You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Predicting randomness
- Thread starter oddiduro
- Start date
Hi Jesse,
By chance I do have that same book. As I used to think like you do, I did some basic research on chaos theory, and the markets before trading this Chaotic Indicator.
Check this paragraph from that same book.
"On the surface, a chaotic system behaves much like a random one. Think about computer models of the Earth's weather system, which are chaotic and so suffer from the butterfly effect. Run the computer model starting from some chosen state, and you get a pleasant, sunny day a month later. Run the same computer model starting from some chosen state plus one flap of a butterfly's wing surely an indistinguishable state in any conceivable practical experiment - and now you get a blizzard. Isn't that what a random system does? Yes but the timescale is wrong. The 'randomness' arises on large timescales - here months. The distinction between determinacy or randomness takes place on short timescales; indeed it should be immediate. After a day that flapping wing may just alter the local pressure by a tenth of a millibar. After a second, it may just alter the local pressure by a ten billionth of a millibar. And indeed in the computer models that's just what happens. It takes time for the errors to grow and we can quantify that time using the Liapunov exponent. So we can safely say that on short timescales the computer model of the weather is not random: it is deterministic (but chaotic). "
I believe what he is trying to say is that on short term, in this case the wheather is predictable but on long term, the error grow exponentially.
Now going back to the markets, and the model I'm trading, check Fridays prediction.
This is the prediction made by the system at 11:15, in black line, and what finally happened in Red.
<img src="http://www.tradingpro.com/images/charts/20060901-NASDAQL.gif">
Clearly, this is not a 2 o 3 months prediction, but it is OK for intraday trading. In this case, I went LONG till 1:00. And then SHORT till 3:00. I had some heat to suffer at 1:00 but nothing to worry about.
Now this is another example from Friday. This is the Russell 2000 prediction, displayed at 12:00. This was the prediction for a 4 hours market path.
<img src="http://www.tradingpro.com/images/charts/20060901-RUSSELL2000L.gif">
I'm far from being a Chaos expert, but I read some books on the theory and how it is applied to market predictions. I care less what the inner workings of these guys black box is, as far as I can make money with it. From my experience their model DO predict the market, and I've seen it many many times.
I've logged their trades during some months at this Journal.
http://www.elitetrader.com/vb/showthread.php?s=&threadid=70418&perpage=6&pagenumber=1
They recomend some books at
http://www.tradingpro.com/Research.html
If you are interested in doing some research I would suggest specially the first
http://www.amazon.com/Predictors-Th...=pd_bbs_1/102-5590381-8988955?ie=UTF8&s=books
and
http://www.amazon.com/Chaos-Order-C...=pd_bbs_1/102-5590381-8988955?ie=UTF8&s=books
I've also seen a film, that can give you a quick inside of the theory.
http://www.amazon.com/Bank-Robert-C...ef=pd_bbs_1/102-5590381-8988955?ie=UTF8&s=dvd
I generaly trust more books, than Films, but anyway the technology described in The Bank is very similar to what PredictionLabs offers.
Take care.
By chance I do have that same book. As I used to think like you do, I did some basic research on chaos theory, and the markets before trading this Chaotic Indicator.
Check this paragraph from that same book.
"On the surface, a chaotic system behaves much like a random one. Think about computer models of the Earth's weather system, which are chaotic and so suffer from the butterfly effect. Run the computer model starting from some chosen state, and you get a pleasant, sunny day a month later. Run the same computer model starting from some chosen state plus one flap of a butterfly's wing surely an indistinguishable state in any conceivable practical experiment - and now you get a blizzard. Isn't that what a random system does? Yes but the timescale is wrong. The 'randomness' arises on large timescales - here months. The distinction between determinacy or randomness takes place on short timescales; indeed it should be immediate. After a day that flapping wing may just alter the local pressure by a tenth of a millibar. After a second, it may just alter the local pressure by a ten billionth of a millibar. And indeed in the computer models that's just what happens. It takes time for the errors to grow and we can quantify that time using the Liapunov exponent. So we can safely say that on short timescales the computer model of the weather is not random: it is deterministic (but chaotic). "
I believe what he is trying to say is that on short term, in this case the wheather is predictable but on long term, the error grow exponentially.
Now going back to the markets, and the model I'm trading, check Fridays prediction.
This is the prediction made by the system at 11:15, in black line, and what finally happened in Red.
<img src="http://www.tradingpro.com/images/charts/20060901-NASDAQL.gif">
Clearly, this is not a 2 o 3 months prediction, but it is OK for intraday trading. In this case, I went LONG till 1:00. And then SHORT till 3:00. I had some heat to suffer at 1:00 but nothing to worry about.
Now this is another example from Friday. This is the Russell 2000 prediction, displayed at 12:00. This was the prediction for a 4 hours market path.
<img src="http://www.tradingpro.com/images/charts/20060901-RUSSELL2000L.gif">
I'm far from being a Chaos expert, but I read some books on the theory and how it is applied to market predictions. I care less what the inner workings of these guys black box is, as far as I can make money with it. From my experience their model DO predict the market, and I've seen it many many times.
