologies

 

[Wysiwyg]

Below are the summarised results for the ASX100 for the week ending 7th August (last week):-

12 made gains from 0.31% to 6.82%
2 made no change
86 made losses from 0.14% to 18.96%

How do you arrive at the conclusion that 22 were profitable for that week?

Thanks. I'm sure I have not missed any system like this. Brokerage, time limit and high risk for small gains = failed consistency.