Random Entry System

[ARTjoMS 10]

It is natural to assume that random entry system wirh risk:reward ratio of 1:1 would be BE. (I am sure that someone has proved it)

 

And so I was wondering... should a random entry system with higher reward:risk ratious be +EV? (Since price tends to trend)

 

Assuming that some risk:rewards (in exact pips) perfrom better.. how could these results be used further to realistically give some edge?

[Perrin 10]

Why would you want to use a random entry system?

 

Also, don't forget about commissions, slippage, taxes, etc. You need to take these into account.

Edited February 1, 2013 by Perrin

[ARTjoMS 10]

Why would you want to use a random entry system?

 

Also, don't forget about commissions, slippage, taxes, etc. You need to take these into account.

 

Where did I say i want to use random entry system? If you read carefully my question ( ''how could these results be used further to realistically give some edge?'') then you probably would not ask such retorick questions or mention kindergarden remindments of slippage.

 

Sorry for being harsh, but i have encountered issue like this countless times and it is really annoying that people don't reead what you have written... or they simply have no deduction skills.. or my english is so bad that they don't understand a word.

 

It is ok to stay away, but why waste our time?

[Perrin 10]

My apologies.

All the best,

Perrin

[Colonel B 10]

It is natural to assume that a random entry system with risk:reward ratio of 1:1 would be BE. (I am sure that someone has proved it)

 

Maybe, but I have yet to see the proof. For all the assertions about this I have yet to see one solid piece of evidence. I have yet to see a completely random test and not some test that claims to be random with out exposing all the parameters so that folks can duplicate the results. Just because someone says its true doesn't mean it is. Or a better interpretation would be that just because someone that is selling something says its true doesn't mean it is.

 

And so I was wondering... should a random entry system with higher reward:risk ratios be +BE? (Since price tends to trend)

 

Assuming that some risk:rewards (in exact pips) perform better.. how could these results be used further to realistically give some edge?

 

They can't and here is why. Arbitrary risk to reward is only 1 part of a much larger equation. It is important without question. However if you are randomly or intentionally buying at bad prices you will get stopped out. If you are buying at bad prices and you think the solution is to use a bigger stop out then you will fail. It is far better to have solid risk management in place then to try to figure out the reward. In reality you need to figure out your risk before you can even begin with figuring out how much you can profit. Also you can only manage your risk you can NEVER manage profit. How close is your entry to the bottom?

 

Also you have to consider the market you are trading. If you tried to use a larger stop out and larger profit target then what the market you are trading will allow then chances are you will get stopped. Assuming your reward is higher then your risk. Or you will simply have to exit at a loss.

 

I can give many examples of how the question is a fallacy in further posts.

[Colonel B 10]

Also the edge has been and will always be knowing where shorts and longs are trapped. Using correlated markets and some sort of bid vs ask software is the most efficient way to do that. There are other ways with certain patterns that will accomplish it however they are much less intuitive. Markets move because of short, shorter, or shorter then you term traders exiting positions. Being able to recognize where short term traders are going to enter or are entering and where they will most likely exit is the supreme advantage a trader has. Knowing that info is the edge and is the only edge and the only edge that counts. This will only cease to be the edge when and if short term traders stop trading like short term traders.

[M.A 10]

It is natural to assume that random entry system wirh risk:reward ratio of 1:1 would be BE. (I am sure that someone has proved it)

 

And so I was wondering... should a random entry system with higher reward:risk ratious be +EV? (Since price tends to trend)

 

Assuming that some risk:rewards (in exact pips) perfrom better.. how could these results be used further to realistically give some edge?

 

It can only be proved on paper and has been proved under concept of probability. Even flipping a coin will not produce 1:1 result in reality. In forex, u cannot assume a breakeven by entering randomly. Here is a detailed explanation of WHY.

Flipping a coin is the best example to understand it (assuming that it produces 1:1). There is no trend type factor in flipping a coin. There is no fundamental or technical analysis (there is said to be one type of technical analysis to judge it which is taken from Cricket) required to judge the result of coin flipped. These reasons make coin-flipping purely random.

