Random Walk Theory

[temptrader]

Why would I care about what you say anyway? You make idiotic assertions like "a market is nothing but the sum of it's constituting participants, each of whom has his own agenda, doing his own thing." and almost in the same breath say

"Next time someone tells you they have a great new system thats gonna beat the markets just give Gamma Jammer a call and tell him to sell one billion worth of EUR/USD while watching your friends pipe dream go up in smoke."

So, could you sell billion worth of EUR/USD ? I couldn't. So, what does that tell you about all the participants?

Let me ask you something, what would you do if Gamma Jammer called you and said he was going to sell one billion worth of EUR/USD in the next 5 minutes? :whistling

LOL!!!! There was a chapter in Pit Bull by Marty Schwartz (I don't actually like the guy just in case you are wondering, think his character is utter reprehensible), where he meets Paul Tudor Jones and other traders and they got round and talked about oil. The next day or so something is acting funny on the oil prices, and at the end Marty realised it was because what was going in the meeting before.:cheesy:

Hey, BSD, you're out of line.

 

[dbphoenix]

terribly sorry, must have been dyslexic there for a minute.

No problem. We're all subject to the occasional brain fart.

 

[BSD]

LOL!!!! There was a chapter in Pit Bull by Marty Schwartz, where he meets Paul Tudor Jones and other traders and they got round and talked about oil. The next day or so something is acting funny on the oil prices, and at the end Marty realised it was because what was going in the meeting before.:cheesy:

Yup, Marty said at the get together that he was bullish on oil.

That night, which was the evening traders dinner at the then Commodity Corp that was eventually taken over by Goldman Sachs, oil was at $12.50 / barrel.

During dinner Marty gave his bullish assessment on oil to all present.

There was like 50% of the then hedge fund money at the time in that room, what with people like Tudor Jones, Bruce Kovner and Michael Marcus attending.

Next day, immediately upon the open, oil started climbing all the way up to $15 / barrel that day.

No problem. We're all subject to the occasional brain fart.

 

[firewalker99]

Perhaps if you guys were to explore the difference between randomness and unpredictability? :whistling

Often a system/event/phenomenon will show a characteristic randomness on a microscopic level, but predictableness on a macroscopic level.

If I throw a dice, I have no idea what it will be (random). But if I throw 1000 dices, I can predict (to a certain degree) how many 6's I'll end up with.

 

[arabianights]

Sure, but you wouldn't be able to tell what the next coin flip would be. A random walk doesn't mean every price is random, it means every price movement is random, although ultimately following a distribution, usually gaussian.

One of the stupidist things that academics who believe in random walks on the markets do is to de trend them, but then give no justification for the trend itself! You can come up with mathematical justifications for doing this but they are ultimately self defeating.

 

[new_trader]

Often a system/event/phenomenon will show a characteristic randomness on a microscopic level, but predictableness on a macroscopic level.

If I throw a dice, I have no idea what it will be (random). But if I throw 1000 dices, I can predict (to a certain degree) how many 6's I'll end up with.

How many 6's will you end up with if you threw 1000 diceseseses?

 

[temptrader]

Often a system/event/phenomenon will show a characteristic randomness on a microscopic level, but predictableness on a macroscopic level.

If I throw a dice, I have no idea what it will be (random). But if I throw 1000 dices, I can predict (to a certain degree) how many 6's I'll end up with.

That's the binomial distribution, and since we have 1000 dices unfortunately we need to approximate using the Poisson distribution.

No, you can only probabilistically predict. The possibility of having NO 6s at all is still there.

Forgot to add: FW is referring here to the disparity in the probabilities. Here we have 5 times more chance of NOT getting a 6 than getting a 6. This means that the probability distribution curve will be skewed to a "favourable" range.

 

[firewalker99]

How many 6's will you end up with if you threw 1000 diceseseses?

No, you can only probabilistically predict. The possibility of having NO 6s at all is still there.

I'm not sure where this is going, but I can't tell how many 6's one will end up with.

I can only make a statement like "there's a xx% chance that you will have at least yy 6's" or "there's a xx% chance the amount of sixes lies in the interval [a, b]".

So, yes new_trader and temptrader I agree.

 

[temptrader]

here's my attempt at sarcasm . . . .

if all dice were fair then the chance of getting all 1000 6s is

1 in 14166102623834861723796252524915224416640471830910191322323547432140618947596486436347661333869287260068907949302029484915942402681211620694598046617844295512220793103312980549591537160959053027940624117598003417503015722697428176155600362263128567590299511776686592862074376328232990325101248680123776914576482815095784568122986221890411837737570098864613342090972756469661488216176894465388028416768338495326989675118087222767384596111351304957869025273802978281783731929966468210579229830069556698928937342508988340792335737744719376598506908977135291983117722648269177947154657697517074993441515526839887073400191797445153760221695723268255006134044062503100710134200414607696976757837002911389023284338696251543694980946202137938610119300450795091488653253649628649410789376

which is highly improbably, but NOT impossible.:cheesy::cheesy::cheesy::cheesy:

 

[new_trader]

I'm not sure where this is going, but I can't tell how many 6's one will end up with.

