There are several very wrong schools of thought which seem popular related to simple probability.
1-"Two outcomes = 50/50", linked to the notion that "random things are always 50/50 or 33/33/33 etc" (aka, dictionary.com-itus)
This is wrong, probably comes from fair die and dice roll examples from school. As stated before, if you play the lottery (and we'll ignore modelling the balls with physics models etc etc). It is random chance whether you win or not. However it is certainly not a 50/50 proposition (unless you buy a lot of tickets . and even then you won't be profitable). Another example would be drawing a heart off the top off a full deck of cards, there are two outcomes, draw a heart, don't draw a heart, but the chance of doing it is 1/4.
I don't know if I can explain it any better than that, providing the maths would just cause further confusion, @Fat Tails or somebody else might have a better approach.
2-"Since you believe you can predict the odds of the very next trade, why do you take losing trades? Just skip those"
I don't even understand this train of thought tbh. Perhaps @Big Mike could expand on it. If something has a 60% chance of happening...it has a 60% chance of happening, that doesn't magically mean I know when it will happen and when it will not.
Again an example. If you roll a fair die, I can "predict" that you'll roll a 2-5 with 5/6 accuracy. I can say you will roll a 2-5 83% of the time. I can bet you that you'll roll a 2-5 every time you roll it, and be correct 83% of the time, thus having an edge if I get good enough odds. I'm not saying I'll know when you'll roll a 1 and won't bet you on that roll though, that's ludicrous. Nor does it mean that rolling a 2-5 on a fair die isn't a random event.
Not trying to have a go at anybody (hence why I quoted Mike, if he takes offence he can ban me), I just think this should be sorted out as I think a basic understanding will help people's trading.
Next time, Monte Hall
Dovie'andi se tovya sagain.
The following 4 users say Thank You to Hotch for this post:
I also would subscribe to the definition of Collins/Wikipedia. A random process is simply a process which is non-deterministic. The outcome of a non-deterministic process cannot be predicted, as it cannot be attributed to one or several causes.
A random process can be typically described by a probability distribution. This does not imply that the probability distribution is known. The definition "... process of selection in which each item of a set has an equal probability of being chosen" does give a general description of a random process, but just refers to the discrete uniform distribution.
I think it is just a joke. Trades are always non-deterministic, and hence the outcome of a single trade is unpredictable. Finding an edge means to identify a probability distribution which is skewed in your favor.
The following 5 users say Thank You to Fat Tails for this post:
One should note that probability quickly gets "complicated" and require an understanding of calculus, but finite/discrete probability is accessible for all.
I would, of course, recommend everyone to take advantage of the excellent math (and physics) courses at MIT's OpenCourseWare (http://ocw.mit.edu/courses/ocw-scholar/), but I understand that not everyone will be inclined to do so. It will be exciting to see if any of those will be offered through MITx this fall.
Last edited by Lornz; March 24th, 2012 at 01:29 PM.
The following 12 users say Thank You to Lornz for this post:
I'm no mathematician, but I think we are confusing 2 different concepts because we are using them interchangeably, Probability and Odds. Probability measures the chances for an event occurring against a total number of occurrences. Odds measure the chance of a singular event occurring as opposed to not occurring. Notice the words 'total number of occurrences' in the description for probability, and 'single' in the description for odds.
The odds of the single next coin flip to be heads or tails, the single next trade to be winner or loser is 1 to 1, some here have expressed it as 50/50.
The probability of the next coin flip to be heads based on the 'total' 1000 flips before it, the next trade to win or lose based on the 'total' 1000 trades observed before it, is variable depending on the distribution of heads/tails, or wins/losses observed in the last 1000 (total) occurrences.
So, single next trade considered by itself, odds 50/50. Single next trade considered in light of the last 1000 total historical trades, win probability 10%, 20%, 60%, 70%...etc.