Risk of Ruin Discussion

[russellhq 10]

Flipping a coin and it landing heads or tails is analogous to entering a trade and it going up or down.

 

may i beg to disagree, pls. this is where your brilliant mind errs enormously.... sorry....

 

for an experienced trader, to go long or short is NOT and IS NEVER.... a 50/50 chance occurrence.... like coin tossing which is always a 50/50, unless of course the coin is loaded....

 

for a profitable trader, before pulling the trigger, he/she already picks out the trade before the price is even getting to that point....

 

his/her selected trade does not occur at random, not at all.... the selected trade therefore is never a 50/50 chance occurrence.... do you see what i mean, russell?

whenever the profitable traders choose to trigger.... he/she choose the trade and price according to his/her pre-tested, pre-proven and pre-determined trading plan, statistics, strategy and r/r management.... which already constitutes tremendous trading edge....

 

imho, experienced traders' trigger is much more than a 50/50 chance occurrence.... rebutt if you like....

 

Hi Nakachalet, sorry I did not intend to make out that trading was a 50/50 chance of winning or losing. Although I believe that it can be possible for trades to be a 50/50 chance of success or fail but the successful trader makes more on the winning side than he loses on the losing side. Therefore being a profitable trader.

 

This concept is called "Expected Value", or EV.

 

The EV of a particular event is equal to:

 

((probability of winning)*(Expected Win Amount)) - ((probability of losing)*(expected loss amount))

 

Take the example of a loaded coin, where 60% of the time it lands on heads. I manage to convince someone to play for even money, or for every head I get $10 and for every tail they get $10. The EV of this game for me is (0.6*10) - (0.4*10) = 2. So, on average, for every $10 I bet, I will make $2, therefore after say 100 games, I will be up on average $200. The game has a +EV for me.

[russellhq 10]

OK, I think it's time to move away from the coins as it seems to be irking quite a few of you

 

Lets use some real life trading results. If anyone wants to participate, I will need the following:

 

Average Profit / Average Loss:

Average Profit / Average MAE:

No. Winning Trades / No Losing Trades:

No. Trades per year:

[nvr735i 10]

Nakachalet -- glad you agree, thanks for commenting.

 

 

 

nvr -- absolutely correct, completely relevant -- but incomplete...

 

1) the 35% win-rate thing is usually related to trend-following systems. A well-trodden road -- considered by some/many to be the only valid strategy-type. Proven to work out but with certain caveats, including time and capital -- coming up...

 

2) trend-following has certain characteristics which can make it difficult, or not a good choice for many:

 

(i) It works best as a fairly long-term system. Which means you can/will be in drawdown for long periods. I personally had an 8-month drawdown in such a system, which wasn't even particulary long-term Other, longer-term trend following systems I have tested have shown past drawdown periods measured in years. That doesn't mean months or years of countinuous loss, it means taking that long to reach a new equity high after a drawdown occurs. Not everybody (not many?) can live with this. I have discovered that I can't...

 

(ii) the low win rate means that you spend most of the time 'underwater' -- psychologically difficult for most but, more significantly, a single trade might provide your profit for the whole year. Miss that trade and you're in trouble. Utter consistency required, complete and utter faith in your system, and no vacations...

 

(iii) Perhaps most significant of all, the guys you mention are all managing huge amounts of trading capital. They are thus able to benefit from the diversification inherent in trading many different markets simultaneously. With a small account you will probably be under-capitalised and reduced to trading a single market -- now your equity curve is really up and down.

 

3) there are other trading models, which have a completely different profile -- larger percentage of wins, usually accompanied by a lower win/loss ratio. Smaller trades but more of them. Shorter timeframes. Less dependency on any one trade. Smaller trades means you might be able to trade more markets simultaneously on a given amount of capital.

 

Shorter timeframes (I am not talking about daytrading) have the reputation of being more difficult to trade, which may be true, but they also have benefits -- including shorter times in drawdown or, put another way, quicker to recover from losses, and quicker compounding of your capital.

