just a few days I notice the duckweed covers a pond I walk by on my
morning walk near my house in Tennessee. I just smile, and keep walking
by it every morning enjoying my walk to the top of the hill.
|Thank you for sharing your process and for sending me the photo.
Marriage as a
Way to Stay Busy
Thank you for your advice back in Jan. Yes, you are right, I only play
in a single deck. I find most single deck in north Nevada. I went to
Reno several times, Reno has the best rule. Western Village and Rail
City are my favorite. Also Boom Town, which is 15 miles west to Reno
also has single deck. In Vegas, I only find El Cortez in downtown has
single deck. Anyway, blackjack is not a good way to make money. You end
up driving around to avoid staying in a house too long. And there are
too few single deck.
Anyway, just an update to my current situation. I had a great ride in
Apr then a big drawdown afterward, and still down 38% from Apr high.
But most importantly , I made a commitment and got married in July.
That makes the drawdown period seems short. I normally distract myself
to other stuff, like fishing, video gaming, during long period of
drawdowns. Being married seems always keep me busy. That is a plus for
Again, thanks for providing a place to sharing with people interested
you for sharing your process.
Dear Mr. Seykota,
I'm 26, German-born and I live in Austria.
I'm working on the dynamic portfolio selection part of my stock trading
Everyday it ranks all US-Stocks based on how strongly/weakly they trend
- just like you show in the "Trends Tutorial" exercise.
The only thing that is missing from my dynamic portfolio selection
system is a way to filter out the thinly trading stocks.
So I implement a rule which filters out stocks whose average daily
cash flow is lower than a threshold (e.g: threshold = 2 000 000).
(Cash flow is calculated as an exponential lag of price * volume). This
cash flow threshold might work pretty well in 2016, but it might not
work that well when applied to stocks in 1950 or any other point in
time. I proceed to test the assumption that the cash flow of stocks
changes significantly through time.
So I calculate the cash flow (time constant = 100) for every currently
trading US-Stock and all the delisted ones as well. Then, for every
trading day from 1950 to 2016, I calculate the median (= 0.5 quantile)
of the cash flow distribution of all US-Stocks that are listed on that
and have a trading history longer than three times the time constant of
the exponential lag (3 * 100 = 300).
I have attached a figure showing the result. The plot indicates that
the cash flow of stocks varies significantly through time. So my
conclusions for dynamic portfolio selection are as follows: Using a
'fixed' cash flow threshold to filter out thinly trading instruments is
flawed. The threshold needs to be variable/dynamic as well.
One possible solution is to define today's cash flow threshold as the
p-th quantile of today's cash flow distribution (e.g.: p = 0.2). This
way each day [p*100] % of the stocks are filtered out.
There are probably many other solutions (and better ones). I'd be very
interested in your opinion on this matter.
Best regards from Austria,
you for sharing your research.
Perhaps you could determine cash flow threshold parameters by back
testing a portfolio that uses your filter.
In your Bernoulli treatise you mentioned that allowing one's hair to
grow out like Einstein's would somehow keep good thoughts from leaking
out of one's noodle.
For most of my life I've worn my hair in the skinhead style and have
noticed that I just can't seem to get it together.
Perhaps there really is something to this notion. Is this a
scientifically testable hypothesis? I sure could use an extra 50 points
or so in the IQ dept.
you for raising this issue.
I don't recall saying anything about thoughts leaking out; perhaps you
can tell me the page and
At any rate, the notion might deserve testing.
Your previous post about instrument universe to trade brought up some
feelings of fear or worry about being left behind.
I need a Tribe to follow or lead. Feelings of being left behind brings
confusion and worry ...
This brings up feelings of fear and anger. Im needing to catch up. I'm
waking from my slumber and I'm worried if it's too late. Do I go to
China, do I stay and try to make things work here. I have to catch up.
I need to be around people. I need a community to connect with and make
money and achieve goals, express myself, gain some independence
I have this thing about being late. Idk why. Late to school, late to
work. Late to wake up. Something I would also like to work on.
So the list of things I would like to work on.
-Staying connected with people and feelings.
-Waking up and getting to work early
-Meeting new people who can help me achieve my goals and vice versa
-Finding a tribe to run with
-Finding work to propel me forward
-Relearning to create, and express myself
-Share and receive feelings
-Learning how fish
-Being on time, making clear agreements and then keeping them
-Finding work and marrying [Name]
While I'm having these some of my thoughts are coming together. I've
been thinking a lot about yin and yang, space and the deep sea, light
and dark, pioneers and, I don't know what the opposite of a pioneer is.
I'm thinking about being left behind and how everything is relative.
When I was dealing with a lot of fear and anxiety when I first
contacted you, I think that's what it was, just being isolated and away
from my Tribe.
you for sharing your process.
Leap from the
I have always enjoyed this little part of a excellent old movie, I
believe the wisdom rings true in trading as well, "right livelihood,
and going with the trend," afterwards it is easy to see. Strange that
it is hidden so well from the "everyday eye."
Indiana Jones and the Last Crusade
you for sharing the link and your insights.
