Random trading systems - "monkey with a pin"

[DionysusToast]

The reasons they do worse than random are fairly obvious and well established.

I'm almost tempted to post the rumpled ones story about a rat in a t shaped maze, for once it would be appropriate.

I haven't seen it - I would love to though.

PM?

 

[DionysusToast]

Yes and so are you, so what is the point you are making?

My point is that you are making a prediction. If the computer is making a random decision to enter AND you are making money, then it is only because you are predicting something else with some degree of success.

And I will put $500 into the Red Cross fund for the Oklahoma Tornado victims if you can prove me wrong.

 

[Liquid validity]

What's to study about entering and exiting at the same price?

Unless you can actually do such a thing, by entering a short at the ask, the whole point is moot. Plus that is only one way of "getting the edge".

On the ES, joining the offer when you think it will hold almost always guarantees you will not get a fill. If there's 2K ahead of you and you see buyers hitting it, then you need them to keep refreshing after you enter and for buyers to keep hitting it but not that much that it moves through.

Not only is that predictive but it is a very narrow prediction and is akin to betting that not only will the market go down but that it will go straight down.

IMO - there are certain times where you can join but if you want to get to b/e very quickly, then you are much better to hit a market order into a level that is about to break.

I know you know this. Join the ask and hit out at the ask is a theoretical discussion, the hands-on trading reality is that this is an extremely narrow prediction that will play out a very small percentage of the time.

Well Ididn't actually say the ask/ask at the same price.
Do that and you are just filling yourself...and paying comms.

What I meant was short entry.
Sit on limit ask (offering liquidity) - get filled (OK successful limit ask fill without
level modification unlikely granted).
Get to exit price (profit or loss) exit with market ask.
You've paid no spread.

The point is its less risky than missing a limit fill on an exit just to earn the spread.
Unless you are working solely within the spread, paying no spread is
probably safer, if still far from easy.

I don't know about you, but I would much rather miss a fill on limit entry than exit.
Fair point that a market exit is prone to slippage, but then so is
a limit order that goes unfilled and then needs amending, or you
pull the plug with market order anyway.
I agree it isn't easy, and probably impossible for most anyway.

 

[cablemonster]

No, the distribution is from trading without any stop or profit target.

Risk management has the effect to narrow the bell curve. A stop loss would narrow it from the left, a profit target from the right.

It becomes really interesting when you do the opposite: not narrow but widen the profit distribution, using money management. You can then get straight rising equity curves just with random trading.

you have just touched on what most numpties dont realise in this game. most retail traders use a smallish stop and smallish profit target (relative to the instrument volatility), they think it is a 50/50 chance of winning but of course it isnt.

 

[Liquid validity]

I haven't seen it - I would love to though.

PM?

Same here, sounds interesting.

 

[Liquid validity]

My point is that you are making a prediction. If the computer is making a random decision to enter AND you are making money, then it is only because you are predicting something else with some degree of success.

And I will put $500 into the Red Cross fund for the Oklahoma Tornado victims if you can prove me wrong.

Yes, correct, I've always said that.
I have a random element, but the foundation is mean reversion.
Without that mean reversion and session timing element, I doubt it would work.
The random element replaces manual price action reading that may have identified
a prime intraday mean reversion opportunity.

 

[the hare]

I haven't seen it - I would love to though.

A rat was put in a T-shaped maze with a few morsels of food placed on either the far right or left side of the enclosure. The placement of the food is randomly determined, but the dice is rigged: over the long run, the food was placed on the left side sixty per cent of the time. How did the rat respond? It quickly realized that the left side was more rewarding. As a result, it always went to the left, which resulted in a sixty percent success rate. The rat didn't strive for perfection. It didn't search for a Unified Theory of the T-shaped maze, or try to decipher the disorder. Instead, it accepted the inherent uncertainty of the reward and learned to settle for the best possible alternative.

The experiment was then repeated with Yale undergraduates. Unlike the rat, their swollen brains stubbornly searched for the elusive pattern that determined the placement of the reward. They made predictions and then tried to learn from their prediction errors. The problem was that there was nothing to predict: the randomness was real. Because the students refused to settle for a 60 percent success rate, they ended up with a 52 percent success rate. Although most of the students were convinced they were making progress towards identifying the underlying algorithm, they were actually being outsmarted by a rat.


By your own admission, most traders will do worse than random. Given that's the case you have to ask why they wouldn't go for a higher probability method of randomly tossing a coin. I think the rumpled ones story goes some way to answering that question, or at least providing an awareness of how humans are programmed to behave.

 

[DionysusToast]

Well Ididn't actually say the ask/ask at the same price.
Do that and you are just filling yourself...and paying comms.

What I meant was short entry.
Sit on limit ask (offering liquidity) - get filled (OK successful limit ask fill without
level modification unlikely granted).
Get to exit price (profit or loss) exit with market ask.
You've paid no spread.