 

IOW, (a) it can be misleading to quote certain statistics out of context -- a trading system needs to be considered in full context and in relation to all of its inter-related properties and (b) there's more than one way to skin a cat.

 

The way that suits you best may not be the one that suited Richard Dennis.

 

Thanks for the reply Nakachalet,

I wasn't implying that trend following is the only way to trade.Jones and Bacon are macro traders,Simmons is an algorithmic HFT trader,william o'neil is CANSLIM.However what i was trying to point out is that the common factor of all these traders is risk management.That's more important than entry,exit and win/loss percentages yet many traders tend to focus on the latter.As to your comments and experience in trend following ,I humbly disagree.I have traded my own modified trend following and run about 50 million (literally!) monte carlos and simulations and have found that:

 

1) Win Loss ratio is even to slightly below even.

 

2)You don't need a very lage amount of capital to trade trend following systems.It's

a function of what you trade and how much risk you take per position.

 

3)Drawdowns are not very long at all if you don't risk a large percentage per position.If you trade like Richard Dennis then they might be,but like you said not everybody wants to trade like richard dennis(including myself!)

 

4)The profit is not predominantly deriveed from a small number of trades.That is only the case when you trade concentrated positions.If your portfolio heat is spread among 20 or 30 instruments with small risk per trade parameters,then you won't see that phenomenon.There are thousands of ETFs that will do the trick if you don't trade futures(Which I Don't).

 

5)Timeframes are set by the trader,not the system.Certainly a system will have parameters that work better than others but anyone can devise or find a system that suits their timeframe.

 

Lastly, I again reiterate that I was not implying that Trend Following is the only way to make money trading.I trade other styles such as mean reversion as well.What i was trying to get across is that your money management algorithm and position sizing strategy are the most inportant factors that should be considered when trading any system.Also,you might want to try running your trend following tests using smaller risk levels and more instruments.You'll probably see that the returns and drawdowns relation to each other improves not assymetricly with the percentage adjustment down in risk.In fact in most cases,risk to drawdown improves nicely.Just food for thought. All the best,Nick

[nvr735i 10]

Nakachalet -- glad you agree, thanks for commenting.

 

 

 

nvr -- absolutely correct, completely relevant -- but incomplete...

 

1) the 35% win-rate thing is usually related to trend-following systems. A well-trodden road -- considered by some/many to be the only valid strategy-type. Proven to work out but with certain caveats, including time and capital -- coming up...

 

2) trend-following has certain characteristics which can make it difficult, or not a good choice for many:

 

(i) It works best as a fairly long-term system. Which means you can/will be in drawdown for long periods. I personally had an 8-month drawdown in such a system, which wasn't even particulary long-term Other, longer-term trend following systems I have tested have shown past drawdown periods measured in years. That doesn't mean months or years of countinuous loss, it means taking that long to reach a new equity high after a drawdown occurs. Not everybody (not many?) can live with this. I have discovered that I can't...

 

(ii) the low win rate means that you spend most of the time 'underwater' -- psychologically difficult for most but, more significantly, a single trade might provide your profit for the whole year. Miss that trade and you're in trouble. Utter consistency required, complete and utter faith in your system, and no vacations...

 

(iii) Perhaps most significant of all, the guys you mention are all managing huge amounts of trading capital. They are thus able to benefit from the diversification inherent in trading many different markets simultaneously. With a small account you will probably be under-capitalised and reduced to trading a single market -- now your equity curve is really up and down.

 

3) there are other trading models, which have a completely different profile -- larger percentage of wins, usually accompanied by a lower win/loss ratio. Smaller trades but more of them. Shorter timeframes. Less dependency on any one trade. Smaller trades means you might be able to trade more markets simultaneously on a given amount of capital.

 

Shorter timeframes (I am not talking about daytrading) have the reputation of being more difficult to trade, which may be true, but they also have benefits -- including shorter times in drawdown or, put another way, quicker to recover from losses, and quicker compounding of your capital.

