Timing The Market Experiment

[Lucky C]

I tried a backtest of this, albeit limited and I'm sure not matching Nomad's algorithm especially since he doesn't spell out all the details. Tried at least a dozen iterations of buying high Sharpe ratio stocks tweaking different parameters.

Results varied wildly between negative ROI to almost keeping up with the S&P 500. I couldn't "data mine" a result that beat buy & hold, though if I kept working at it for several hours I might be able to get something that looks good in a backtest, but probably doesn't mean anything in a live situation. I would guess that implementing the rules (the way I set them up anyway) would result in lower returns than the market without benefiting from lower volatility. High Sharpe ratio is what investors go after to achieve a diversified lower-volatility portfolio, so it seems odd to me to chase a high Sharpe applied to individual stocks, where you'd be performance chasing stocks with high volatilities that are "hidden" by the even-higher numerator of the Sharpe ratio.

Again I'm sure Nomad's rules are different in some ways than the ones I tried, but generally it looks to me like high Sharpe ratio of an individual stock over the past X months is not indicative of outperformance in the following month.

[Nomad]

@alphaville
I am trading with cash money but only just over 1% of net worth at the moment (with this strategy).
Note this is my third and most controversial strategy.
As the experiment progresses this will increase. Could be 2-3% of net worth within a year.

[Nomad]

I tried a backtest of this, albeit limited and I'm sure not matching Nomad's algorithm especially since he doesn't spell out all the details. Tried at least a dozen iterations of buying high Sharpe ratio stocks tweaking different parameters.

Results varied wildly between negative ROI to almost keeping up with the S&P 500. I couldn't "data mine" a result that beat buy & hold, though if I kept working at it for several hours I might be able to get something that looks good in a backtest, but probably doesn't mean anything in a live situation. I would guess that implementing the rules (the way I set them up anyway) would result in lower returns than the market without benefiting from lower volatility. High Sharpe ratio is what investors go after to achieve a diversified lower-volatility portfolio, so it seems odd to me to chase a high Sharpe applied to individual stocks, where you'd be performance chasing stocks with high volatilities that are "hidden" by the even-higher numerator of the Sharpe ratio.

Again I'm sure Nomad's rules are different in some ways than the ones I tried, but generally it looks to me like high Sharpe ratio of an individual stock over the past X months is not indicative of outperformance in the following month.

I am not basing the Sharp ratio on consecutive prior months, the time difference is not equally spaced.
Yes, I realise that I am being somewhat vague on the algorithm details.
In my backtesting on portfoliovisualizer.com I was getting ridiculously high returns... one year was certainly above 100%.
However, massive volatility is what I see with this until you get a few good months to start the ball rolling.

[Alphaville]

@alphaville
I am trading with cash money but only just over 1% of net worth at the moment (with this strategy).
Note this is my third and most controversial strategy.
As the experiment progresses this will increase. Could be 2-3% of net worth within a year.

ah, nice, you can’t blow up that way then; even with the worst possible results you won’t owe anyone. it’s the price of learning, i suppose—we do pay closer attention when money is on the line.

have you read taleb’s “fooled by randomness”?

if you haven’t—you should right away. i’d post you quotes here, but they’re way too long

if you have read it—what do you make of it?

[Nomad]

@alphaville
I've back tested my algorithm from 2008 on a monthly basis to today, that is 142 steps.

The backtest massively out performs the Nasdaq and even all the individual companies that are used in the algorithm?

That says to me that the algorithm could have significant merit.
Hence the reason for the experiment with real cash.

It won't be plain sailing however, I am down about 9% or there abouts at the moment.

[Alphaville]

yes but... have you read him or no?

[Nomad]

@alphavilla
I have not as yet. I will probably read it but it may or may not apply to this...

[Alphaville]

@alphavilla
I have not as yet. I will probably read it but it may or may not apply to this...

oh hahahahaha it absolutely applies. that book is crazy. cheaper than the $500 loss.

i’m trying to post you a quote that doesn’t run afoul of the censors but it’s a bit complicated. there are too many, they’re too long. basically he says that track record means nothing when you blow up. he makes fun of a lot of people. i’m not qualified to make fun of anyone because i have little experience, but his book attacks many myths about trading i’ve heard about before everywhere. attacks? skewers, more like.

anyway by cash i didn’t mean “real money” as opposed to a simulation—i meant that you control shares equal to your funds invested, as opposed to a leveraged trade where you control a larger lot than you paid for, and are therefore working with debt issued to you by your broker. if it’s the 2nd, you could end up owing more than originally chose to invest.

