[mathematician]

Originally Posted by mikkom

Here you go. I always post this link when the randomness discussion starts - I doubt anyone ever reads it but the proof is there (full text available at the link)

http://press.princeton.edu/titles/6558.html

Yeah, I'm not going to pay $50 to read it. But I have heard that there have been proofs that markets do not follow a random walk (in particular Lo's work and Mandelbrot's work). I just mean that I have no proof because I have not actually read any of the proofs in detail, I'm just taking someone's word for it.

So all of this about random walks is sort of a mute point. Nevertheless, it is interesting. This link that I posted, shows a strategy that can profit from a random walk.

http://www.isigmasystems.com/implications.html

[mikkom]

Originally Posted by mathematician

Yeah, I'm not going to pay $50 to read it.

It's free - just click "full text online" - link.

[mr.marketz]

-------

[mathematician]

I'm just going to talk about the zero-sum part here. I'll look at the rest of your post later.

Originally Posted by nubcake

mathematician you have made a mistake with the spread. the true spread is literally not money going to anyone. not your broker, not a bank, not anyone. the real spread is the gap between the highest value being offered to buy at versus the lowest value being offered to sell at.
...

Oh yes, my mistake. They don't necessarily get to keep the entire spread; they're just trying to take part of it. I got it now.

Originally Posted by nubcake

i still don't take it as zero-sum, because to me a zero-sum game means a situation where one side is either winning or losing, but the fact is in forex your broker is always winning quietly on the sideline. if there were no brokers and people could trade directly between themselves then that would be a zero-sum game because one side either takes the other sides money or loses their own money to the other side, and no other money goes in any other obscure direction.

If a game (forget markets etc, just in general...) consists of players and each player gets a score, then a zero-sum game is one where the sum of all the players' scores is zero. It doesn't just have to be two parties. It can be three parties or whatever.

The reason why I consider trading a zero-sum game is because every trader, broker, dealer, bank, etc. involved either takes money from the other participants or loses money to the other participants.

Originally Posted by nubcake

your second example is completely flawed. without the broker there isn't the extra pippage docked from both parties as commission for trading through xyz broker. your second example simply should not have the '-1' inside each bracket. i know why you did this and think this way, but once you understand what the spread actually is instead of what you've believed it to be then you will see why what you wrote is wrong.

Of course, if the brokers were not a part of the picture at all, it would be:

(+10) + (-10) = 0.

But when I said "Without the broker as part of the game: (+10 - 1) + (-10 - 1) = -2", this is what I mean: Some people consider the only "players" of "the game" to be the traders, not the brokers. However, even if you consider it this way, the brokers are still part of the picture. The traders still have to pay transaction costs. In that case the total scores of the players (i.e. just the traders) is:

(+10 - 1) + (-10 - 1) = 9 - 11 = -2

That minus 2 is the money that goes to the brokers who is, as you put it, quietly winning on the sideline. So when you consider the only players to be the traders, then it's a minus-sum game.

So it comes down to how you define the game. (I prefer the definition where everyone involved is a player in the game, so there are not just two sides.) No matter how you slice it, though, money doesn't just vanish. When someone losses money, it's going into someone else's pockets. This fact is part of what makes trading so competitive.

The reason why I made the zero-sum feature to be Axiom #1 in my first post on this thread was to emphasize the extremely competitive nature of trading. I think it's important to consider all the interested parties and what they're trying to do.

[BlackMage]

Originally Posted by nubcake

holy crap that's long, and yes took a long time to type. you're welcome

That was a solid post.

Just a quick note: Many high frequency strategies are centered around providing liquidity - not taking it. Thus they will try to earn (a part) of the spread or at least avoid paying the spread. This way the "spread risk" can be minimized - but of cause it opens up for other types of risks.

[nubcake]

cheers blackmage!

Originally Posted by mathematician

I'm just going to talk about the zero-sum part here. I'll look at the rest of your post later.


Oh yes, my mistake. They don't necessarily get to keep the entire spread; they're just trying to take part of it. I got it now.

cool. whoever told you the spread = broker fee needs to be slapped.