But you did say "I can predict (to a certain degree) how many 6's I'll end up with."

I asked you how many, which I thought was a simple question. Imagine you are on a game show and you have to lock in an answer using 4 characters or less. So, for the chance to move on to the next round and win £1 Million, your answer to the question:

How many 6's would you end up with if you threw 1000 dice? Is....:

 

[firewalker99]

FW, I wasn't pressing a point. I didn't think that by saying you agree is the same as saying I was right.

Well, not that it matters, I would've thought that was pretty obvious...

As for your rather rhetoric question... If I can only pick one number, then you already know what it would be. So why press me into saying 166? We all know there are only probabilities, not certainties.

 

[new_trader]

Well, not that it matters, I would've thought that was pretty obvious...

As for your rather rhetoric question... If I can only pick one number, then you already know what it would be. So why press me into saying 166? We all know there are only probabilities, not certainties.

How can it be obvious? I wasn't sure I was right :whistling

 

[firewalker99]

A chart generated from random phenomena doesn't itself has to be random. A true randomwalk like a toss of a coin will produce a chart that will look like a stock chart because the principles gooverning the two are similar. This is, by the way, knowable a priori. A truly random event is serial and you have to look at each event to see the randomness. If you bunch the events in larger chunks, which is what charts do, then due to large diviations which take time to disappear, you will get recginisable and tradable patterns. This is the reason why trading is not gambling: the data are not (under normal circumstances) rigged.

It is misleading to look at random charts and compare it to a random walk chart like a coin toss because you really do not trade the random walk itself but the chart of the random walk which needn't be itself random.

I've had to give this some thought, but you're right: randomness of a phenomenon does not imply it is in itself random. You seem to have a good knowledge on the matter, perhaps we should bring the proposition of determinism into this... one thing's for sure, this has got my braincells working

 

[tune]

Anyway, there is an obvious absence of moderators in this thread even though there are now more than when I first joined. This thread is starting to get ugly and chances are I'LL be the one who is banned...

Apart from all this stuff - I've been wondering what your style of trading is. Please don't request that I wade through a million posts to find out cos life is kinda short. If you don't want to reply that's fine. It's just that I would genuinely like to know what your bag actually is. And no, I'm not after details. Just a general area. Mine is regular Wyckoff style price and volume.

 

[Splitlink]

This intellectual stuff that FXScalper2 and FW are on go a bit beyond the ability of the normal Ladbroke punter, like me, to follow so I'll pass!

My previous ramblings about randomness were not meant to be more than a suggestion that randomness, in itself, had to be included in the general scheme of TA. It, certainly, should not be ignored.

Someone mentioned, earlier, that followers of the Random Walk Theory refuted the theory by becoming trendline followers after a price direction had been decided randomly.

Why not? When, after a period of congestion, the random behaviour has decided on a certain direction that the price is to follow, the traders, who are not, in themselves random but are trend followers, jump on the band waggon and this new course could last for several hours or thirty minutes---no one knows, though, for sure.

Split

 

[barjon]

sorry not to have been paying proper attention - done some severe housekeeping (not had to do that in a while!!).

as far as the topic is concerned isn't it about money flow. ie: If there's a net increase in money in the market then prices will rise and if there's a net decrease they will fall. You could say that there's some randomness about money flow round the edges, of course.

cheers

jon

 

[Profitaker]

There is a subtle, but significant difference, between randomness and predictability. Very few events are truly random, but that doesn't mean all non-random events are predictable.

As far as markets go, if future prices were predictable, there wouldn't be a market. It is the constant disagreement by individuals over price, that causes the price itself to fluctuate. This shouldn't be too difficult to grasp.

 

[new_trader]

There is a subtle, but significant difference, between randomness and predictability. Very few events are truly random, but that doesn't mean all non-random events are predictable.

As far as markets go, if future prices were predictable, there wouldn't be a market. It is the constant disagreement by individuals over price, that causes the price itself to fluctuate. This shouldn't be too difficult to grasp.

Strange, I thought a transaction can only occur when individuals AGREE on a price to exchange. If the market is as unpredictable as you imply, why should there be any disagreement on price? Both are making an exchange at a price they agree is fair. The difference is that one thinks it is cheap at the current price and the other thinks it is expensive.

 

[theroguetrader]

Nah!
Its an auction. How can an auction be random!?

well any trader will tell you that the random walk theory is a load of rubbish of course and always raises a few eyebrows.. !!
" prediction " - and the markets leaving " signposts " everywhere are two completely different things of course... !!
Even stocks maintain an upward trend over time -

 

[Splitlink]

Most traders base their trading policies on a particular price action that works over a period of time. I am no exception. Statistical evidence is difficult to refute, but I still think that randomness is a fact of life in short term trading, particularly in trend turning spots. Once the trend is established. more traders come in, making the trend more predictable as time passes.

Why is it that we wait for breakouts? Because we don't know, that's why!

Split