Viewing the market through f(x)
(binned per thread starter’s request) 200MA f(x) Trading
is it right by system? {image}
I am in that same short.
Viewing the market through f(x)
Once spreads tighten again after the close will be looking to get short NZDUSD. Set up looks really nice... {image} Need to free up some margin though so we will see...
Can almost see the 3rd leg of what most would call a 123 pattern getting ready to form. Shorting the pullback rollover with the trend is always a decent bet!
Viewing the market through f(x)
Space, Time & Ergodicity
When you first look at a price chart, you see that we are watching price evolve over 2 axes, price and time.
And as you watch the price trade, it appears that these 2 axes are linked and are moving together.
But… are they really?
When you step back and think about it, you will find that the 2 axes are completely different, in terms of their drivers.
The time axis moves along at whatever constant rate you arbitrarily selected while the price access is entirely driven by actual orders trading over the bid-ask spread.
For some, it seems trivial but it begs the question, do these differences have implications that we need to be aware of?
Yes, they do, especially in terms of gathering market statistics that you plan to trade on.
The biggest implication is, this means the process of trading can be considered non-ergodic under certain circumstances.
This means that what we measure in space, can be completely devoid from what we actually will most likely observe in time.
Something can be considered ergodic if the time average of an individual observation equals the average of the ensemble of observations.
An Example of Non-Ergodicity
Let’s consider we have a trading strategy that has a 50% win rate, but we must risk 40% of our account in order to make 50% on every trade.
Would you trade this system?
Let’s see, playing this game should net us a positive expected value of +5% of our equity for each trade.
EV% = (.5 * .5) – (.5 * .4) = 0.05
So mathematically, yes this game makes sense to play, and we should play as often as we can since we can expect to make a 5% return for each round we play, on average.
Unfortunately, the payoff, exposure of f(x) is actually non-ergodic, meaning that, even though this game makes perfect sense on paper, in space, it falls apart in time, and one will actually almost never see a profit all if they keep playing.
To demonstrate this, I have simulated playing this game for x rounds, 10000 times. x increases by 1 so we can tease out the variance to see what the payoff looks like if we were to play more than x rounds.
The first chart demonstrates what we are likely to see on average as a return, playing x rounds on the horizontal axis. This is what we should see in terms of measuring space.
The second chart demonstrates what we actually are most likely to see on average as a return playing x rounds.
The third chart is the probability of you observing a return that is closer to the space average, vs the time average. You can see that if you play long enough, almost no one makes money and the very few who do, make it all.
If this doesn't blow your mind you should. How could a system that yields a positive expected value of 5% per trade, lose almost 99% of the time over the long haul?
Because the process is non-ergodic and the likelihood of actually winning based on the probability of you stepping down a positive path or a negative one.
The catch is, once you are on that path, you can't ever get off it.
The losers never come back to be winners, the number of winners just gets smaller and smaller.
Making decisions based entirely on space measurements, with no regard to their time counterparts, is one of the deadliest things you can do.
When you first look at a price chart, you see that we are watching price evolve over 2 axes, price and time.
And as you watch the price trade, it appears that these 2 axes are linked and are moving together.
But… are they really?
When you step back and think about it, you will find that the 2 axes are completely different, in terms of their drivers.
The time axis moves along at whatever constant rate you arbitrarily selected while the price access is entirely driven by actual orders trading over the bid-ask spread.
For some, it seems trivial but it begs the question, do these differences have implications that we need to be aware of?
Yes, they do, especially in terms of gathering market statistics that you plan to trade on.
The biggest implication is, this means the process of trading can be considered non-ergodic under certain circumstances.
This means that what we measure in space, can be completely devoid from what we actually will most likely observe in time.
Something can be considered ergodic if the time average of an individual observation equals the average of the ensemble of observations.
An Example of Non-Ergodicity
Let’s consider we have a trading strategy that has a 50% win rate, but we must risk 40% of our account in order to make 50% on every trade.
Would you trade this system?
Let’s see, playing this game should net us a positive expected value of +5% of our equity for each trade.
EV% = (.5 * .5) – (.5 * .4) = 0.05
So mathematically, yes this game makes sense to play, and we should play as often as we can since we can expect to make a 5% return for each round we play, on average.
Unfortunately, the payoff, exposure of f(x) is actually non-ergodic, meaning that, even though this game makes perfect sense on paper, in space, it falls apart in time, and one will actually almost never see a profit all if they keep playing.
To demonstrate this, I have simulated playing this game for x rounds, 10000 times. x increases by 1 so we can tease out the variance to see what the payoff looks like if we were to play more than x rounds.
The first chart demonstrates what we are likely to see on average as a return, playing x rounds on the horizontal axis. This is what we should see in terms of measuring space.
The second chart demonstrates what we actually are most likely to see on average as a return playing x rounds.
The third chart is the probability of you observing a return that is closer to the space average, vs the time average. You can see that if you play long enough, almost no one makes money and the very few who do, make it all.
If this doesn't blow your mind you should. How could a system that yields a positive expected value of 5% per trade, lose almost 99% of the time over the long haul?
Because the process is non-ergodic and the likelihood of actually winning based on the probability of you stepping down a positive path or a negative one.
The catch is, once you are on that path, you can't ever get off it.
The losers never come back to be winners, the number of winners just gets smaller and smaller.
Making decisions based entirely on space measurements, with no regard to their time counterparts, is one of the deadliest things you can do.
Viewing the market through f(x)
Space, Time & Ergodicity When you first look at a price chart, you see that we are watching price evolve over 2 axes, price and time. And as you watch the price trade, it appears that these 2 axes are linked and are moving together. But… are they really? When you step back and think about it, you will find that the 2 axes are completely different, in terms of their drivers. The time axis moves along at whatever constant rate you arbitrarily selected while the price access is entirely driven by actual orders trading over the bid-ask spread. For some,...
?( 0.5-0.5)? what is first 0.5 ? and second 0.5?
(0.5-0.4) what is 0.5?
and what is the most o
important in this comment?
{quote} Uhh. I understand some things in your comment. Could write this in normal people lamguage
You have an expected value of 5% because you have a 50% chance of making 50% minus a 50% chance of making 40%.
Viewing the market through f(x)
- #60
- Jan 15, 2021 12:22am Jan 15, 2021 12:22am
{quote} Could write this in normal people lamguage
The first .5 is the probability of winning in an event.
The second .5 is what we get when we win.
The third .5 is the probability of losing.
The .4 is what we give up when we lose.
So the Expected Value is (the probability of winning X What we get when we win) minus (the probability of losing X what we give up when we lose).
~ Bill