Whereas market movements r based on real factors. Even if u ignore the spreads and other costs (coz they can be adjusted by adjusting TP and SL), result can never be 1:1.

 

I think best way is to make a simple program to trade randomly. Review results, adjust settings and re-test and then all this cycle again. If by adjusting results, u get close to what u r trying to acheive (which is not possible in my opinion), then u r doing right thing.

[Gekko78 10]

Here is something I have really never been able to grasp about questions like this......

 

I always assume that ALL trades are 50-50 ........ The price can only go 1 of 2 ways

 

1) it either goes down and hits your stop

 

2) It goes up and hits your target.

 

Yes price can go sideways but I do not count that as it will eventually get to one of the above mentioned.......

 

Even if you have this "great system" that shows when a set up happens it will work 90% of the time .....it still always has a 50-50 shot once the trade is placed.

 

Just like the coin flip is always 50-50 , yes it can land on heads/tails 10 times in a row but when the flip actually happens it is still always 50-50

 

If someone could explain otherwise to me I would appreciate it.

 

Thanks

[ARTjoMS 10]

Just like the coin flip is always 50-50 , yes it can land on heads/tails 10 times in a row but when the flip actually happens it is still always 50-50

 

If someone could explain otherwise to me I would appreciate it.

 

Thanks

 

Guys here are questioning things that are mathematically self evident. If on big sample size with R:R 1:1 one still doesn't get BE (EV=0) then he siply has wrong code.

[Gekko78 10]

Guys here are questioning things that are mathematically self evident. If on big sample size with R:R 1:1 one still doesn't get BE (EV=0) then he siply has wrong code.

 

:confused: what?

[ARTjoMS 10]

My idea was that from results obtained one could draw some conclusions of how trending the market is on particular time (pip) frame.

 

Just maybe... it migh give some edge in deciding if it is worth to keep profits ride or close the trade.

[ARTjoMS 10]

:confused: what?

 

I said your thinkiing process was right.

[Gekko78 10]

I said your thinkiing process was right.

 

Oh ok LOL thanks!

 

 

 

I am not "mathematically nor statistically inclined" so I just wanted to make sure

 

Thanks again

[Colonel B 10]

Here is something I have really never been able to grasp about questions like this......

 

I always assume that ALL trades are 50-50 ........ The price can only go 1 of 2 ways

 

1) it either goes down and hits your stop

 

2) It goes up and hits your target.

What you mean is The trades you take can only go 1 of 2 ways. Price can go up then down then up again before it hits your stop.

So 3) price goes close to your stop with in 1 tick then goes in the direction you thought.

4) it goes with in 1 tick of your target and then goes after your stop out.

If you choose to pull the trade at 1 tick within your target either way it effects the 50-50 thinking.

 

Yes price can go sideways but I do not count that as it will eventually get to one of the above mentioned.......

Eventually. If and only if you have stops and targets that are reasonable for that market. Again its how you trade price not price itself. What about my profit target on the spoo at 3000? Or my target on the 6E at 4.9? I still have my first target in apple 1200 just like the folks on TV said and my next target is at 3000. And my stop out for all 3 is 0.

Even if you have this "great system" that shows when a set up happens it will work 90% of the time .....it still always has a 50-50 shot once the trade is placed.

 

Just like the coin flip is always 50-50 , yes it can land on heads/tails 10 times in a row but when the flip actually happens it is still always 50-50

 

If someone could explain otherwise to me I would appreciate it.

 

Thanks

*This example is pertaining to futures, forex, and other types of markets such as these. I can't fully comment on baseball cards or comic books since I don't trade or collect either of those.*