If a game (forget markets etc, just in general...) consists of players and each player gets a score, then a zero-sum game is one where the sum of all the players' scores is zero. It doesn't just have to be two parties. It can be three parties or whatever.

The reason why I consider trading a zero-sum game is because every trader, broker, dealer, bank, etc. involved either takes money from the other participants or loses money to the other participants.

Of course, if the brokers were not a part of the picture at all, it would be:

(+10) + (-10)...

i think this is starting to get too much into semantics. i still don't agree and this is why. you mentioned how it depends on how people define the 'players', which is true. but trading isn't a game of 3 players. it's a game of 2 players (or sides) and a third party. it's like when 2 people want to play tennis, there are the 2 players and then you have the people who take a fee to use their tennis court for an hour. in trading the people who rent you the tennis court = the broker. they don't have to play. they just have to exist and have customers to profit. what then happens between the actual players is of no consequence to them.

unless the brokers also put their own trades in and in a manner which then affects it's clients then this puts a whole new spin on things, but even then this becomes a situation you can't really quantify because you have no way of measuring the brokers involvement in any price action (if any exists). add this onto any trickery the broker may or may not be playing with your price feed and you end up fighting ghosts. i think it's of no value trying to fight the broker (if you believe they are in the battle against you in any way) because you simply have no way of knowing what is and what isn't your brokers actions. apart from any obvious things like trade execution freezing and weird slippage then all you have to measure is the price on the chart in front of you. if the brokers actions are somehow affecting the chart you see, then if you have always been of the mindset that you need to make a strategy purely based on the chart then you are already in the mindset to battle your broker without having to specifically frame it that way in your head. you are just still simply building strategies based on what you see.

i've never experienced anything like this, so maybe someone else can step-up and put a new light on this for me?

but really, i guess at the end of the day if a certain mindset gets you to the goal of being profitable it then doesn't matter how you view anything. as long as it's working for you.

[mathematician]

Originally Posted by nubcake

firstly i don't think the term 'investment' should be found anywhere near anything forex related, ever. like oil and water the two are completely different and don't mix at all.

That's funny, you're literally judging a book by its cover.

[mathematician]

Originally Posted by nubcake

that doesn't mean you should then go to a weekly, monthly, yearly, decade chart just to minimize the spread as much as possible.

Perhaps you have the wrong idea about what I mean by long term. I don't look at the weekly or monthly charts. I use the daily and 3-hour charts. Daily and 3-hour for the main direction and sometimes the 1-hour chart for entries and exits.

[mathematician]

Originally Posted by nubcake

what risks you ask? the risk that your trading plan will simply take too long to play-out over time before you realise it is not a winner. the risk of trying to trade larger lot sizes to overcome the slower pay-rate. the risk of trying to trade smaller lot sizes to minimize damage done by failed trades, effectively giving yourself a pay-cut on top of still only MAYBE being paid at some later date.

It seems like you're assuming a trading protocol of: 1. entry signal, 2. enter, 3. hold until stop, target, or exit signal. You're forgetting about pyramiding. I think it's a good idea to scale into a trade. Start with a small risk. You can add to it if it's starts moving in your direction. With pyramiding, you certainly aren't taking a pay cut. And if you count the initial entry and the scale-in entries, it might be more frequent trading than you think.

[mathematician]

Originally Posted by nubcake

...after a few years of this i would eventually discover that suddenly my strategy isn't the tits and is in fact only a fluke or a 'specific market behaviour' optimised strategy that only works during certain market conditions which occurred for a period of time only once so far.
...
people on the slower timeframes are not afforded this benefit at all. it may take a month, a year, a decade before they discover that the market is no longer willing to accomodate them and their snail tactics.

With a short-term strategy, you run into the same problem of not discovering when the system fails as a long-term strategy. Let's look at a daily system and an hourly system:

You say that the risk of longer-term strategies is that it might take years of data before you discover that the strategy doesn't work. And why might it take years? Because the market happened to be in a bull run and so all the results were skewed? When that bull run was over the system broke down? Sure that could happen, but it's the same for the short term. Consider this:

6000 bars of hourly data = 250 days = about one trading year (no weekends)

An shorter-term hourly system might very well have the same kind of skewed results simply because the market was in an up trend that year. On the other hand...