 

IOW, (a) it can be misleading to quote certain statistics out of context -- a trading system needs to be considered in full context and in relation to all of its inter-related properties and (b) there's more than one way to skin a cat.

 

The way that suits you best may not be the one that suited Richard Dennis.

 

BTW,sorry for the name mixup,but I was doing 3 things at once!LOL

[russellhq 10]

I ran a few simulations based on the following premise:

 

The game is breakeven ie EV is 0

The success rate is varied from winning 10% of the time to winning 90% of the time

We play the game 100 times and accumulate our results after each play.

The accumulation of 100 plays is repeated 1 million times and the MAE is counted each time.

For each winning %, I look up the MAE that would occur roughly once in 10,000 times.

 

The results look like this:

 

 

So basically this confirms what we already suspected. High win% games have low drawdown and low win% games have high drawdown for the same EV. This is shown by a MAE of only 14 when we have a 90% win rate and a MAE of 89 when we have only a 10% win rate. (remember that the outcome on average for each win% is 0 i.e. EV=0)

[MightyMouse 12]

but mightymouse....

 

many supposedly trade forex without stops....

 

at least that is what they claim....

 

perhaps, most are brokers advising their acct holders that it is better to trade without stop.... so they can make more money, quicker and easier.... :crap:

 

really, in terms of reality, i do not really know who is right or who is wrong....

 

and not that i care much if anyone is trading any product with or without stops....

 

i can only say that i trade with only 9 tics stop and sometimes try to scale down to a 6 tics stop too.... lol YES, I USE STOP TO LIMIT MY RISK EXPOSURE FOR EACH AND EVERY TRADE.... AND I AM STILL HERE TRADING.... HAPPILY, PROFITABLY AND CONTINUOUSLY.... hopefully for many more years to come....

 

i guess it must be mighty exciting to go into any biz without knowing what the limit is going to be and what or how much you are personally responsible for.... yes, that is exciting.... for many.... :missy:

 

When I say "trade" I am referring to scalping, day, or short term trading. If you trade long term, then I consider that more of a investment than trading.

 

Day trading without stops is just an accident waiting to happen. If you wait long enough, it will happen. Or put another way; the probability of something not happening is only relevant in the short run. In the long run, everything happens.

[nvr735i 10]

I ran a few simulations based on the following premise:

 

The game is breakeven ie EV is 0

The success rate is varied from winning 10% of the time to winning 90% of the time

We play the game 100 times and accumulate our results after each play.

The accumulation of 100 plays is repeated 1 million times and the MAE is counted each time.

For each winning %, I look up the MAE that would occur roughly once in 10,000 times.

 

The results look like this:

 

 

So basically this confirms what we already suspected. High win% games have low drawdown and low win% games have high drawdown for the same EV. This is shown by a MAE of only 14 when we have a 90% win rate and a MAE of 89 when we have only a 10% win rate. (remember that the outcome on average for each win% is 0 i.e. EV=0)

 

Forgive me if I missed something but if the payout is always the same,isn't it obvious that

if you increase the win percentage,then you'll earn more and drawdown less?I guess i'm just missing the significance of the graph.How about if you have a game that has an 80% win rate of $.10 and a 20% loss rate of $.50.You have a high win probability but negative expectancy.

[russellhq 10]

Hi nvr735i, I mentioned the EV was 0, I never mentioned the payout was the same. That is, for the 90% game, for every dollar you win, you lose $9. And again, for the 20% game, for every dollar you win, you lose only $0.25. I hope that is clear now.

[tommaso 10]

Guys,

All I can say is that trading without stops is almost a sure way of guaranteeing you will be wiped out at some point.I've done alot of research on this topic.In this thread i see the QUOTE]

 

You have done no research. Trust me.

If so, tell me how can we trade without stops ?

Can you tell me anything alternative to using stops?

