Probability/Odds Trading the Currency Markets ie. Gambling

Hi Guys...

I've been really struggling to get my head around this question and I hope one of you can help me out here.
I trade the Forex currency markets (which is essentially gambling with the odds in my favor... well.... most of the time)... Anyway, my overall Win/Loss ratio is 60% ie. I consistently win 6 trades out of 10.

My question is this: Given that 6 out of 10 trades is a win, and my last 6 trades are (eg. W, W, L, W, L, W): What is the probability/odds that my 7th trade is going to be a win? And if that trade is a "Win" OR a "Loss", what is the probability/odds that my 8th trade is going to be a "Win"? And then the 9th trade etc....

I'm really interested in the probability of trades 7, 8, 9 & 10 given the history of trades 1 to 6....

I hope that kinda makes sense and I'd love to hear what some of you have to say about this.... and, it would be really nice to be able to reduce my "bet" when the probability/odds are not in my favor....

Thanx in advance....

Toon

it seems to me that yet again you have a case of the past being unable to forecast the future in gambling.

Your odds are 6 out of 10 each time [you say]. Thus by your own definition, the past does not matter. You would have to stipulate that, after a win, say, the next time around you feel your odds have changed.

In other words, under the parameters you give, patterns will mean absolutely nothing.

Surely the fact that I have X amounts of winning trades in my immediate history has an effect on my immediate future? For eg. lets change my previous trade history to W, W, W, W, W, W.... given the fact that I consistently win 6 out of 10 trades ie. lose 4 out of 10, the chances of me winning the next 4 trades is a lot less than loosing them (or a portion of them)....

I'm not trying to forecast the result, but I am trying to get an indication of the probabilities/odds of a losing trade coming up.... Using that information, I can then adjust my "bet" size: Increasing it when the chances of a winning trade are good and decreasing it when the chances of a loosing trade are good - obviously.

Or am I totally off the mark here?

Quote: Toon

Or am I totally off the mark here?

Yes, it seems you are totally off the mark.

Let's use roulette as an example. You have a 1 in 38 chance of hitting any given number. Let's say you choose Red 7. In the last 37 spins it hasn't hit. Should you increase your bet because it's now more likely to happen? The answer is no, it's still a 1 in 38 chance of hitting.

I would imagine your Forex trades are a lot more complicated than spinning a wheel to determine outcome, so if it were me I'd analyze the various parameters of the winning and losing trades and see if a trend can be spotted there and use that as a guide as to when to risk more money (on more favorable trading circumstances).

I'm not sure you can compare trading with roulette. There are three major differences:

In roulette (or any other game of chance) you have intrinsical access to all relevant probabilities and can calculate basically everything you wan't to know, including your "edge". The dynamics of trading is fundamentally different from that.

Further, the game of roulette is stable: it is the same game today and will be the same game in ten years. This is definetly not the case in trading, as you trade against the market (and not against some rules). Strategies that work today might not work tomorrow.

And the third: In roulette every event is statistically indepenent from any other event. Again thats not neccessarily the case in trading (for example if you trade against a single agent and he drops out of the market).


BTW, how do you come tthe 60% figure of a Win/Loss percentage ? It is most surely an estimate from your past performance. And it better not from your 3 wins on 5 trades. Because 3 hits on 5 tries means nothing - make it 1000 tries and one can speak again.

For all the reasons stated above, in my opinion all this trading stuff is pure nonsense when your only source of value comes from fluctuations of the market price.

You guys are 100% correct... but I think you are missing my point/question here.... I'll try to rephrase....

Just as a matter of interest:

1. I understand the whole story about probabilities and the fact that the probability of each new trade has nothing to do with the previous one etc.
2. I accept that trading and a game of chance have nothing in common which is why I stated in my first post that trading is "gambling with the odds in my favor" ie. I will not enter a trade unless I am confident that I am entering the trade from a position of high probability (which I gauge by a whole range of variables etc. etc.)
3. The one thing that I feel IS given is my win/loss ratio of 6:4 (the period is irrelevant, but it stretches back to well over 1000 trades).

So, given that I win 6 out of every 10 trades (the amount won or lost is irrelevant), how can I get an indication of WHEN the probability/odds (whatever you want to call it - I am not being clinical about the definitions of each - which is where I think you guys are getting me wrong) are favorable or rather indicating that the chances of a losing trade coming up is good...etc.

I think I am getting somewhere with the Odds Ratio and the Likelihood Ratio as well as a bit of Standard Deviation.... SO, I'm going to throw all of this into a bag, shake it up and see what I come up with....

I'd appreciate it if you guys will keep an eye on this thread to check on my logic when I come up with something.... who know's, this could be revolutionary...!!! Heh

Thanx for your comments thus far...