6000 bars of daily data = 24 years of data (6000/250).

Now it's true that this could just be a lucky 24 years and if I had gone up to weekly or monthly charts, we have the same problem again. However, I'd rather see 24 years of data where things can change: interest rates, inflation, business cycles, etc. All these economic factors which cause these bull runs and bear markets, they are more likely to be seen over a longer period of time. So I would say that looking at a longer-term strategy is actually less risky.

[mathematician]

Originally Posted by nubcake

i think this is starting to get too much into semantics. i still don't agree and this is why. you mentioned how it depends on how people define the 'players', which is true. but trading isn't a game of 3 players. it's a game of 2 players (or sides) and a third party...

Yes, let's just agree that it's semantics.

[nubcake]

i think pyramiding is a large part (or maybe almost all) of the key to being profitable, not just in the fact that when you win you win more, but more to do with the maths behind entwining this with trailing stops and the shifting nature of risk vs reward.

in fact, it'd be interesting to know if there is ANYONE out there who considers themselves profitable who DON'T pyramid. is there a thread already for this that someone can point me to? i've never heard of a winning trader who doesn't pyramid.

probably last note on zero-sum vs negative sum : with all the talk of broker being the spread / broker not being the spread / what the spread is / etc etc etc, i lost sight of the actual reasoning behind things. what your figures were showing was simply where money was going, and to who. the reason why trading is a negative-sum game is because no matter what an INDIVIDUAL trader does s/he will always be buying at the higher price of the bid/ask and selling at the lower price of the bid/ask, and therefore as soon as a trade is entered the very first hurdle to overcome is the spread. if a trader buys then instantly sells, or sells then instantly buys, the trader loses money (the spread). nothing can change this, whether a broker is taking pips or not the spread puts every single trader in negative profit the instant they make a trade. If someone can figure out some way of making this untrue then they literally have themselves a money printing press. for everyone else though we have to wait for the prices to move +/- spread before we are at break-even.

this then comes back to why pyramiding seems to be part of the key to profits. as long as there is enough movement in price to allow a pyramiding strategy to enter more than just the initial trade then the act of pyramiding has changed the risk vs reward. if you had a 50/50 chance of the price going up or down, then to be stopped-out on 1 trade loses you whatever amount... but if the price moved in your favour and you are able to enter a new trade (pyramid into a trade) and aren't increasing the initial risk (simple example would be if your initial trade has its stoploss trailed into break-even already so the only risk is the new trade pyramided losing) you have now got a setup where the amount of reward gained with each positive pip movement is greater than that of which you lose any time you lose a trade. the problem with pyramiding is that you need the price to move further before a pyramided trade is at overall break-even and a free trade, so this is where volatility comes into play as far as i see it.

now the problem as i see it with regarding an overall larger timeframe trend being a factor in short-term trading strategies being profitable or just flukey is that the degree to which a long-term trend might be (regarded as) positively affecting price movement in the strategies direction is minimal. when it comes down to the bare-bones of things price moves up or down (no not sideways). each time you zoom-in on from a larger timeframe you still see peaks and valleys, and areas of mini-trends within larger trends, and then as you zoom-in further you see the same again, and again, and so on.

the further problem with this is that it's all subjective, and candlestick charts are a poor bearer of price information. if you break each section into, let's say, 8 minute candles, what you see is a totally different representation of the exact same data that is represented by a 5min chart or a 10min chart. to even shift the data on a 1min chart half a minute forward or back would completely change the picture on the charts. some people don't realise this, and that's a little scary to know that others don't see how ambiguous and subjective charts are. Twoblink has touched on this hard in some of his posts, and since Twoblink came to this realisation he has since disregarded open and close values and uses a high/low chart only as this removes a part of the ambiguity of charts. point and figure charts also remove some ambiguity also.

