Said Bitar Daily Journal

[LEONARDZIR]

Said, a couple of questions about WTT. Your 11% return in September is pretty spectacular. I backtested WTT using Nick’s rules in his book and came up with a-1.5% loss in September for the WTT. Are you using different rules or did you optimize? Nick raised the issue of selection bias with the WTT. How have you worked around the bias?

[SaidBitar]

11% are based on closed positions so if you bactested it most probably you saw these 11% in another month.
as far as i remember i am using everything as is.
the thing is that there were many entry signals in this year and i was lucky with them
here are the stocks that gave me these results
CDE 155% after commissions
HL 90.7%
AMBR 61.5%

there are other 6 stocks but they are small losses or small winners, but the above 3 are the monsters

[SaidBitar]

here is single run from 2016 till today

on WTT

[JulianCohen]

Said Bitar wrote:

here is single run from 2016 till today

on WTT

And that’s on the Russell 3000 right? $1 minimum?

[LEONARDZIR]

Thanks Said. It looks like a pretty good run this year on my back testing as well.

[SaidBitar]

Julian Cohen wrote:

Said Bitar wrote:

here is single run from 2016 till today

on WTT

And that’s on the Russell 3000 right? $1 minimum?

yes
price limiters from 1 to 10

[LEONARDZIR]

Said, about the WTT. You mentioned that there were a lot of signals in the WTT this year. You also mentioned that you were “lucky” to have picked some really big winners and avoided at least one big loser. This raises the question of selection bias. Do you have any concerns about selection bias?

[LEONARDZIR]

I might add selection bias using the Russell 1000 with the price limiters

[SaidBitar]

Len Zir wrote:

Said, about the WTT. You mentioned that there were a lot of signals in the WTT this year. You also mentioned that you were “lucky” to have picked some really big winners and avoided at least one big loser. This raises the question of selection bias. Do you have any concerns about selection bias?

Not at all, because i have ranking mechanism that will tell me what to buy when there are more signals than positions required.
Also I trade R3000 so it is a large universe so even with price limiters still the universe is decent

[LEONARDZIR]

Said. Never thought about using a ranking to sort through the stocks..I have been hesitant to trade WTT because even with price limiters on Rusell 3000 there are still too many trades for the available capital.

[LEONARDZIR]

Said, how exactly do you include ranking in selection? For example let’s say you come to the end of the week and have capital to buy 10 stocks and the system generates 10 trades. Will you take all 10 trades or do you have ranking criteria that excludes some of them?
Also over the last 10 years what did you get for CAR and drawdown?

[SaidBitar]

If I have cash for 10 positions and i have 10 signals i will take all of them simply because this is what my backtesting did, and i think if you want to get similar results to your backtest trade exactly like it did .
If I have csh for 10 positions and i have 15 signals then i will check the ranking (positionscore) i posted about this few months before but the result is it doesn’t matter what you choose as ranking mechanism all are the same on the long run.

here is the result of backtesting for the last 10 years from 1/1/2006 until 12/31/2015 using commissions of (0.2% min of 25$).

if you use IB commissions the results will be looking much better.

[Nick Radge]

One thing you could look at Len is rank the universe first and select just the top 100 momentum stocks.

This becomes the new universe to then apply the WTT system.

You may find a much better risk/adjusted return with significantly lower selection bias issues.

[LEONARDZIR]

Thanks Nick and Said. I like the idea of ranking the universe and the top 100 momentum stocks as the tradeable universe

[ScottMcNab]

Nick Radge wrote:

One thing you could look at Len is rank the universe first and select just the top 100 momentum stocks.

This becomes the new universe to then apply the WTT system.

You may find a much better risk/adjusted return with significantly lower selection bias issues.

How often would you re-select the top 100 stocks Nick ? Would this be something done once a year and based over the returns over the last 3-5 years for example? Probably the selection of intervals isn’t the crucial part but rather implementing the concept in some form ? Should also work with MRV systems too I guess ? (lot of question marks in there)….
Thanks for posting

[Author]

Posts