I have seen this with gurus like the famous FT71. This is not true. Parts of it are true. How you trade makes the percentage. Your trades are 50-50. Not mine. There are fundamental facts about price and markets that make this thinking not true. When you have enough collected short term traders collected in 1 direction and their stops are triggered then 100% of the time price moves in that direction. Not 50% not 90% not 99% but 100%. How far and for how long? It depends on a ton of factors. But suffice it to say you can get the direction 100% of the time when stops are triggered. What is the direction a month from now? Don't know. My answer is what are the positions of retail traders a month from now? But you said you could predict the direction 100% of the time!!! No. Make no mistake you can predict the direction at the time the stops are hit and a little before. An hour before there is no way of knowing and there is no way to trade it an hour after. So your trades are 100% winners then? No. Because you can still be too slow to get on and miss it or you can still hold it for too long thus turning a winner into a loser. And also there is no way to know with certainty where it could end up and how long it will take to get there. It could go 2 ticks then come back to your entry and even stop out before it goes to your target. So trades can have 100% chance of winning in certain situations but anyone can ruin a perfectly good trade. Just because you can get the direction right doesn't mean that you know the direction 100% of the time. You can for sure know the direction after enough short term traders have been collected in 1 direction and their stops are hit.

 

50-50 is relative because there is actually more then 2 possibilities of what can happen in a trade. Some clarification could be maybe someone has small profit targets and takes profit right away. Or maybe some one has a target of 3000 in the S&P and a stop out of 700 and is still in since last May. Again maybe your trades are 50-50.

 

Your stops and targets and how you exercise them determines your percentage. If you don't move a stop or a target then yes its going to be 50-50. A 90% system may just mean they are using a large stop and a small profit for targets. These are the reasons for the percentage. Not the market. The applications of trades to the market make the percentage. I have had lots of trades where I was in a winning trade and it wasn't to my target and I pulled it. It never went to my target and would of turned into a loser. In this situation it defies the 50-50 ideas. I kept a winner from turning into a loser and with out adding or averaging. How do you quantify that? Yes if I traded with a 50-50 mindset and wasn't dynamic and flexible in the market then yes it would of been a loser. However it wasn't and the money went into my account.

 

Ok so you could say that every trade is a winner or a loser. And maybe that was the point all along. However what about a scratch trade? Winner or loser? What about a trade that has some winners and a loser? Do you account for a stop moved to your entry after you take some off or do you keep it at its original place? Are those winners or losers?

 

In closing you can increase your odd of winning or losing by how you exit the trade. When you flip a coin you have to wait for the coin to settle and accept the results. In trading its not that way. You can actually cut the trade when it is at any point of winning and its counted as winning. The equivalent to this with a coin is being able to see what the side of the coin is while its spinning and push it over with your finger.

[BlueHorseshoe 10]

Guys here are questioning things that are mathematically self evident. If on big sample size with R:R 1:1 one still doesn't get BE (EV=0) then he siply has wrong code.

 

Hi ARTjoMS,

 

Ignoring costs such as spread, slippage, commissions etc, then this is correct, yes.

 

If you run lots of tests of this using market data, however, the results won't be distributed in the same way as they would for a coin-toss as quickly. In other words, you'll need a larger sample size to start seeing the bell curve, and you'll see larger than expected 'runs' in the data than you would expect with a coin-toss. I can't tell you why.

 

For this reason, I have a theory (which I have never gotten around to testing because it is not likely to be practically trade-able) that with a position-sizing algo that shuts off to drawdown more readily than it opens up to run-up of the equity curve, such a random entry system that produces profitable runs could be long term profitable.

 

Unfortunately the other possibility from your opening post isn't viable: as you decrease the risk/reward ratio the frequency with which you are stopped out will increase to keep things at break-even over larger sample sizes. Conversely, if you increase the risk/reward ratio the opposite will happen over the long run - you'll be stopped out less frequently (higher win rate), but your losses will be correspondingly larger, leading to breakeven with a large sample size.

 

Note that in both these instances, although the end results will be similar to the the 1:1 ratio results, the equity curve would be far less smooth. Theoretically, this increases risk of ruin and lends itself less to successful position sizing, which tends to thrive on consistency of returns.

 

I hope that answers your question.

 

BlueHorseshoe

[Colonel B 10]

My idea was that from results obtained one could draw some conclusions of how trending the market is on particular time (pip) frame.

 

Just maybe... it might give some edge in deciding if it is worth to [let] profits ride or close the trade.

 

How about this. What causes trending? Time? Price? The moon and stars? Any idea? If the market is trending up what should you do? Buy it? Look for places to sell it? Any idea?