so, when you see someone shorting on a lower timeframe when you are going long because your higher timeframe shows a bull trend you see them going against an 'obvious' bull trend... but what's to say that the bull-trend you see is actually going against a bear trend in an even HIGHER timeframe than what you are looking at. this expanding/contracting of timeframes and trends etc can go on forever. at what point do you stop zooming out to see the 'true' trend? you can't really. it's also the opposite on the flip-side of things, if you look at lower timeframes they can show a reversal already formed while a higher timeframe is still appearing to be just retracing... it's all a case of when you decide your beginnings and endings are. the differing timeframes simply warp the data. this is where understanding exactly what information you are being given by whatever chart you look at is so important because to know what is given is to also know what IS NOT given (or perhaps i should say hidden?).

this is why 'chart patterns' are subjective. if everyone traded all chart patterns exactly the same and also looked at only a specific timeframe and never changed that either... then and only then could chart patterns be reliable. but all it takes is for a proportion of people to change their chart timeframe for the reliability of the patterns to now skew, and then add on to that the fact that traders can decide which patterns to trade or to ignore, when to sleep or go eat lunch, etc etc etc. people also die, and new traders also enter the game. so many variables affect something as simple as a chart pattern.

charts are liars. reading between the lines is an advantage most don't realise.

[mr.marketz]

Glad to see this discussion developing. I took down my earlier post because I felt I was overstepping my bounds and not allowing others to draw their own conclusions. Which is why I stopped posting on (and reading) forums for a while.

Pyramiding as nubcake calls it (or scaling in, or averaging) is a tremendous edge. I call it position building... I made a mention last night of taking positions, and not trades. Again, people should feel free to act accordingly. There are those folks that say a trade consists of a single view point... you either get the direction right, or not. That simple.

However, any traders that I've known who have employed that static approach typically flat-line their equity curves... or eventually dissolve them over an extended period of time. Price action IS random... people are not. They consistently drop hints as to when they may want to take their profits, or cut losses. Those are the windows of opportunity for savvy independent traders to profit. So you start building your "position" at those opportune moments.

You don't need to see Citibanks order flow data to cash in... you simply need an approach which adds dynamics to your trading. Averaging a position is one of the primary ways to do that.

Just my 2 cents.

MM

[mathematician]

I think we're starting to come to some common ground now.

Originally Posted by nubcake

i think pyramiding is a large part (or maybe almost all) of the key to being profitable...

Yep, pyramiding is the name of the game.

Originally Posted by nubcake

this then comes back to why pyramiding seems to be part of the key to profits....

That's a great way to explain it!

Originally Posted by nubcake

the further problem with this is that it's all subjective, and candlestick charts are a poor bearer of price information....

Yes! I once had the privilege of using TradeStation. It's great because you can set an arbitrary number of minutes you want to use for your candles/bars. It will just slice it up in whatever size you want. We were using 55 minute bars just to be slightly different from everyone else who uses 1-hour bars.

Originally Posted by nubcake

so, when you see someone shorting on a lower timeframe when you are going long because your higher timeframe shows a bull trend you see them going against an 'obvious' bull trend... but what's to say that the bull-trend you see is actually going against a bear trend in an even HIGHER timeframe than what you are looking at....

Yes, and at some point you just have to make a decision and go with it.

Originally Posted by nubcake

this is why 'chart patterns' are subjective....

And it's a good thing they're subjective. If everyone agreed, there'd be no market.

[nubcake]

mr.marketz jump on in the waters fine.

mathematician i guess this is why i keep saying that you just need to overcome the spread. once you have a strategy that does that then you can decide when your beginning and end is, since it is all subjective. what isn't subjective is how much the spread is. as you know the spread is relative... to the subjective timeframe. so having a strategy that churns as fast as possible while overcoming the spread is where i want to be, and not be trading higher timeframes which churn slower and might be setting me up to be knifed in the back by a strategy that will someday fail.

i guess you either profit fast or slow, or you bleed money fast or slow. not necessarily dollar values but the turnaround on which way the individual trades play out.

i guess it's like having a clock that clicks each second, or one that only dongs every hour. it's easier to know when the clock that clicks each second has failed than the clock that only dongs once an hour.

edit : following from the first paragraph to mathematician... i guess you could say that in this subjective world the spread is the one truth that can be reasonably relied on. the spread is the anchor point to then swing-off into the subjective realm of price action. to anchor onto anything else is to be swimming in darkness upside-down with whirlpools distributed randomly and also shifting.