Predicting randomness
- Thread starter oddiduro
- Start date
Quote from Equalizer:
I just love it when analogies are mis-used!
But tell me something Hank, is this really analogous? I mean, a coin flip is just that, some bored geezer tossing a coin. If it is a fair coin, and fair tossing method, and we assign a value of +1 for heads, and -1 for tails then the long term expected value is 0, i.e. 50% heads, 50% tails,
Price, on the other hand, can move by 0.25pts up or 0.75ps down, then gap up 2 points, etc, etc, and unfotunately, it isn't generated by some geezer tossing a coin, rather it is a consensus amongst a group of people, or computers programmed/operated by people, having different needs, different views, different utility and risk-profiles.
And lets add other info, like volume, transactions per unit time, volume per unit time, B-A-spread, etc.
"Ahh, yes" you might say, "but the ensemble effect appears to be random - don't we model prices by GBM?".
The key words being, "model", "appears".
The key questions being, "What is price", "what is random", "how can I profit from this"?
Something to think about...
hey equalizer,
thanks for the thoughts. however they are misapplied to my example. the flipping of the coin is analogous to price movement in a specific direction constituting a trend. trend is defined by those who believe that such things can be succesfully traded, generally by pointing at the large traders who claim to be trend followers, who have met with success, the positive black swans so to speak , within the distribution of traders, as--- higher highs and lower lows or the opposite for such a length of time so that profit can be realized. this is all contingent upon the entry being correct in the initial direction and the timing of the entry within the context of traders seldf imposed time frames.
what a run on sentence! but i think you know what i mean. allow another example--price moves 5 units up, will the next unit of time/price be up? does the 5 units of price in an increasing amount mean that the 6th unit of time/price be greater than the 5th unit? ofcourse not. heads/tails is the perfect analogy.
best,
surfer
To begin with, Price Oscillations must be specifically defined to be labeled correctly. Price must be viewed as well in a non-varying/specific volume bar or contract bar format generated by a data provider attuned to building the bars accurately and consistently.
Next provide the following needed information generated from "A" single chart, based on (1 bar = Average Daily Volume divided by 140) using the above parameters.
What was the previous extreme oscillation top in relation to the current price? (Be specific in relation to price oscillation outcomes)
What was the previous minor oscillation top in relation to the current price? (Be specific in relation to price oscillation outcomes)
What was the previous extreme oscillation bottom in relation to the current price? (Be specific in relation to price oscillation outcomes)
What was the previous minor oscillation bottom in relation to the current price? (Be specific in relation to price oscillation outcomes)
What was the sequential outcomes of the previous 5 or 6 extreme oscillations? (Be specific in relation to price oscillation outcomes)
What was the sequential outcomes of the previous 5 or 6 minor oscillations? (Be specific in relation to price oscillation outcomes)
Provide me the following information and I will answer your question with better than 90% accuracy. Without it and I agree with your head & tails analogy. Provide me the specifics of what I ask and your head & tails analogy flys out the window so fast it seemingly disappears.
If you do the work, the points I've requested will flawless plot out so no fudging.
Not only interesting but thought provoking as well. What makes the environment even more difficult is you've got program trading along with manual trading going on simultaneously. I have personally experienced what you pointed out albeit in different stocks. I've posted bids a dollar away from the inside bid and have been filled at lightning speeds - comparatively speaking - anywhere from a few seconds to a minute or thereabouts. Also the price in a lot of instances goes throuigh your price and keeps spiralling down - as you mentlioned - and the reverse for offers notwithstanding external market conditions.
In my opinion price is formed of signal plus random noise. The smaller the time frame you're charting, the lower is the signal to noise ratio, and the more random the price movements. To extract the price signal you need to filter out the noise on both axes, time and amplitude. TA indicators help you doing this, but introduce a lag. The smaller the time frame you're trading, the more you're trading the noise, and the more your results are affected by sudden random market reactions.