 

What causes trending to stop? Time? Price? Your stop out?

 

Is a market trending on a 5 minute time frame? Is a market trending only on a 30 minute?

 

So it went up 100 pips in 10 minutes. Do you buy it? Do you look to sell it? Is it going to go up another 100 pips? Any ideas?

 

Do you have something that will allow you to get on the trend in the first place? Start with this first. Where you get in is important in that it gives you a better chance to win. But getting in is not the only thing. Forex is the worst market for any entry.

 

If you don't have anything to get you into the trend what are you using to fade it? Can you fade it or do you need it to be trending to make money?

 

The only way to know if its worth it to keep it on or take it off is to bet more. You bet more. Once you put more on you can take them off and let others ride. Its simple if you have more on. So the answer to the edge you are looking for is BET MORE. Put more contracts on.

 

What do you do once you have done that? Cut your winners quick and cut your losers even quicker. +25, +50, +75, +100 for 4 contracts. 8 would be 2x+25, 2x+50, 2X+75, 2x+100. But what if it goes 600 pips? My response is, does it normally go 600 pips or did it do that this time? What about the times it goes 100 pips and turns around and comes back? But I use a 300 pip stop out and I cant take +25. Well then change your stop to -50 or -25 instead.

 

Taking big swinging d!#* trades is sexy. That is the way retail traders with small accounts trade and think. You get good and stay in this game by taking smaller bites with more on. Not taking 600 pips with 1 or 2 contracts.

[ARTjoMS 10]

In other words, you'll need a larger sample size to start seeing the bell curve, and you'll see larger than expected 'runs' in the data than you would expect with a coin-toss. I can't tell you why.

 

This is interesting, but very hard to believe. In a random entry system you enter short with 50% chance and enter long with other 50% chance, so you can imagine trades as a series of determined outcomes and your result depends entirely of whether system has picked buy or sell, but .. since it is effectively 50-50 I don't see how there can be longer runs than in coin flipping situation.

 

Unfortunately the other possibility from your opening post isn't viable: as you decrease the risk/reward ratio the frequency with which you are stopped out will increase to keep things at break-even over larger sample sizes. Conversely, if you increase the risk/reward ratio the opposite will happen over the long run - you'll be stopped out less frequently (higher win rate), but your losses will be correspondingly larger, leading to breakeven with a large sample size.

 

Are you saying that a random entry system with various rick:reward ratious will still be BE in a long run? Now I don't know how close these numbers are in long run... but that shouldn't be true. You can imagine that as a combination of two 1:1 trades. You win 1st one and then risk 2:2.. so 1:3 in total.

 

If both of them were 50-50 trades then 1:3 trade would be BE, but since 2nd trade contains information that market has moved in one direction it affects probability of outcome. It should still be close to 50-50, but not exactly 50-50.

[Gekko78 10]

What you mean is The trades you take can only go 1 of 2 ways. Price can go up then down then up again before it hits your stop.

So 3) price goes close to your stop with in 1 tick then goes in the direction you thought.

4) it goes with in 1 tick of your target and then goes after your stop out.

If you choose to pull the trade at 1 tick within your target either way it effects the 50-50 thinking.

 

 

Eventually. If and only if you have stops and targets that are reasonable for that market. Again its how you trade price not price itself. What about my profit target on the spoo at 3000? Or my target on the 6E at 4.9? I still have my first target in apple 1200 just like the folks on TV said and my next target is at 3000. And my stop out for all 3 is 0.