Hi, I'm a 23-Year-Old Trend Follower from Cape Town, South Africa.
If the trends I follow are on weekly or daily charts,
is my trying to take the trade on the hourly chart so that I can have a
tighter stop and thus more leverage just me not accepting that I can't
time the market?
Much Thanks in Advance,
you for raising this issue.
In general, tighter stops enable larger positions and more frequent
You might consider finding a sensitivity that suits you by back testing.
Note: stops do not guarantee a fill at your stop price so you might
consider including some execution slippage into your computations.
More on Start
Thank you so much for your reply and your willingness to share your
I really appreciate you taking the time to enlighten me. You have given
me much food for thought.
I love how you have named the various strategies and I notice you have
given the problem very careful, original and considered thought.
Thank you so much again.
Best wishes and warmest regards
you for acknowledging me and the work.
I find that the fear of feeling or appearing foolish can stop people
from thinking things through.
Some of the most inventive people I know enjoy the feelings of living
in the mystery and readily share the contours of their ignorance.
If someone shows you they have a better solution, you can move forward
by replacing yours with theirs.
I appreciate your answers.
Regarding point 4., two further questions arose in my head.
The system back tests included (other than the entry/exit rules and
1) volume filter (considering the relationship of the account size and
traded volume in the given markets)
2) constraints on allowable position number per sector in the portfolio
3) a maximum aggregate allowable risk
4) a maximum allowable margin.
All of the above constraints were determined by one or two discretional
value. As a consequence, a lot of entry signs were skipped by the
1) Do you consider a system with such discretional filters an
Removing all of the filters above, the system backrests give better
results (higher Bliss numbers), however more distinct ones (0% skid /
1.47 Bliss, 100% skid / 0.31 Bliss) - fees were also set this time to
0. In this case I get around 1900 trades (20 years, 58 products) -
around 2 trades per year per product (instead of one trade per one
product for every three years).
On the other hand - again without the filters - using longer term
entry/exit rules, the back test results are poorer, however not that
volatile (0% skid / 0.58 Bliss , 100% skid / 0.39 Bliss) - here getting
around 400 trades for the same portfolio/same 20 years. (one trade per
one product for every three years)
2) According to your experience, are these latter Bliss values still
rather volatile for a longterm system?
you for raising
1) You might ask your clients (perhaps including yourself) if they feel
the system has proper optimization and proper volatility.
You might consider taking your feelings about <establishing
with your clients> to Tribe as an entry point.
Thank you so much for your reply clarifying "Selection Bias" for me and
introducing me to the term "Start date Dependency".
I get it now. It is not the best term to use in the sense that all
trading systems have an intentional selection bias on what to trade and
when to trade.
I feel foolish, I shake my head and laugh at myself.
I am ever so grateful for you help to rid me of this confusing and
I like that you remind me to watch my language carefully.
You are the only person I have ever known that has such pinpoint
accuracy on dissecting and using language. I really like this trait of
your and I am really impressed by your language abilities.
On the concept of start date dependency, I wonder if you could share
any wisdom on how I might deal with the problem of start date
dependency while attempting to develop and evaluate a trading system.
Many thanks again.
continuing this issue.
Start-Date Dependency arises from trading a very large universe of
Shortly after inception, the portfolio fills up with instruments that
give the next signals; this soon precludes buying anything else.
Furthermore, many of the instruments stop moving shortly after
selection; they do not move higher and generate profits and they do not
move lower and exit the portfolio.
Thus, you wind up with a portfolio of stale non-moving instruments and
do not have resources to acquire fresh instruments.
You can address this situation in a number of ways:
1. Disclose the situation and live with it.
2. Pre-screen the universe of instruments to reduce it to what the
account can handle, eg.: hyper thrusters; fast food joints; stocks
beginning with the letters, "AM."
3. Assemble a long-term index fund.
4. Raise more money; here, the new customers wind up providing
liquidity for pre-existing ones - kind of a virtual Ponzi.
4. Include an aggressive purge strategy, to make room for new
a. Bulimic purge: get rid of everything on a
b. Fractional purge: get rid of a fraction of
on a regular basis.
c. Caribou purge: let the wolves have any members
herd that don't keep up with the others.
Let me know if you get any interesting results back testing these, or
Note: you don't generally see much about start-date dependency or purge
strategies or virtual Ponzi in the disclosure documents of major funds.
You might consider taking your feelings about <foolish>
I notice stocks that trend higher.
I notice some stocks are trend gently and have gentle ATRs, while
others trend to move violently with wild swings pulling this way and
that way with larger ATRs.
I imagine these with violent ATRs, like a fighting fish hooked on a
line, while more difficult to deal with, they might be more satisfying
I would like to fish for, and catch some stocks with strong trends and
large ATR characteristics.
I am considering somehow trying to normalize the trend for ATR in order
to rank the strength of trends among stocks and help with actual stock
selection, instead of simply using ATR just in position sizing and
I am curious if you have any views, suggestions or advice on this? I
seek your wisdom and guidance to help me shine some light on my path
- - - - -
I feel somewhat stupid.