The point is its less risky than missing a limit fill on an exit just to earn the spread.

It's not less risky at all. It is only less risky in the theoretical world.

Enter at market - you need the market to go your way.
Enter at limit - you need the market to go against you a specific amount and then go your way.

Entering at limit is only less risky when the chance of the 2 scenarios above are the same.

Unless you are working solely within the spread, paying no spread is
probably safer, if still far from easy.

I don't know about you, but I would much rather miss a fill on limit entry than exit.

If I see 5000 offered and 300 bid, and that bid was 2000 30 seconds ago, I would MUCH rather enter with a market order.

In this scenario, your read is that the 300 will break. Joining the 5000 is foolish at this point.

 

[cablemonster]

The reasons they do worse than random are fairly obvious and well established.

I'm almost tempted to post the rumpled ones story about a rat in a t shaped maze, for once it would be appropriate.

please post it i need to brighten my day

EDIT - thanks read it now.

 

[DionysusToast]

Yes, correct, I've always said that.
I have a random element, but the foundation is mean reversion.
Without that mean reversion and session timing element, I doubt it would work.
The random element replaces manual price action reading that may have identified
a prime intraday mean reversion opportunity.

Ok - so you do not trade randomly.

You have a predictive system with a random component.

Calling it random trading is like calling me a playboy bunny because I have a set of rabbit ear deely boppers on.

 

[Liquid validity]

Ok - so you do not trade randomly.

You have a predictive system with a random component.

Calling it random trading is like calling me a playboy bunny because I have a set of rabbit ear deely boppers on.

Well that would depend on whether you believe a PC detects mean reversion on
a predictable or random basis...
I prefer to call it semi random as a PC is crap at reading price.
Each to his own regarding the bunny suit, does the missus know or is it hers anyway

 

[DionysusToast]

A rat was put in a T-shaped maze with a few morsels of food placed on either the far right or left side of the enclosure. The placement of the food is randomly determined, but the dice is rigged: over the long run, the food was placed on the left side sixty per cent of the time. How did the rat respond? It quickly realized that the left side was more rewarding. As a result, it always went to the left, which resulted in a sixty percent success rate. The rat didn't strive for perfection. It didn't search for a Unified Theory of the T-shaped maze, or try to decipher the disorder. Instead, it accepted the inherent uncertainty of the reward and learned to settle for the best possible alternative.

The experiment was then repeated with Yale undergraduates. Unlike the rat, their swollen brains stubbornly searched for the elusive pattern that determined the placement of the reward. They made predictions and then tried to learn from their prediction errors. The problem was that there was nothing to predict: the randomness was real. Because the students refused to settle for a 60 percent success rate, they ended up with a 52 percent success rate. Although most of the students were convinced they were making progress towards identifying the underlying algorithm, they were actually being outsmarted by a rat.

Like it.

By your own admission, most traders will do worse than random. Given that's the case you have to ask why they wouldn't go for a higher probability method of randomly tossing a coin. I think the rumpled ones story goes some way to answering that question, or at least providing an awareness of how humans are programmed to behave.

Because randomly tossing a coin, without any other predictive element will cause you to do nothing more that pay the spread and commissions.

I do agree that over-thinking it will lead to excessive losses. The problem is you seem to be saying that the flip side of the coin is that not thinking at all will lead to gains. It will not. Not thinking AKA random trading will, over time lead to you paying the spread plus fees.

Randomly entering a directional trade will leave you to decide which trades to cut and which trades to run. This is a skill. Some will be better than others. It is the absolute opposite of random.

Hence all those people pouring billions into HFTs and not just going out and buying a shiny new coin.

 

[Liquid validity]

It's not less risky at all. It is only less risky in the theoretical world.

Enter at market - you need the market to go your way.
Enter at limit - you need the market to go against you a specific amount and then go your way.

Entering at limit is only less risky when the chance of the 2 scenarios above are the same.



If I see 5000 offered and 300 bid, and that bid was 2000 30 seconds ago, I would MUCH rather enter with a market order.

In this scenario, your read is that the 300 will break. Joining the 5000 is foolish at this point.

For you, and the way you trade, I completely agree.
You look for micro entry with the flow.

Mine quite frequently takes trades against the micro flow, so what you say is
not an issue in that instance.

 

[new_trader]

A rat was put in a T-shaped maze with a few morsels of food placed on either the far right or left side of the enclosure. The placement of the food is randomly determined, but the dice is rigged: over the long run, the food was placed on the left side sixty per cent of the time. How did the rat respond? It quickly realized that the left side was more rewarding. As a result, it always went to the left, which resulted in a sixty percent success rate. The rat didn't strive for perfection. It didn't search for a Unified Theory of the T-shaped maze, or try to decipher the disorder. Instead, it accepted the inherent uncertainty of the reward and learned to settle for the best possible alternative.