How do people who don't use stops, hedge ?

 

Do you think they open a trade and let it go until ruin ? haha

 

Trading with no stops has nothing to do with trading without any form of "protection" or risk management. On the contrary, hedging is achieved in a much more advanced way, by "strategy overlay".

 

Clearly, if we start with the mindset of coin tossing. Make an entry and place a tp and stop, we certainly don't go anywhere.

It is easily proven that any algorithmic strategy based on that is unprofitable.

 

Smart discretionary traders can survive for a while because a human has a better capability to determine a good entry point (local high/low).

 

But if we talk about strategies and algorithm, stops have really no place.

We achieve hedging by overlaying strategic layers, and this can be proven effective with rigorous controlled experiments. Forget the coin toss. It's like talking to a 2 yo child.

 

Clearly, if you tell him, he will not know what you are talking about .... so i will take anything you say. But just remember, it may come useful in a far future ...

 

Tom

[MightyMouse 12]

Guys,

All I can say is that trading without stops is almost a sure way of guaranteeing you will be wiped out at some point.I've done alot of research on this topic.In this thread i see the QUOTE]

 

You have done no research. Trust me.

If so, tell me how can we trade without stops ?

Can you tell me anything alternative to using stops?

How do people who don't use stops, hedge ?

 

Do you think they open a trade and let it go until ruin ? haha

 

Trading with no stops has nothing to do with trading without any form of "protection" or risk management. On the contrary, hedging is achieved in a much more advanced way, by "strategy overlay".

 

Clearly, if we start with the mindset of coin tossing. Make an entry and place a tp and stop, we certainly don't go anywhere.

It is easily proven that any algorithmic strategy based on that is unprofitable.

 

Smart discretionary traders can survive for a while because a human has a better capability to determine a good entry point (local high/low).

 

But if we talk about strategies and algorithm, stops have really no place.

We achieve hedging by overlaying strategic layers, and this can be proven effective with rigorous controlled experiments. Forget the coin toss. It's like talking to a 2 yo child.

 

Clearly, if you tell him, he will not know what you are talking about .... so i will take anything you say. But just remember, it may come useful in a far future ...

 

Tom

 

The coin toss allows you to examine certain aspects of trading in a controlled environment. No one but you is suggesting that it is a duplicate of reality.

[nvr735i 10]

Tom

You can trade as you like,but you should know that algorthmic trading is based on determining a mathematical exit or stop.Every major disaster from long term capital to lehman brothers was due to the lack of risk management.Tell me something ,if you have no stops,how do you decide where to get out?If you decide mentally than you're still using a mental stop.You can trade without stops obviously.However history has shown that trading on emotions (fear,greed) has been the worst way to trade.I'll send you the chart of inflows and outflows for the last 100 years if you like.By the way,no algorithmic trader would ever run a monte carlo sim on a coin toss for an actual system.It's a total waste of time.If you're using a strategic overlay as you said and that tells you where to get out,it's a stop.Just because you don't place it on the books it doesn't mean it doesn't exist.By the way also if you have a loss and don't sell,it's still a loss.Your alternative to trading without stops IS to trade without stops.if you're comfortable with that,i'm happy for you.I have no idea what you're talking about with the coin flip and getting nowhere comment.Nobody trades on a coin flip.We use actual intraday market data to determine if a system will be profitable or not.The good discretionary trader is a human with feelings.That's a disadvantage when it comes to trading.It doesn't mean you can't trade discretionarily.Lastly,it has been proven that entry/exit is one of the least important factors when trading.It's only about 10% of a system.Do some empirical research on the topic and i'm sure you'll come up with the same conclusion.Good luck!

P.S. By the way ,i've done several years of research on this ,contrary to your belief.I have

an office full of data that i can show you.We can compare your results to my results on

stops if you like!I hope you have empirical data to present your case.

Edited August 17, 2011 by nvr735i

[nvr735i 10]

By the way most algorithmic trading also sets up for entry,and position sizing as well asdefining system parameters.

