Anti-martingale with MTP

[jtrade]

Matt,

In the article linked in your Daily Report received this morning Van Tharp writes :

Anti-martingale systems, where you increase your risk when you win, do work.

Do you think this applies to positive expectarncy, relatively low win-rate methods such as with MTP ? I can see the substantial potential benefits from increasing size whist winning & reducing whilst losing, but have never tried this with real money in real time because it seems to me the only prudent environment in which to do this would be with at least 50% winners.

Very interested in your opinion, please.

Best,

J.

[tombtrader]

Hi J

Isn't this what position sizing is? As you win, your account size increases and the 2% you risk will be more dollars?

[jtrade]

Hi J

Isn't this what position sizing is? As you win, your account size increases and the 2% you risk will be more dollars?

No : anti-martingale means that during winning streaks you actually increase your risk, eg. from 2% to 3% per trade, or incrementally more as the streak continues, and vice-versa for losing streaks.

J.

[Eddo]

No : anti-martingale means that during winning streaks you actually increase your risk, eg. from 2% to 3% per trade, or incrementally more as the streak continues, and vice-versa for losing streaks.

J.

If I could put my two penneth worth in here - I believe that position sizing based on a %age (i.e. 2%) of the CURRENT capital value - which is what we all believe here I hope - is an anti-martingale system - in that we are increasing the size of our position as the pot gets bigger and decreasing it when the pot gets smaller.

So as far as I am concerned this IS an anti-martingale system - and it works :)

[jtrade]

Hi, Chris,

I think what you and tt are referring to is simple (and very good) Fixed % MM. Whilst this does indeed increase position size with account growth, my understanding is that the term anti-martingale is used to describe something very much more aggressive, which attempts to increase total leverage exponentially, rather than maintain that fixed %.

In his MM papers, not all of which I have read, VT gets into some quite interesting angles on MM - eg. greatly increasing leverage on the part of one's account balance consisting of profits above the intial equity. In the context of this thread, I think he is talking about leveraging up winning streaks beyond that which takes place with a Fixed % MM.

But, then again, I could have this all back to front (it wouldn't be the first time...).

:D

J.

[chrisg]

I just wanted to point you guys to a system that uses this method. On your first entry you risk 1% if it is a winning trade you then risk 2.5% the if the next is winning you risk 5% next 7.5% then 10% after 4 consecutive winning trades. If you lose one trade then you start back at 1%. This does have a positive expectancy when back tested on some systems. To look at a system that uses this you can go to www.wealth-lab.com. One system is called dynamic position sizing. When tested over a 10 year period on their selected stocks it had a net return of 366.17%. I have looked into this method of position sizing but have stayed away from it mainly due to the fact most methods your risk is not the distance from your entry to your stop but the whole amount you accualy place on the trade. For example if you have a $100,000 account you only can buy $1,000 worth of stock. I hope this helps


chris:rolleyes:

[Steve Griffiths]

Hi Guys,

A very interesting thread here......

From my Angle would suspect that this is a dangerous approach to take, because I know (from experience) that after a few winners you are more likely to have a losing trade, so in this example, you will have a larger initial risk when this loss comes through, therefore making it larger than necessary. I could see this wiping out all the previous losses.

As such, I would think, while it may be tempting, in reality it would mean that losses would not be kept small and under control.

Just my thoughts

Steve

[chrisg]

Believe it or not the thinking behind this method is once you have a winning trade the probability of the next being a winner is greater than 50%. I have to admit that I do not buy into this completely. In my real world experience I usually do not have more than three consecutive winning trades. My advice is to keep it simple keep losses as small as possible and let profits run.


Chris

[Matt Bowen]

Hi Jtrade

No : anti-martingale means that during winning streaks you actually increase your risk, eg. from 2% to 3% per trade, or incrementally more as the streak continues, and vice-versa for losing streaks.

Just be careful how you have this worded...

As you win the bet size increases (in per portion to the account size). He's only betting more because he won on the last trade, if he lost, he decreases the position size (amount not the percentage). Just remember the position size percentage never changes, the amount or size of lots changes and that is what decides whether he add or decreases the next bet size.