Maybe I should not be using the terms Probability or Odds.... I'm trying to establish the relationship between my (call it) 10 previous trades and my (lets say) 5 future trades....

There must be some kind of relationship because I win 6 out of every 10.... otherwise I would be winning all my trades, which would be great but that is not the reality...

Quote: Toon

how can I get an indication of WHEN the probability/odds [...] are favorable or rather indicating that the chances of a losing trade coming up is good...etc.

Again: given only your past performance - you can't.

Whether the trade is favourable is part of your own *research*, models, parameters and stuff. This has nothing to do with the mechanics of probabilities, it has all to do with understand your market.

If you ask "which index to use", then you did not fully grasp the problem, because this index is the very solution to your problem.

Asking a question like "I wan't to solve this (vague) problem, where do I find its solution?" doesn't make you look like you know what you are doing.

think of card-counting. Is anything similar to that, such that a previous win or loss would be altering the basics.

Just trying to see patterns per se is very human, but in gambling a known fallacy.

Forget I even mentioned trading.... consider that I am playing Blackjack

There is a certain level of skill that helps me win hands but after a few winning hands, the chances are going to increase that I am going to start loosing... I therefore want to reduce my bet size over this period. After loosing a few hands, I increase my bet again as the chances that I am going to win increase....

I know you can't say for definite when you are going win or lose, I'm just trying to establish a relationship (if there is one)....

I'm making progress with research and will hopefully be able to show you what I am talking about soon....

Fair enough...

Don't show it to us (actually we don't care that much). If you are confident on what you are doing, feel free and show it to the market.

But if you are asking for advice, you should listen. The first question you should answer from your research is: are your trades statistically independent.
If they are, there is no reason to alter your bet sizes based on past performance, because then past performance does not indicate anything.

If your trades are not statistically independent, find the correlation between them. If you have a positive correlation increase your bets after each win, reduce after each loss (i.e. play a "positive progression"). If your correlation is negative, decrease your bets after each win and increase your bet after each loss (play a "negative progression").

But whatever you do, you must be *sure* that (1) all your trades are statistically dependent, and (2) you quantify the correlation correctly.

Blackjack games from the same shoe are slightly negative correlated, since winning hands are likely to consume good cards, which are then not available in the shoe the next hands are drawn from. But you see this is a statement which follows from the analysis of the game, not from past results.

Roger that MangoJ.... thanx for that (I must read up a bit more on Blackjack)

The interesting thing is that I have just come up with a solution to my question. Well, I think I have. I'm just busy testing the results and will post them shortly as they are proving to be quite interesting and in my opinion very useful...

I think I put you all off the mark from the beginning.... What I have been trying to establish (and now realize) is that this has nothing to do with the actual business of gambling/trading.... it has everything to do with the money management thereof....

What I have established (or rather.... think I have) is that there is indeed a relationship between past history and future trades/bets in relationship to relative bet size and the end of day profit ie. bank balance....

I just need to clean up my calculations and put it all together in a way that it makes sense without going into detail....

Watch this space....

Quote: Toon

What I have established (or rather.... think I have) is that there is indeed a relationship between past history and future trades/bets in relationship to relative bet size and the end of day profit ie. bank balance....

What?

As a few posters have tried to explain, if the success of an individual trade is independent from previous trades (which it almost certainly is), then the past has no bearing on your future trades.

If you Win or Lose a series of trades, you will not regress to your expected win average over the next few trades. The market does not care how you the individual trader is doing. If anything, you should be questioning whether your 60% average is correct if you go on a losing streak.

You should be making bets in proportion to your expected advantage, this is how you manage your money. But your expected advantage only depends on the factors affecting the one trade you are about to make, not on recent history.

Quote: Toon

Forget I even mentioned trading.... consider that I am playing Blackjack

There is a certain level of skill that helps me win hands but after a few winning hands,
the chances are going to increase that I am going to start loosing...

Sounds like the gambler's fallacy to me.
Not for independent events that you are actually thinking about.
You are correct that over the next "many X number of trials" your winning percentages can converge to your expectation,
not just the next one or 5.
How big is X?

You seem to be very concerned about the results from a distribution of X trials.

You should probably be more concerned learning about the arc-sine law as it applies to trades.
You are more likely to be ahead or behind more times in a series of trades than say 50% ahead/50% behind or 60%/40%
and nothing you can actually do about it unless you can increase your winning percentages.

Dr Thorp is a very well know mathematician and trader. (there are others)
He is very wealthy from both and even has a fancy office in Newport Beach, CA.
Learn from him

Why do you think
you can re-invent the wheel??

Quote: Toon

I therefore want to reduce my bet size over this period.

Why?
You are in a winning streak or trend.
Ride it out, bet your advantage.
you could be the lucky one and win big and retire and just watch DVDs all day.