[rdhtci 10]

 

The coin toss allows you to examine certain aspects of trading in a controlled environment. No one but you is suggesting that it is a duplicate of reality.

 

...but people are constantly taking results derived from coin-toss examples and using those results to make decisions about their trading. Which is attempting to correlate it with reality.

 

The coin-toss is useful for demonstrating the variance of returns. Whilst we all know that over a high enough number of coin tosses the result is going to come out very close to 50-50, it is truly surprising how far away from 50-50 the experience can be over small numbers -- in other words, strings of consecutives wins -- or more importantly, losses -- can be far, far bigger than you would intuitively expect.

 

Somebody else in this thread referred to the same phenomenon on a Vegas roulette table -- something like 16 straight wins/losses.

 

It is simple enough to calculate the probability of a given number of consecutive results for any win rate percentage, but you had better not put your hard-earned cash on it. For example, the probabilty of 10 consecutive wins or losses on a 50-50 proposition (a coin-toss) is 0.98% -- less than 1% probability.

 

Personally, nothing would induce me to use that figure for anything related to my trading capital.

[nvr735i 10]

Very ,Very true! The supposedly once in a lifetime event seems to happen every few years!

[russellhq 10]

 

Whilst we all know that over a high enough number of coin tosses the result is going to come out very close to 50-50, it is truly surprising how far away from 50-50 the experience can be over small numbers -- in other words, strings of consecutives wins -- or more importantly, losses -- can be far, far bigger than you would intuitively expect.

 

I don't think that is true at all. Let me ask you, what is the probability that you will break even after 1,000,000 coin tosses? And, if we increase the number of tosses, are we converging to break even?

[rdhtci 10]

I don't think that is true at all. Let me ask you, what is the probability that you will break even after 1,000,000 coin tosses? And, if we increase the number of tosses, are we converging to break even?

 

Both statements are completely true.

 

The law of large numbers states that actual results become closer to the expected as the number of instances increases. In other words a coin-flip produces closer to 50-50 the more times you flip. This has absolutely nothing to do with breakeven. The concept of breakeven introduces another factor, the size of a bet, which allows the distribution of returns to affect the result. If you introduce beakeven, your subject is no longer the number of heads or tails produced, it has become how much money has been made or lost.

 

Conversely, over a small number of coin tosses, a string of consecutive heads or tails can be far larger than the expected outcome would suggest -- leading to a short-term result further away from 50-50 than you might expect.

 

I have witnessed a small experiment where a number of series of 100 coin flips was performed and recorded. The strings of consecutive heads and tails were eye-opening. I have also traded long enough to see such strings in my own results. Sooner or later, everybody will experience this. Anyone who hasn't seen it has probably not consistently traded the same method for long enough.

 

This latter phenomeneon, the larger-than-expected string of losses, is precisely what connects the coin-toss to concepts of risk management. That and the position size element are precisely what combine to produce ruin. Put simply, a position size which is too large can combine with a larger-than-expected string of losses and wipe you out.

 

There really is no argument about this, and it embodies the precise problem faced by those trying to calculate the largest possible position size. The black swan is a much used phrase which has come to be incorrectly used as an analogy for a bad day in the market. It's not that at all - it refers to an event which has not previously happened in our experience. In other words, not just a larger-than-expected string of losses, but a string of losses which is so large that it has never been known to occur before.

 

That event is out there, it's in your future. Any calculations related to risk need to acknowledge this.

 

Again, it's importance is related to your personal circumstances -- specifically, it's related to account size. Without being in any way flippant <g>, it's a different matter trading an account that was funded with $500 from an extra job flipping burgers than it is trading a much larger account that represents 20 or 30 years of carefully accumulated capital. Or a large amount of capital representing other people's money. One is much more easily replaced than the other, and you might be happy accepting larger risk.

 

Which is precisely why professional traders trade at low risk levels. Inexperienced traders tend to push out the boat because they are unaware of these factors. They also tend to blow out their account.