 

*This example is pertaining to futures, forex, and other types of markets such as these. I can't fully comment on baseball cards or comic books since I don't trade or collect either of those.*

I have seen this with gurus like the famous FT71. This is not true. Parts of it are true. How you trade makes the percentage. Your trades are 50-50. Not mine. There are fundamental facts about price and markets that make this thinking not true. When you have enough collected short term traders collected in 1 direction and their stops are triggered then 100% of the time price moves in that direction. Not 50% not 90% not 99% but 100%. How far and for how long? It depends on a ton of factors. But suffice it to say you can get the direction 100% of the time when stops are triggered. What is the direction a month from now? Don't know. My answer is what are the positions of retail traders a month from now? But you said you could predict the direction 100% of the time!!! No. Make no mistake you can predict the direction at the time the stops are hit and a little before. An hour before there is no way of knowing and there is no way to trade it an hour after. So your trades are 100% winners then? No. Because you can still be too slow to get on and miss it or you can still hold it for too long thus turning a winner into a loser. And also there is no way to know with certainty where it could end up and how long it will take to get there. It could go 2 ticks then come back to your entry and even stop out before it goes to your target. So trades can have 100% chance of winning in certain situations but anyone can ruin a perfectly good trade. Just because you can get the direction right doesn't mean that you know the direction 100% of the time. You can for sure know the direction after enough short term traders have been collected in 1 direction and their stops are hit.

 

50-50 is relative because there is actually more then 2 possibilities of what can happen in a trade. Some clarification could be maybe someone has small profit targets and takes profit right away. Or maybe some one has a target of 3000 in the S&P and a stop out of 700 and is still in since last May. Again maybe your trades are 50-50.

 

Your stops and targets and how you exercise them determines your percentage. If you don't move a stop or a target then yes its going to be 50-50. A 90% system may just mean they are using a large stop and a small profit for targets. These are the reasons for the percentage. Not the market. The applications of trades to the market make the percentage. I have had lots of trades where I was in a winning trade and it wasn't to my target and I pulled it. It never went to my target and would of turned into a loser. In this situation it defies the 50-50 ideas. I kept a winner from turning into a loser and with out adding or averaging. How do you quantify that? Yes if I traded with a 50-50 mindset and wasn't dynamic and flexible in the market then yes it would of been a loser. However it wasn't and the money went into my account.

 

Ok so you could say that every trade is a winner or a loser. And maybe that was the point all along. However what about a scratch trade? Winner or loser? What about a trade that has some winners and a loser? Do you account for a stop moved to your entry after you take some off or do you keep it at its original place? Are those winners or losers?

 

In closing you can increase your odd of winning or losing by how you exit the trade. When you flip a coin you have to wait for the coin to settle and accept the results. In trading its not that way. You can actually cut the trade when it is at any point of winning and its counted as winning. The equivalent to this with a coin is being able to see what the side of the coin is while its spinning and push it over with your finger.

 

I see what you are saying here but , I still think in 50 -50 . When I put a trade on I know it can only go up or down eventually .....scratch trades I suppose would count but only if you move your stop up to BE +costs.

 

I know we try and create strategies so that will win more often than we lose but when we trade its all 50-50 to me......perhaps there is no right answer I suppose it doe snto really matter anyway .......you either win over time or you don't.

 

Take the lottery for example ....lets say the odds are 1,000,000:1 . Even with those high odds if I actually buy 1 ticket I still have 50% chance of winning....I either win or I don't.

I either get killed driving my car to work or I don't....I either die in a plane crash or I don't......

[BlueHorseshoe 10]

This is interesting, but very hard to believe.

 

Hi ARTjoMS,

 

As far as I'm concerned, this is the right attitude. You're getting purely anecdotal evidence from some random person on a forum who claims to have done some tests which might be riddled with errors, mis-interpreted, or just a plain fabrication.

 

The proper answer to your question is that you need to test this for yourself.

 

As for why I think that runs occur, it's because the markets are not random in the way that a coin flip is - as you state in your original post "price tends to trend". And I think that ultimately the outcome will still be as random because I think that such trends represent 'jumps' in a stochastic process.

 

You can imagine that as a combination of two 1:1 trades.

 

I've imagined all sorts of things in the past, and nearly all of them were wrong. So I stopped imagining and went to the data. In answer to this and any other question you may have, I would urge you to do the same . . .

 

Hope that helps,

 

BlueHorseshoe

[BlueHorseshoe 10]

Even with those high odds if I actually buy 1 ticket I still have 50% chance of winning....I either win or I don't.

 

This sounds great - what lottery do you play?