[Matt Bowen]

Hi Chris,

If I could put my two penneth worth in here - I believe that position sizing based on a %age (i.e. 2%) of the CURRENT capital value - which is what we all believe here I hope - is an anti-martingale system - in that we are increasing the size of our position as the pot gets bigger and decreasing it when the pot gets smaller.

So as far as I am concerned this IS an anti-martingale system - and it works

Spot on :D

[Matt Bowen]

Hi Chrisg,

I just wanted to point you guys to a system that uses this method. On your first entry you risk 1% if it is a winning trade you then risk 2.5% the if the next is winning you risk 5% next 7.5% then 10% after 4 consecutive winning trades. If you lose one trade then you start back at 1%. This does have a positive expectancy when back tested on some systems.

I think you would be better off spending 50 bucks and go ride a rollor coaster. I don't think you are going to find very many professional money managers using this and I highly doubt their clients want to take a Xanax every time they read their income stements.

It reminds me of Van Tharp's secrets of the masters position sizing you will get a taste of how this works on levels 9 and 10 when some of the variables are removed. He does this not so you will employ the technique, but rather it forces you to understand what you are doing.

I think if a lot more people went through this exercise they would comprehend risk management (much more than the think they do). The levels 1-4 (free) don't really show you anything other than the basic calculations. It's the discipline to keep applying the position size in the face of a losing streak that makes them a winner and that's what he's trying to teach.

[stylist]

How often do you need to change your account balance in the software to make the best of position sizing after every trade? or after every trading day ?

[Matt Bowen]

Hi stylist,

How often do you need to change your account balance in the software to make the best of position sizing after every trade? or after every trading day ?

After each trade...

[Matt Bowen]

There maybe some confusion here as to what exactly is a Martingale and Anti-Martingale bet strategy:

Professional gamble's have long claimed that there are two basic position-sizing techniques – martingale and anti-martingale. With a martingale strategy the gambler will increase the bet size when equity decreases (when the person loses money). In anti-martingale you do the opposite; you decrease the bet size when you lose money. In the simplest form of a martingale strategy you double your bet size every time you lose.

The goal with position sizing is to increase your exposure as your account grows (you do not not increase the % on the position size) and limit your exposure when you go through a series of losing trades. This technique is called an anti-martingale strategy.

I think the best way for people to under stand the difference between Martingale and Anti-Martingale is to look at the insanity of how the Martingale works in a casino.

For this example we will assume you are playing on a five-dollar table. First of all most casinos have a maximum bet. Many five-dollar tables have a maximum bet of $500. If you bet using the martingale where you double after every loss the progression would look like this:

Your next bet is $10. If you lose:
Your next bet is $20. If you lose:
Your next bet is $40. If you lose:
Your next bet is $80. If you lose:
Your next bet is $160. If you lose:
Your next bet is $320. If you lose:
Your next bet is $640, which exceeds the maximum bet for the table.

It would only take eight losses to exceed the table maximum bet. It doesn't matter how many times you double up your bet. You will up you will only win five dollars if you eventually hit. This takes into account all your progressions up to that point. Even if you could exceed the $500 maximum bet, on the eighth bet you are risking $640 to win $5. You have already invested $635 for your previous seven bets. If you lose that one you are out $1,275. Can you imagine risking over a thousand dollars for a chance to win five dollars? The casinos know that if someone had unlimited resources they would eventually win. That is why the set maximum bets at the table.

Remember: House odds in a casino are very different for each game For example: The roulette wheel has no memory...so if you spin 12 red numbers the odds do not increase because you are now on the 13th spin. Now, if you happen to be playing Blackjack you can turn the odds in your favor by playing basic stategy and counting the face cards in a six shoe deck, but this is also a good way to get thrown out of a casino because if they catch you...they will throw you out. My point is that each game has different odds.

(notice that the bet is a zero sum game in a casino) Trading is not a zero sum game...it's a minus sum game because you still have to cover commissions.

(Now you know why they have table limits in Casinos). The casino is not stupid, they know that at some point you are going to get a winner if you have big enough bankroll they will eventually lose to a martingale player without any table limits.