 

One of the beautiful things about trading is that you create your own circumstances. But it's a double-edged sword. I am simply suggesting that anyone considering risk management should be aware of these simple elements.

[russellhq 10]

rdhtci, both statements are completely true but completely irrelevant

 

What is relevant, and what you skipped over, is the accumulation of results (which is what my subject has been all along).

 

It's this accumulation that affects traders and we see it in our draw down.

 

So, I'll ask again. If I flip a coin 1,000,000 times and starting at 0, I add 1 for every head and subtract 1 for every tail, what's the probability of my sum returning to 0? And something more relevant to draw down, what's the probability that at some point during the count my count will reach -100 or less?

[tommaso 10]

 

...but people are constantly taking results derived from coin-toss examples and using QUOTE]

 

[that wasn't me actually, but a quote from the powerful mouse ;-) ]

[tommaso 10]

Tom

You can trade as you like,but you should know that algorthmic trading is based on determining a mathematical exit or stop

...

By the way,no algorithmic trader would ever run a monte carlo sim on a coin toss for an actual system.It's a total waste of time.If you're using a strategic overlay as you said and that tells you where to get out,it's a stop.Just because you don't place it on the books it doesn't mean it doesn't exist.By the way also if you have a loss and don't sell,it's still a loss.Your alternative to trading without stops IS to trade without stops

....

.

 

Right, coin toss is a too simplified and useless model to make reference to.

A more useful model could be the GBM with mean reversion, and then one could just start making some sense, at least with the precaution to test the strat with **different volatility** levels (as mkt as a non-costant volatility obviously):

 

ModelAssist for ModelRisk

 

Algorithmic trading stretches much beyond a naive immagination can reach. There are ways to "overlay" different strategies (for instance a trender against a countertrender) in such a way that one strategy "protects" the other one, and both run without stops. There is also the concept of "artificial option", a conceptual device created using the folio instruments. No need to say that this is much better, and can be proved experimentally, than attempting to use stops at single trade level, which is an absolutely sure way to leave all the money in the mkt, in the long run (i am talking of algorithms). As said, smart traders can still trade with stops and survive a little longer (sometimes more than their lifetime), due to a better choice of high probability "trades" (news processing, past experience, instinct, understanding of international economy, local high/low perception, etc.).

 

Tom

Edited August 18, 2011 by tommaso

[rdhtci 10]

 

...but people are constantly taking results derived from coin-toss examples and using QUOTE]

 

[that wasn't me actually, but a quote from the powerful mouse ;-) ]

 

Ha ha - no, tommaso, I know it wasn't you -- I was quoting the mouse quoting you and actually responding to the mouse

[rdhtci 10]

rdhtci, both statements are completely true but completely irrelevant

 

What is relevant, and what you skipped over, is the accumulation of results (which is what my subject has been all along).

 

It's this accumulation that affects traders and we see it in our draw down.

 

So, I'll ask again. If I flip a coin 1,000,000 times and starting at 0, I add 1 for every head and subtract 1 for every tail, what's the probability of my sum returning to 0? And something more relevant to draw down, what's the probability that at some point during the count my count will reach -100 or less?

 

 

russellhq -- we're going round in circles here. I responded to your post which said 'I don't think that's true...' by explaining exactly why what I said is true. Now you say it's true but irrelevant. Maybe you should make up your mind if you think it's true or not. Given that we're talking about the law of large numbers, that much is pretty simple - it's unequivocally true. If you think it's irrelevant, that's your prerogative, but I would caution you to think again.

 

And I haven't skipped over what you term 'accumulation of results' - quite the reverse, it's exactly what I have been talking about. My posts have made several references to 'distribution of returns' and 'variance of returns'. In this context, 'distribution of returns' means how your wins and losses are grouped, 'variance of returns' refers to how far away from the average your returns are at any one time. Seems to me that these concepts are exactly what you are discussing.