I review your excellent TSP project work on Trends again. I experience
an aha! I sense an answer may simply be found within that exercise.
I now feel a solution to dealing with trend volatility may be that the
lag simply and cleanly takes care of the volatility by smoothing it. I
feel rather foolish now. I am still extremely interested to hear your
thoughts, just in case I learn anything else on the merits of
normalizing trends or otherwise.
Many thank again for putting up with me.
you for raising
In general, you make the most return from instruments that start out
with low volatility and then pick up volatility as the trend gets
You might like to take your feelings about <foolish> to
an entry point.
I enjoy reading FAQs and I continue to learn from contributors and your
Please keep up the good work. I try to move myself forwards and not
wallow in self pity or excuses.
I am very interested in your views, and I wonder if you might share
your wisdom, for dealing selection bias?
I wonder how you suggest one might deal with the problem of testing and
aligning a trading system and instrument universe with one's available
My understanding of selection bias is as follows.
Selection Bias: The problem that occurs when one has a large universe
of trading instruments, such as say constituents of Russell 2000 stock
index , and not enough capital available to take every instrument's
When the system is "switched on" The system will load up positions, by
selecting the first instruments that have a valid entry signal. Once
the total risk budget is consumed, no further trades can be taken until
further risk budget becomes available.
A system with this problem seems largely dependent upon the date of
switching it on, which makes it difficult to evaluate if system is
feasible or not.
The system might load up on the first twenty shares that come along and
then not be able to trade subsequent signals which makes the system
evaluation results highly variable.
I feel uneasy and not sure if I can trust the system results and I feel
unsure if the system is valid or not, as each time I picks a new start
date, I get a different result.
In short: Selection Bias is a problem that occurs when there are:
* Too many instruments to trade
* Too many signals to trade
* Not enough capital to take all of the trading signals for all of the
In my mind's eye, utilizing a radio analogy. I visualize my
grandfathers old radio. When I was a young boy I would sit in front of
it and change the tuner dial to find different radio stations.
I recall the dial round and round, turning it from one side to the
other. I would search through the static noise and find all the radio
stations, some I could tuned so I could clearly listen to the music,
while others would never be clear, no matter how much I finely adjusted
the tuner dial. I would repeat this exercise over and over, going
through the entire frequency range, from one side to the other, after
doing this a while I got to know which frequencies had the "good" radio
stations. Good being the clear radio stations with the best music. We
even marked these on the radio frequency car
I wonder if there is a way to do something similar with the trading
system so that I am left with fewer, but better trading signals for
fewer but better instruments.
I am not sure how to apply filters so that they discriminate against
"bad" instruments and "bad" signals, while not excluding the "good"
For example: I could reduce the instrument universe by trading only the
S&P500, but I wonder if perhaps there are instruments in the
Russell 2000 that could be much more profitable than trading just the
As I am struggling with the selection bias in my system testing result,
I seem to be stuck and can't even begin to understand how I could
evaluate the answer to this question.
Your thoughts appreciated
you for raising
I might suggest calling this issue "start-date dependency" rather than
A system without selection bias selects everything or nothing.
You might consider taking your feelings about <confusion>
Tribe as an entry point.
In the Tribe meeting I express my goal of less drama in my family
relationships, decrease medicinal eating and it also becomes my goal of
increasing self-love after identifying with another Tribe member.
I observe and identify the games and disconnect from incorrigible
family members playing destructive games.
I express my feelings when games are presented and I expose the
I buy a guitar and learn to play.
I communicate and share feelings with my son and I notice
progress. I still eat medicinally but less often. I
start a fund. I work on organizing capital to use for initial
performance record. I see progress financially over this
Thank you Ed for working with all of us in Tribe and me in particular.
I see incremental progress and I see I am a different person than when
I begin tribe.
you for sharing
Thank you, Ed.
I can not believe that I said the man, whom I consider a genius. You
are a man of a great country, where trading is developed. I'm a guy
from a small village [Name]. I live in the cold pole - Yakutsk.
Our trading industry is not developed. Of course, I realize that I need
more than indicators. I learned how to trade in futures - joining the
beginning of a trend. With the trend of trade I almost broke the record
of [Trader]. Made 8800% for 1.5 years.
I have no problem with discipline. Always comply with trade rules. I
want to share with you my ideas on trading. I have enough mathematical
training, but use in matstatistiku trade.
I found how to use matstatistiki could catch the bottom of the
investment during the financial crisis of 2008. I can not understand
why there the price has stopped? Some of the big players out there to
Using this tool, you could buy at the bottom for Sberbank, Gazprom,
VTB, Rosneft and other assets during the crisis of 2008, and then sell
at the top. It seems very big players know where to buy and sell using
mathstatistic. I think correctly?
Judging from the responses, you are very clever. Are you left-handed or
you for sharing
You might like to consider back testing your system to find optimal
values for position sizing and portfolio selection over a longer sample.
A system that buys in right at the bottom during 2008 might work really
well the next time we get to the bottom in 2008.
|To Top of Page