 

You get one ticket this week, and I'll spring for the other; we'll have a 100% chance of winning between us!

 

I think you're confusing the odds of a particular outcome with the probability that there will be an outcome divided by the number of possible binary outcomes. They're two very different things. Probability is a bugger!

 

BlueHorseshoe

[ARTjoMS 10]

As for why I think that runs occur, it's because the markets are not random in the way that a coin flip is - as you state in your original post "price tends to trend". And I think that ultimately the outcome will still be as random because I think that such trends represent 'jumps' in a stochastic process.

 

Sorry, but did you read what I said? I said that it has nothing to with markets, outcomes could be whatever they are but since you are picking sell with 50% chance and buy with other 50% you are effectively getting BE.

 

Imagine that coin toin toss has 100% chance to get heads and 0% to get tails (I believe this is what you mean by markets are not random). But you randomly pick heads or tails with 50% chance. In result you get 50% chance to win and 50% chance to lose in each coin flip. And this is purely random just like classic coin flip. I hope i shed some light on this.

 

This type of stohastic process (binary+discrete) is called a "Bernoulli process" you can google that for more information and it has nothing to do with jump process.

 

As far as I'm concerned, this is the right attitude. You're getting purely anecdotal evidence from some random person on a forum who claims to have done some tests which might be riddled with errors, mis-interpreted, or just a plain fabrication.

I don't see anyone claiming to have done tests here. But in a sense you are right. It was a bad idea to open this thread.

[BlueHorseshoe 10]

Sorry, but did you read what I said?

 

My goodness, you are persistent and obnoxious, aren't you?

 

In result you get 50% chance to win and 50% chance to lose in each coin flip. And this is purely random just like classic coin flip. I hope i shed some light on this.

 

I see exactly what you mean. But when testing, this wasn't the assumption that I made; I assumed that there was 50% chance of each outcome exactly as with a coin-toss (and consequently did some lazy programming and consistently entered one side of the bet rather than trying to build a randomizing function). So, my results were either:

 

1. derived from a flawed understanding of probability
   
2. made an incorrect interpretation of the results
   
3. aren't relevant to your particular premise
   

Incidentally, going on my argument, the market is not like a coin that is weighted 100% one side, but a coin on which the weighting shifts from one side to the other with random periodicity. This is NOT a Bernoulli process (wikipedia just told me so!).

 

Nevertheless, that still doesn't impact on the scenario you describe.

 

This type of stohastic process (binary+discrete) is called a "Bernoulli process" you can google that for more information and it has nothing to do with jump process.

 

If you're going to start quoting Bernoulli processes at me (with which I'm not familiar, by the way, so thanks for that) then you're evidently a pretty intelligent person (though a little rude). So you should go figure out the answer rather than asking someone with a poorer understanding than yourself.

 

I don't see anyone claiming to have done tests here. But in a sense you are right. It was a bad idea to open this thread.

 

Maybe I wasn't clear - I have done tests.

 

The thread might have been a bad idea for you, but I have learnt something, so I hope that makes it feel a little worthwhile?

 

BlueHorseshoe

Edited February 2, 2013 by BlueHorseshoe

[ARTjoMS 10]

My goodness, you are persistent and obnoxious, aren't you?

 

I am aware of that.

So you should go figure out the answer rather than asking someone with a poorer understanding than yourself.

 

Yes, but that was not my question... this part (that i called as assumption) was "mathematically self evident" to me.

 

Sometimes it is hard to explain yourposition without sounding at least argumentative.

[HARDMAN_GAZ 10]

It is natural to assume that random entry system wirh risk:reward ratio of 1:1 would be BE. (I am sure that someone has proved it)

 

 

So you are just assuming...that someone has proved it!?

 

And under this assumption (that you're not really sure about, and you can't seem to prove), you're going to start acting like the forum math genius.

 

:rofl:

 

Crack on Einstein! I'm loving it!

[TL Staff 10]

Hi guys

 

Let's try to keep everything as civil as possible.

 

While debating an idea is fine, we will not tolerate rudeness, insulting posts, personal attacks, or purposeless inflammatory posts.

 

Thank you.