The experiment was then repeated with Yale undergraduates. Unlike the rat, their swollen brains stubbornly searched for the elusive pattern that determined the placement of the reward. They made predictions and then tried to learn from their prediction errors. The problem was that there was nothing to predict: the randomness was real. Because the students refused to settle for a 60 percent success rate, they ended up with a 52 percent success rate. Although most of the students were convinced they were making progress towards identifying the underlying algorithm, they were actually being outsmarted by a rat.


By your own admission, most traders will do worse than random. Given that's the case you have to ask why they wouldn't go for a higher probability method of randomly tossing a coin. I think the rumpled ones story goes some way to answering that question, or at least providing an awareness of how humans are programmed to behave.

The 'rat' experiment is not an analogy to trading! Making observations of the behaviour of things and then trying to determine whether there is definite cause and effect makes trading analogous to scientific studies like Physics. Observation and experiment, even trial and error can lead to predictable outcomes. I wouldn't be able to trade if markets were random. A good understanding of supply and demand is what you need, I would even go so far as to say it's ALL you need to know. The effect of these two forces on the market is observable even if at first you don't know how to interpret it. This takes lots of study and practice.

 

[DionysusToast]

Well that would depend on whether you believe a PC detects mean reversion on
a predictable or random basis...

It will be based on rules you defined and therefore is not random. It relies on specific market conditions. I can give you a completely random exit generator if you want to go full retard, I mean full random...

Tropic Thunder (5/10) Movie CLIP - Never Go Full Retard (2008) HD - YouTube

I prefer to call it semi random as a PC is crap at reading price.
Each to his own regarding the bunny suit, does the missus know or is it hers anyway

Just the ears...

 

[Liquid validity]

Because randomly tossing a coin, without any other predictive element will cause you to do nothing more that pay the spread and commissions.

I do agree that over-thinking it will lead to excessive losses. The problem is you seem to be saying that the flip side of the coin is that not thinking at all will lead to gains. It will not. Not thinking AKA random trading will, over time lead to you paying the spread plus fees.

Randomly entering a directional trade will leave you to decide which trades to cut and which trades to run. This is a skill. Some will be better than others. It is the absolute opposite of random.

Hence all those people pouring billions into HFTs and not just going out and buying a shiny new coin.

You've already mentioned in this thread one reason a random entry can work:
http://www.trade2win.com/boards/tra...trading-systems-monkey-pin-6.html#post2132824
A trailing stop.
That is extremely crude, it does demonstrate, as I've said numerous times,
that random alone does not work.
Its all about running winners and keeping losses in check.
Trade randomly and cut your winners and run your losses proves random alone
is not enough.

 

[DionysusToast]

For you, and the way you trade, I completely agree.
You look for micro entry with the flow.

Mine quite frequently takes trades against the micro flow, so what you say is
not an issue in that instance.

Then why enter a limit at a specific price? Why not let it flow 3 ticks against you and enter there?

Your entering at limit argument only counts if the market moves your way from that point - otherwise my micro entry wins by a mile.

An entry short at limit 1668.50 is absolutely worse than my market order short 5 minutes later at 1670.

It is ONLY at the micro level that it makes a difference.

 

[Liquid validity]

It will be based on rules you defined and therefore is not random. It relies on specific market conditions. I can give you a completely random exit generator if you want to go full retard, I mean full random...

Come on your just here for the lulz now aren't you
Its been said plenty of times that random entry has to be coupled with a tightly
defined and quite likely complex exit strategy.

 

[the hare]

The problem is you seem to be saying that the flip side of the coin is that not thinking at all will lead to gains. It will not. Not thinking AKA random trading will, over time lead to you paying the spread plus fees.

100% agreed. However, I advocate random trading for a number if reasons

1 your probably better off than following the usual garbage promoted in books and forums

2 it gets you to understand the random distribution of gains and losses, and if you approach the development sensibly, there's definitely lessons worth learning

3 it focuses you to think about trade management and exit.

4 its quite a nice tool for highlighting the true nature of an edge, and cutting through the general bs about the magnitude of edges that are required or available

By the time you have 1,2 and 3 sorted out you generally don't need much more.

 

[DionysusToast]

You've already mentioned in this thread one reason a random entry can work:
http://www.trade2win.com/boards/tra...trading-systems-monkey-pin-6.html#post2132824
A trailing stop.
That is extremely crude, it does demonstrate, as I've said numerous times,
that random alone does not work.
Its all about running winners and keeping losses in check.
Trade randomly and cut your winners and run your losses proves random alone
is not enough.

Correct - and it takes a lot of skill and experience to know when to cut a trade.

A fixed set of rules about exiting trades applied in all conditions will just see you pay the spread.

It is only your predictive skills - about how far a move will go OR when a move is over that will see you end trades at a good location.

If using a fixed set of rules, it is only your predictive skills that will see you enabling the system when you predict that complimentary market conditions will prevail.

And if you have those predictive skills - you should put them to use at your entry point if trading outright positions.