I've logged their trades during some months at this Journal.
http://www.elitetrader.com/vb/showthread.php?s=&threadid=70418&perpage=6&pagenumber=1
They recomend some books at
http://www.tradingpro.com/Research.html
If you are interested in doing some research I would suggest specially the first
http://www.amazon.com/Predictors-Th...=pd_bbs_1/102-5590381-8988955?ie=UTF8&s=books
and
http://www.amazon.com/Chaos-Order-C...=pd_bbs_1/102-5590381-8988955?ie=UTF8&s=books
I've also seen a film, that can give you a quick inside of the theory.
http://www.amazon.com/Bank-Robert-C...ef=pd_bbs_1/102-5590381-8988955?ie=UTF8&s=dvd
I generaly trust more books, than Films, but anyway the technology described in The Bank is very similar to what PredictionLabs offers.
Take care.
Quote from steve46:
If on the other hand we have what is statisticians call a "fair coin" equally weighted on both sides. Then I am still going to bet on heads. Although the flips are independent, there is no reason to rule out a streak of heads, and in practice (I have already done this) streaks of 5 to 10 in a row are not unusual..Depending on the odds offered me, I will take that bet IF the risk to reward ratio is attractive. In other words it is all about evaluating the risk to reward vs the odds that a run of heads will continue...
Every loser at a roulette table has tried that one, cant believe you are so naive
They wait for black to come up 5 times in a row then bet on red, guess what their strike rate is? yep 50%, and they get paid 48.5% because of the zero (house number).... bye bye Steve
Sorry Dano:
Your in the wrong neighborhood..These folks are still working on basic stats...Your going to have to wait a few years to get your point across...LOL
Back to calculating the odds of a coin toss (for a "fair coin")
Each toss has 1/2 odds...1/2 heads, 1/2 tails
Odds of getting heads on two tosses = 1/2 ^ 2
Odds of heads on 5 tosses = 1/2^5 (2 to the 5th power) or
.03125 or 3.125%
Not great odds, so I would like to get a nice payday if I hit that bet..and I am going to bet very conservatively if at all.
But that is not the actual bet..
Since each flip is independent of the last (or previous) flip, when we get to the fourth and fifth I am still going to bet on heads....Remember you've already "given me" three heads in a row as the starting condition...
So the actual bet is "will the next two flips come up heads"
and the odds of that happening are
12.5%
Much better odds and I will take that bet...
Your in the wrong neighborhood..These folks are still working on basic stats...Your going to have to wait a few years to get your point across...LOL
Back to calculating the odds of a coin toss (for a "fair coin")
Each toss has 1/2 odds...1/2 heads, 1/2 tails
Odds of getting heads on two tosses = 1/2 ^ 2
Odds of heads on 5 tosses = 1/2^5 (2 to the 5th power) or
.03125 or 3.125%
Not great odds, so I would like to get a nice payday if I hit that bet..and I am going to bet very conservatively if at all.
But that is not the actual bet..
Since each flip is independent of the last (or previous) flip, when we get to the fourth and fifth I am still going to bet on heads....Remember you've already "given me" three heads in a row as the starting condition...
So the actual bet is "will the next two flips come up heads"
and the odds of that happening are
12.5%
Much better odds and I will take that bet...
Quote from trader28:
Every loser at a roulette table has tried that one, cant believe you are so naive
They wait for black to come up 5 times in a row then bet on red, guess what their strike rate is? yep 50%, and they get paid 48.5% because of the zero (house number).... bye bye Steve
Sorry this is entirely different. READ and think before you post.
This is not about a single event. It is about a setup that the original poster presented..
Apparently for you thinking has nothing to do with it
Again for those who may be a little slower on the uptake.
Can one have a run of 5, without having a run of 1, 2, 3, and 4?
The answer is NO
If one were to use a monte carlo engine to get a result you would over the long run see a reversion to 1/2 odds as statistical laws suggest. Over the shorter run however, ANYTHING can happen, BUT again that run of 5 heads CAN'T HAPPEN UNLESS THE FIRST FOUR FLIPS ARE HEADS...
So Einstein, in the gentleman's example the first three flips are heads..Most of the work has been done for me.. I have a 12.5% chance of getting two more heads. If I flip the coin again and it comes up heads, the odds of my winning the bet jump to 50% as YOU pointed out...and again on a short term basis I know that runs of 5 to 10 heads happen with regularity...I will take that bet
Again for those who may be a little slower on the uptake.
Can one have a run of 5, without having a run of 1, 2, 3, and 4?
The answer is NO
If one were to use a monte carlo engine to get a result you would over the long run see a reversion to 1/2 odds as statistical laws suggest. Over the shorter run however, ANYTHING can happen, BUT again that run of 5 heads CAN'T HAPPEN UNLESS THE FIRST FOUR FLIPS ARE HEADS...
So Einstein, in the gentleman's example the first three flips are heads..Most of the work has been done for me.. I have a 12.5% chance of getting two more heads. If I flip the coin again and it comes up heads, the odds of my winning the bet jump to 50% as YOU pointed out...and again on a short term basis I know that runs of 5 to 10 heads happen with regularity...I will take that bet