 

I am trying to point out to you that no amount of mathematical calculation will stand you in any stead when your 'distribution' or 'variance' are suddenly a mile away from anything you have previously encountered, or anything you were able to predict, or anything your math calcs suggested... Or anything you were prepared for. If you don't work out how to prepare for that, you're out of the game. Ruin. If you think that's irrelevant to what you are discussing, I can only wish you luck.

 

As for your question about the results of 1,000,000 coin tosses under specified conditions -- I don't know. There's no point in calculating that because THAT is irrelevant. Why is it irrelevant? Because of everything I have been saying -- you have absolutely no way to determine your future returns -- not their average, not their distribution, not their variance. There is no correlation between real-world trading and anything concerning a coin toss EXCEPT that there will be variance in both. Which, in future trading, can only be unknown until after the event. That is precisely why approaching this problem from the angle of calculating the largest position size which will not cause ruin is fundamentally flawed...

 

I think I'll leave it there - I've tried to clarify, I don't think I can do any better -- if you think it's irrelevant, best of luck -- sincerely.

[russellhq 10]

Hi rdhtci, what I am getting at is; it is not the long run of losers or long run of winners that affect the variance but the probability of winning that has the biggest impact. I tried to demonstrate this by graphing the variance of different winning probabilities with 0 EV. It show the variance of 1 in 10,000 events. For example, using a fair coin and flipping it 100 times, there is a 1 in 10,000 chance that at some point along the way you will be down 38 or more points. You don't need 38 losers in a row to achieve this, just 38 more losers than winners at some point during the 100 flips, this makes things far more likely!

 

I said the law of large numbers was irrelevant to the trader because the trader will never trade long enough for it to make a difference. The variance experience over a finite set is far more relevant and is what I am discussing.

 

I am trying to point out to you that no amount of mathematical calculation will stand you in any stead when your 'distribution' or 'variance' are suddenly a mile away from anything you have previously encountered, or anything you were able to predict, or anything your math calcs suggested... Or anything you were prepared for. If you don't work out how to prepare for that, you're out of the game. Ruin. If you think that's irrelevant to what you are discussing, I can only wish you luck.

 

I completely agree being prepared is essential.

 

Because of everything I have been saying -- you have absolutely no way to determine your future returns -- not their average, not their distribution, not their variance.

 

This I have trouble with. If you have no way to predict future events then how can you possibly asses your risk and make a low risk trade?

 

As for your question about the results of 1,000,000 coin tosses under specified conditions -- I don't know. There's no point in calculating that because THAT is irrelevant. Why is it irrelevant?

 

It is relevant in the same way the coin toss experiment you witnessed was an eye opener. It gives you a whole new perspective!

[nvr735i 10]

Tom,

I don't see how having a stop as an exit can leave more money in the market than trading without stops!The very purpose of the stop is to have an exit that will help you quantify your risk.

As to your comments that a good trader can decipher news,trade on instinct,etc...,if that's your belief,i wish you well.As far as i'm concerned noone can predict or decipher anything that will

tell them where a security price will be in the future.All we can do is have a setup for entry,layout our best case and worse case scenario,hopefully preset our risk,and wait for the trade to either go one way or the other.That's how i trade and will continue to trade.I'm very happy knowing my portfolio heat at any given moment(barring a black swan or fat tail event).It helps me sleep at night and live to fight another day.So,I don't see a case for argument.I say to each his own and good luck!

[nvr735i 10]

Both statements are completely true.

 

The law of large numbers states that actual results become closer to the expected as the number of instances increases. In other words a coin-flip produces closer to 50-50 the more times you flip. This has absolutely nothing to do with breakeven. The concept of breakeven introduces another factor, the size of a bet, which allows the distribution of returns to affect the result. If you introduce beakeven, your subject is no longer the number of heads or tails produced, it has become how much money has been made or lost.

