Here is a bit more detail on this concept as I know many of you are still unclear on how it is possible that 'More contracts == Less risk', especially those who have read books by those such as Van Tharp, etc.. The concept that more contracts can yield less risk goes against everything he and other MM 'experts' teach, in fact goes against everything in standard probability and statistics. And yet there are certain situations where 'More contracts == Less risk' is not only possible but highly probable (you can never say absolutely, as risk is uncertainty and thus inherently uncertain)
Defining risk (uncertainty analysis) is an extremely complicated subject because the nature of uncertainty is chaotic, nonlinear, imprecise, constantly changing. That presents a huge problem for statisticians who like to fit things into nice little all-encompassing models as true uncertainty cannot be modeled in this way (at least not given our present understanding).
One thing which throws a lot of people off is that many educators in the trading community (and also in other fields related to statistics/probability) try to stuff things into these simple models such as binomial (ie tossing a fair die) when they do not really fit. This is why for instance people who fully understand the situation constantly rail against the black scholes pricing model with its gaussian assumption.
It is a very similar situation with MM, there is no such thing as a fair die in the markets, the situation is much more akin to choosing from a selection of loaded die wherein you know there is bias on one side of the current die but there is uncertainty as to which side is favored. Standard statistical methods are incapable of dealing with this problem. That is why if you are looking at it from the frame of standard probability theory, you will not be able to grasp why it is possible that more contracts can yield less effective risk.
(see [0901.2987] Maximum Entropy: The Universal Method for Inference for more thorough discussion if interested)
Essentially what it boils down to is that there is an unknown bias in the expected value. In a multifractal situation like financial markets there are different biases at different levels. Obviously predictability varies widely depending on the market mode (ie trending, ranging, consolidating) but in general it is much easier to predict what will happen on a medium/longer term (ie trend will continue, or trend has ended and will reverse being the most obvious ways to predict)
There are certain situations however where the short term future bias is also quite predictable. So the goal is to find situations where both the short term bias and the longer term bias are deemed predictable and in alignment with each other.
In such a situation more contracts can mean less effective risk because it allows you to get into a position early and capitalize on a potential large movement of hundreds of ticks while at the same time affording you the opportunity of partially scaling out early and thus controlling/limiting your risk in the case that new information reveals your analysis of the longer term bias to be incorrect and you have to close those positions quickly, possibly with slippage.
The key insight here (and what is expounded upon in the Maximum Entropy paper linked above) is that risk is relative and not constant, there is no such thing as fixed risk in the real world. You can choose to look at it that way if you want but that is a limited perspective and not indicative of reality. I think we have all seen the results of what happens when you become too certain of inaccurate models of risk.. ie the Gaussian copula applied to assign risk to CDO's otherwise known as Mortgage backed securities
The following 11 users say Thank You to sefstrat for this post:
Great info Sefstrat. Hope others take the time to read this thread.
Puzzler of the Day:
You are in a trade that has setup nicely and moved 2 ticks in your favor. All of a sudden the market quits moving. Trades are still taking place, but the price is not moving. This goes on for 30 seconds. Has your risk changed?
The following user says Thank You to hondo69 for this post:
BTW in my previous post, I was not trying to imply that standard MM techniques such as those espoused by Van Tharp and others are invalid or do not work, they are valuable techniques and they do work. Its just that they are an oversimplification of the true dynamics of risk/uncertainty and IMO there are more effective ways of looking at the situation.
The risk has changed as it is always changing but with that information I don't think it is possible to quantify exactly how much or in what way, too many other factors to consider.
The following 2 users say Thank You to sefstrat for this post:
selfstrat: some really solid, clear, insightful thinking/exposition there. Thxs.
I am not sure if I get it though. It seems to me that it is not that you reduce 'risk' by trading multiples, rather improve the risk-reward equation, i.e. overall profit expectancy.
On the one hand you initially take more risk, i.e. 3 versus 1 * stoploss. That's obvious.
On the other hand, you greatly increase the chances of getting a greater percentage of the favorable excursion of the trade by scaling out. With 1 contract you have to ensure that your profits exceed losses, whereas with multiples, although of course you still have to do this, you do not have to exit as soon as an initial profit level is reached, initial being that level which, over time, more than pays for stop loss exits and also provides regular profit. With the multiple approach you can take some profit out of the market and then if there is no further extension of the move in your favour, you can get out at break-even. But with one unit, if you were going for more, that break-even exit would mean the entire trade made nothing.
So I think what you are really advocating is this: the ability to scale out of exits improves the risk-reward equation.
Does that sound right or am I missing something here?