[Alphaville]

here’s a short quote that explains what he’s going to do...,this is not the full reasoning obviously, just the introduction

In this chapter, I discuss some well-known counterintuitive properties of performance records and historical time series. The concept presented here is well-known for some of its variations under the names survivorship bias, data mining, data snooping, over-fitting, regression to the mean, etc., basically situations where the performance is exaggerated by the observer, owing to a misperception of the importance of randomness.

and here’s one of his zingers:

Some people carry their data mining activities into theology—after all, ancient Mediterraneans used to read potent messages in the entrails of birds.

the whole book of course is an exercise in skepticism and epistemology and a call for true scientific thinking when it comes to trading in the financial markets. of course he looks (and laughs) at faulty heuristics and biases that operate in the real world.

he’s brilliant really, and superentertaining. it’s very hard for me these days to finish a book, but with this one i just want to begin again hahahahahaa.

but seriously.

[Nomad]

@alphaville
Try this thought experiment regarding book makers.

They create odds for every football game based on the previous games.
So if you had to put a bet on / invest in a football team for the next game what would you do?

Anyone sensible would put their money on the team(s) in form.
Do team ALWAYS stay the best times? No, form comes and goes.
When a team loses form or another gains form, you would change your choices.

From my view point, passive investing is like betting on every team - even ones that are having a terrible season.
Active investing is putting money on the teams near the top of the table as per the thought experiment.
It is not a perfect analogy but you get the idea.
The approach I am using models this analogy and over time beats the index/average.

How could it not beat the index? Only if the good teams and bad teams continously cycle from good to bad in rapid succession.
Luckily on planet Earth - this does not happen for football teams or companies.

One of the funds I put money into is Fundsmith Global Equity and that is almost ten years old.
Originally it was £1 per unit and now a unit costs £5.32 or an annualised profit of more than 18% per annum after charges are deducted.
Incidentially, I started buying it 5 years ago when it was already a market leader.

So, from my perspective the naysayers and skeptics are flat wrong - stock picking can and is being done succesfully.


Q1) So, you read this brilliant book, has that book made you any money?
Q2) Does the book say 'low cost passive trackers' as the implicit subtext?

[Alphaville]

Q1) So, you read this brilliant book, has that book made you any money?
Q2) Does the book say 'low cost passive trackers' as the implicit subtext?

hahahaha! how...? i just finished it 2 days ago

and no, this book is not about stock picking for amateurs, but rather about the limits of knowledge, the place of cognitive biases in dealing with randomness, and how the randomness rules the financial markets, especially when it comes to highly unexpected events.

his examples are professional traders though, who are leveraged, hence they all can blow up by definition.

again, if you’re doing this purely on a cash basis, without any leverage, it’s a different game completely, because the odds are completely different too, given that you can’t go negative. worst possible case scenario you’d just go to 0. very hard for that to happen. a margin call is altogether a different story though.

since i don’t know which one is you case, i’m unable to offer anything on your strategy regardless of “track record”, plus there’s always the problem of induction + solon’s warning.

eta: as for the track record itself, and assuming cash and no risk of blowout, i’d say also there is no way to judge it without comparing vs 30 random selections from the same pool of stocks (30 selections, not 30 stocks) and also maybe the nasdaq 100 and the nasdaq composite?

ps- i’m not against stock picking, or your method of picking, or any of that.

[Nomad]

@alphaville
I am not sure that this book is relevant to what I am doing. When you talk about randomness, I don't see a lot of randomness?
I don't consider that Jeff Bozos is lucky at running a company or that Warren Buffet and Terry Smith are lucky at stock picking. They are
just very knowledgeable and skilled. People would not suggest that Paul McCartney is a lucky songwriter or Ronaldo is a lucky footballer.
Somehow, consistant success in most fields is put down to skill, but in stock picking it is put down to luck?

[Alphaville]

eh, don’t worry about it man