 

Conversely, over a small number of coin tosses, a string of consecutive heads or tails can be far larger than the expected outcome would suggest -- leading to a short-term result further away from 50-50 than you might expect.

 

I have witnessed a small experiment where a number of series of 100 coin flips was performed and recorded. The strings of consecutive heads and tails were eye-opening. I have also traded long enough to see such strings in my own results. Sooner or later, everybody will experience this. Anyone who hasn't seen it has probably not consistently traded the same method for long enough.

 

This latter phenomeneon, the larger-than-expected string of losses, is precisely what connects the coin-toss to concepts of risk management. That and the position size element are precisely what combine to produce ruin. Put simply, a position size which is too large can combine with a larger-than-expected string of losses and wipe you out.

 

There really is no argument about this, and it embodies the precise problem faced by those trying to calculate the largest possible position size. The black swan is a much used phrase which has come to be incorrectly used as an analogy for a bad day in the market. It's not that at all - it refers to an event which has not previously happened in our experience. In other words, not just a larger-than-expected string of losses, but a string of losses which is so large that it has never been known to occur before.

 

That event is out there, it's in your future. Any calculations related to risk need to acknowledge this.

 

Again, it's importance is related to your personal circumstances -- specifically, it's related to account size. Without being in any way flippant <g>, it's a different matter trading an account that was funded with $500 from an extra job flipping burgers than it is trading a much larger account that represents 20 or 30 years of carefully accumulated capital. Or a large amount of capital representing other people's money. One is much more easily replaced than the other, and you might be happy accepting larger risk.

 

Which is precisely why professional traders trade at low risk levels. Inexperienced traders tend to push out the boat because they are unaware of these factors. They also tend to blow out their account.

 

One of the beautiful things about trading is that you create your own circumstances. But it's a double-edged sword. I am simply suggesting that anyone considering risk management should be aware of these simple elements.

 

You summed it up perfectly.We all probably haven't seen our worst drawdowns.However we can lower the probabilty of ruin by controlling position size and portfolio heat.

[nvr735i 10]

russellhq -- we're going round in circles here. I responded to your post which said 'I don't think that's true...' by explaining exactly why what I said is true. Now you say it's true but irrelevant. Maybe you should make up your mind if you think it's true or not. Given that we're talking about the law of large numbers, that much is pretty simple - it's unequivocally true. If you think it's irrelevant, that's your prerogative, but I would caution you to think again.

 

And I haven't skipped over what you term 'accumulation of results' - quite the reverse, it's exactly what I have been talking about. My posts have made several references to 'distribution of returns' and 'variance of returns'. In this context, 'distribution of returns' means how your wins and losses are grouped, 'variance of returns' refers to how far away from the average your returns are at any one time. Seems to me that these concepts are exactly what you are discussing.

 

I am trying to point out to you that no amount of mathematical calculation will stand you in any stead when your 'distribution' or 'variance' are suddenly a mile away from anything you have previously encountered, or anything you were able to predict, or anything your math calcs suggested... Or anything you were prepared for. If you don't work out how to prepare for that, you're out of the game. Ruin. If you think that's irrelevant to what you are discussing, I can only wish you luck.

 

As for your question about the results of 1,000,000 coin tosses under specified conditions -- I don't know. There's no point in calculating that because THAT is irrelevant. Why is it irrelevant? Because of everything I have been saying -- you have absolutely no way to determine your future returns -- not their average, not their distribution, not their variance. There is no correlation between real-world trading and anything concerning a coin toss EXCEPT that there will be variance in both. Which, in future trading, can only be unknown until after the event. That is precisely why approaching this problem from the angle of calculating the largest position size which will not cause ruin is fundamentally flawed...

 

I think I'll leave it there - I've tried to clarify, I don't think I can do any better -- if you think it's irrelevant, best of luck -- sincerely.

 

once again,i agree.That's why many traders stopped using the Kelly Criterion to determine optimal position sizing to maximize return.It's way too aggressive in terms of sizing as far as i'm concerned.