Trading Lessons/Insights From Coin Flipping

[tradingjournals]

I did a study of coin flipping from the perspective of trading. The aim was to gain some (new/alternative/existing) understanding of trading via an analysis of coin flipping. I am wondering whether others in here have gained similar insights and lessons that they learned from thinking and reading about coin flipping as a game. Both games share some peculiar common aspects.

Write your point of views, and what you know below.

Please avoid engaging in insults of others. I am stating this, because I was insulted few days ago by a person I never interacted with. It was very vicious, that I was tempted to reply or even take a more serious action. I also considered leaving this forum for good.

 

[walterjennings]

The only thing coin flipping can teach you about trading is that humans are easily fooled into thinking there are patterns in true randomness.

 

[masterjaz_99]

I too have pondered coin flipping as my current 'system' is basically a coin flip. I enter long/short at the open and exit the position at the close. I use a simple criterion to determine system direction. It runs basically 50-50, hence, the thinking about coin flipping.

There was a great thread from long ago that basically did the same as my system but used a coin rather than a criterion. The thread was a thought experiement on how to improve the 50-50 condition of the coin and the fact that given 50-50 odds, the system will be a net loser AFTER slippage, commissions, etc over a long sample period. (Found the thread: http://www.elitetrader.com/vb/showthread.php?threadid=179402&perpage=6&pagenumber=1 BUT the thread is heavy on insults, moderator intervention, and poor focus of participants).

Currently I am looking at how entries and how exits can be managed, without over massaging the data, in order to improve the 50-50 odds. Currently, I am looking at a couple ideas. 1) wait for a 1pt move against the direction at the open before entering (only to be concerned about missing the move), 2) holding for a point of additional profit after the close. Trading the NQ's, by the way...

So, while I have contributed nothing here, I will watch with diligence to see if anything interesting comes of this thread.

 

[circadian]

Quote from walterjennings:

The only thing coin flipping can teach you about trading is that humans are easily fooled into thinking there are patterns in true randomness.

Just don't underestimate the persistence of "random process."

 

[fickletrader]

Quote from walterjennings:

The only thing coin flipping can teach you about trading is that humans are easily fooled into thinking there are patterns in true randomness.

Not true at all.

You can set up an interesting, slightly positive expectation game like this:


* Start with 10 poker chips

* Stake as many of your chips as you like

* Flip 2 coins to make 1 of the following 4 combinations: HH, HT, TH, TT

* If either of the flipped coins is a tails (T), then the amount staked is lost

* If the outcome is HH, pay out 3 additional chips for each coin staked, and return the original bet to the bettor. (This last part makes it positive expectation)


This game will somewhat resemble actual trading equity swings as long as your trading rules have a slightly positive expectation and allow winning trades to be bigger than losing trades.

You can get creative trying variants, such as a fixed income of 1 chip per coin-flip iteration, or betting real $$ against your friends on who can reach a chip count first, and or last the longest, etc.

 

[walterjennings]

Quote from fickletrader:

Not true at all.

You can set up an interesting, slightly positive expectation game like this:


* Start with 10 poker chips

* Stake as many of your chips as you like

* Flip 2 coins to make 1 of the following 4 combinations: HH, HT, TH, TT

* If either of the flipped coins is a tails (T), then the amount staked is lost

* If the outcome is HH, pay out 3 additional chips for each coin staked, and return the original bet to the bettor. (This last part makes it positive expectation)


This game will somewhat resemble actual trading equity swings as long as your trading rules have a slightly positive expectation and allow winning trades to be bigger than losing trades.

You can get creative trying variants, such as a fixed income of 1 chip per coin-flip iteration, or betting real $$ against your friends on who can reach a chip count first, and or last the longest, etc.

E(X) = 0.75*(-X) + 0.25*(3*X) = 0. Not quite sure why you think it has positive expectation? Am I missing something? You cant count the return of the original bet as a gain in equity because it was already part of your equity. If you wanted to, you treat it like a fee to play, and return it in the winnings. E(X) = 1.00*(-X) + 0.25*(4*X) = 0 or Expectation = Fee + prob of winning * gain from winning

 

[intradaybill]

Quote from walterjennings:

The only thing coin flipping can teach you about trading is that humans are easily fooled into thinking there are patterns in true randomness.

There is no true randomness in the world since the universe started with some initial conditions and all motion obeys Newton's laws, unless you can prove to us otherwise. The universe is deterministic, and according to Einstein even the past, present and future coincide to a degree depending on speed and reference frame.

Marxists at some point in time tried to enforce this idea in universities that chaos was related to randomness and that everything became from this chaos. Today we know that specified complexity of the type found in human eye and other living organisms and parts of them thereof cannot be explained by chaos and randomness and it is due to careful and meticulous design. Yet, the Marxists cannot shallow that and continue with their randomness crusade. There can be no world without repeating patterns. They are all around you. Only a fool cannot see them. Actually, being able to see which patterns are relevant and which are not is directly related to intelligence.

 

[fickletrader]

Quote from walterjennings:

E(X) = 0.75*(-X) + 0.25*(3*X) = 0. Not quite sure why you think it has positive expectation? Am I missing something? You cant count the return of the original bet as a gain in equity because it was already part of your equity. If you wanted to, you treat it like a fee to play, and return it in the winnings. E(X) = 1.00*(-X) + 0.25*(4*X) = 0 or Expectation = Fee + prob of winning * gain from winning

Thanks for the correction, I was typing from memory and it's been a long time. The correct payout should be 4 chips per chip staked, plus the original stake returned to the bettor.

 

[walterjennings]

Quote from intradaybill:

There is no true randomness in the world since the universe started with some initial conditions and all motion obeys Newton's laws, unless you can prove to us otherwise. The universe is deterministic, and according to Einstein even the past, present and future coincide to a degree depending on speed and reference frame.

Marxists at some point in time tried to enforce this idea in universities that chaos was related to randomness and that everything became from this chaos. Today we know that specified complexity of the type found in human eye and other living organisms and parts of them thereof cannot be explained by chaos and randomness and it is due to careful and meticulous design. Yet, the Marxists cannot shallow that and continue with their randomness crusade. There can be no world without repeating patterns. They are all around you. Only a fool cannot see them. Actually, being able to see which patterns are relevant and which are not is directly related to intelligence.

To an educated theorist, the assumption of determinism is as much of a hail mary as the assumption of non-determinism. If fact in science, no theory can ever be proven, only disproven, and unfortunately, there is no way to disprove determinism or disprove non-determinism. But, recent research in the field of quantum physics leans towards the non-determinism side of the debate, where sub atomic particles seem to behave in a probabilistic manner, existing in infinite states at any given point in spacetime.

 

[walterjennings]

Quote from fickletrader:

Thanks for the correction, I was typing from memory and it's been a long time. The correct payout should be 4 chips per chip staked, plus the original stake returned to the bettor.

Then it does not teach anything about betting in a 50/50 game, merely betting in a game where the payouts are not fair for both parties.

Much like, I flip a coin, if its heads, you give me 2$, if its tails, I give you 1$, has positive expectation.