Oh, I see, you KNOW you are going to start to lose sooner or later because you should be only winning 60% of the time.
The Law of Large Numbers!!!

Quote: Toon

After loosing a few hands, I increase my bet again as the chances that I am going to win increase....

The Law of Large Numbers!!
This is good stuff

Quote: Toon

I know you can't say for definite when you are going win or lose,

Then why worry or think about it, Study and bet your advantage.

Quote: Toon

I'm just trying to establish a relationship (if there is one)....
I'm making progress with research and will hopefully be able to show you what I am talking about soon....

Classic

Good Luck to you

ps
read more about Dr Edward Thorp
http://en.wikipedia.org/wiki/Edward_O._Thorp

100%... I agree with you totally.... but I said "relative bet size" not the actual making of the bet...

As I mentioned earlier.... I put you all on the wrong track initially and I think you guy's are all misunderstanding where I am trying to go with this.

What I am trying to establish has absolutely nothing to do with the actual trading/playing (Blackjack) of the game.... I am trying to prove that if you change your bet size during certain conditions (ie. the chance being good that you are going to win/lose a bet) based on your prior performance then you will make more money at the end of the day - which is what we do is all about.

The results I am getting so far are really interesting (and are actually proving my point), but I am yet to prove my results sufficiently enough to convince all of you.... (I'll get lynched if I post some crap after all this) hehehhh

I'll keep you posted....

Thanx 7Craps.... I'll definately read up about Dr Thorp and look into the Arc-sine wave.... hav'nt heard of that before.... sounds interesting...

Cheers...
PS: Still going to try to hammer out my hypothesis... it's become personal

Quote: Toon

I am trying to prove that if you change your bet size during certain conditions (ie. the chance being good that you are going to win/lose a bet) based on your prior performance then you will make more money at the end of the day - which is what we do is all about.

The general consence is (for independent trials), that you change your bet size proportional to your advantage. This will maximize bankroll growth, meaning you will double your bankroll earlier.

In fact this goal is different from maximizing probability being "ahead at the end of the day". Why would you want to be ahead at the end of every day, why not ahead at the end of the year or your career ? Being ahead every day only looks good for potential investors but is to much of a constrain for your gambit. In fact this phrase is worth nothing if your competitor performs better at the end of the year.

First of all: Define your problem correctly. Make assumptions or gather them from previous experience; yours or that from others!!! Sorry to be rude, but you sound like all those people who think nobody ever thought before them.

So, get a textbook about econometrics and time series.
One question is: are your trials independent?
Well, the answer is not on this forum. It is in the statistical testing of your time series. Look up "autocorrelation" in your textbook's index.

If the result is zero (no autocorrelation), then you are playing roulette (or craps, or blackjack, or baccarat): the past deviations are no information for the future, stick to the constant average.

If there is autocorrelation, use the info to adjust your expected value. If negative, DON'T BET. This idea of "reducing the bet size" is preposterous. If positive, bet. How much? Depends on your stash... You may use Kelly's criterion, though it has its flaws.

If you were trading for a company, I gather they would have coached you. So I guess you are on a personal errand. My advice: keep away from the market, it is an institutional way for big fish to extract money from the small fish that we are.

Quote: Toon

Maybe I should not be using the terms Probability or Odds.... I'm trying to establish the relationship between my (call it) 10 previous trades and my (lets say) 5 future trades....

There must be some kind of relationship because I win 6 out of every 10.... otherwise I would be winning all my trades, which would be great but that is not the reality...

This is either a troll account or you are just very stupid. They have carefully explained why this thinking is bad.

Also, do you really think that your small winning or losing streak as any effect on currency movement? The world just changes all of a sudden because you pick one currency over another and are due for a loss?

The only way for you to know in my opinion is to start keeping a trading diary. I trade off off point and figure patterns and they all have specific probabilities.... I think to know your probabilities you need to know a few things:

1. The reason you made the trade...do you initiate the trade by going long or short ( sorry I dont trade forex..only equities, options and index options so i dont know how you "enter" the market with your trades). I would analyze long and short trades separately.
2. was it based on a pattern or a range, an indicator/oscillator.... If so define it.
3. Are the trades in the same product, if not look at them separately
5. did you have rules on the trade... are your losses always "stop outs" or do you sometimes close the trade before a stop?
6. On winning trades... only count wins if your objective on the trade was primarily met... for example if you think the USD/JPY offers 2% upside with only 1% downside but you took a small profit on a 10bps move I dont really count that as one in the win column.

Its work to analyze this and I do it constantly on spreadsheets but once you know where your greatest strength and weakness are you can gain a lot of confidence and learn to manage your impulses that lead to bad trades and situations where you should have been more ballsy too.
I think you will find you trade a particular product and set up better than other... that's the probability you want to nail down.