The following user says Thank You to cclsys for this post:
Yea you have the idea right it sounds like.. the primary diference is in the way I define risk versus the way most traders do. I define risk like an insurance company whereas most traders look at risk as if they were gambling.
In other words, an insurance company bases risk primarily on uncertainty and is weighted on the likelihood of actually losing the full investment, or in their case having to pay out the full premium. Hence why it is based on uncertainty as the less certain you are about possible outcomes the higher the likelihood of losing your capital is, anyone who has ever had an insurance policy knows that the likelihood of them paying out the full premium is extremely low. Insurance companies would not exist if this were not the case.
On the other hand, a gambler looks at it the situation from a pessimistic perspective, 'I have no clue whats going to happen so lets assume the worst'.
Of course, the reason an insurance company can work the way it does is because it has knowledge the gambler does not have, they do have a clue about what will happen based on past experience and understanding of human nature. If you were to simply increase your contracts and still 'place bets' like the gambler, you will lose all your money rather quickly.
In another thread I made a similar analogy between two blackjack players, one who is a card counter and one who is not. It is the same scenario there, the card counter is still technically risking his entire bet with each hand.. however, his knowledge of the cards remaining in the deck allow him to know when to bet big and when to play it safe.
So basically, my point is learn to think like an insurance company, or a card counter.. the pessismistic 'black/white' view of risk employed by the gambler is very limiting in ways that few traders understand. If you study how the market works and understand how to accurately evaluate the true risk of a given situation, you will have a huge advantage over other traders, actually much bigger than the advantage the card counter has over the house or the other blackjack players.
Moreover, you don't have to be a math genius to be able to do this.. in fact it requires no math at all, all it requires is the understanding of how the markets work. Ie, how the operators, banks, market makers and other institutional players make money and even more importantly you must understand how the average trader thinks (ie the gambler). That part you will have to piece together for yourself, many people make reference to various aspects of how the markets work but you will find very few if any who will paint a whole cohesive picture of it, once you fully understand how the players work it will be immediately obvious to you why this is so.
(first rule of fight club..)
The following 8 users say Thank You to sefstrat for this post:
Also understand how your trading style affects risk. I'll use two examples to illustrate how risk changes over time.
#1] A person enters short with 1 contract thinking the market will move down a few ticks quickly for an easy scalp. After 30 seconds there is little movement in price and the trader has the option to get out at break even now or stay in the trade. The risk on this trade has changed since entry.
#2] A second person has entered shrot at exactly the same time with 3 contracts looking to scale out as the price moves down. Since the market has "frozen" since entry this trader too must evaluate how to manage the change in risk since entry.
In example #1 the trader must make their decision as "all or none" since they only have 1 contract in play. But the 2nd trader has more options and may take 1, 2 or all contracts off the table due to change in risk.
Given the fact that trader #2 has more options available during the trade simply means they have a lower risk over time. It's really no more complicated than that. And if their trading style selects entry points that lend themselves to quick market moves that will in turn reduce their risk exponentially.
The following 2 users say Thank You to hondo69 for this post:
The issue with taking contracts off before it reaches say the same level as your initial risk is a difficult one.
eg you risk 10 pips and take your first lot off at 5 pips trading 3 lots.
What you have to acquire is the no of times the mfe goes 6 ticks ( allowing for the bid offer spread) in the direction of your trade.
Because at 5 vs 10 even if you win 66% of the time you won't make any money on that lot.
That is NOT a problem if you know that 75% of the time your entry is good for 6 pips.
But then you must also look at what if - what if I just kept my lots on and took the further exit(s)?
I almost guarantee that you will make more money but your win/loss rate will fall and the drawdown's will be larger.
What scaling out does do is reduce the no of outright losses to 'fails' ie you make 5 ticks but lose 10 on the other 2 lots - you have actually only lost 15 ticks instead of the 30 you originally took. ( which is 50%)
It's all a question of balance but one that is statistical - you must KNOW on average " I can expect eg 8 ticks from my trade 70% of the time" and therefore I manage my trade accordingly and keep an eye on the stats to make sure that figure doesn't change over time.
It is simply insufficient to know that you win eg 55% of the time.
And if you spot the error in this post I can go one step further...
The following user says Thank You to Mindset for this post:
This is probably not what you were referring to, but using a 10 tick stop is an error in my book. The probability of 10 tick adverse excursion which immediately reverses and moves in your favor is extremely high compared with a larger movement (given that you understand how to identify asymmetric risk entry situations where the likelihood of extended adverse excursion is extremely low)
That is closely related to what I was referring to in the post above when I said you must know how the players make money. If you are using a 10 tick stop you are an 'easy target' for the stop hunters..
The following user says Thank You to sefstrat for this post: