For me, the optimum way to trade is hybrid trading: a mixture of algo and manual.
What this means is that I have an algo that I can turn on and off, and more. The more is that I have buttons for the different components of all my trading pictures and trade management and I can select which parts of which picture I want to use in a particular context.
I do not believe that its possible to achieve the same profitability with an algo as a proficient discretionary trader. However, with a hybrid algo I can achieve a larger daily consistent profit for less risk than discretionary trading of just one market.
It is not possible for me to program all of the context into an algo and that is why hybrid trading - turning on and configuring the algo for each trade - works best for me. You can see more about this in previous posts and pictures.
Wednesday, January 13, 2016
Tuesday, January 5, 2016
Technology Wins! Algos Take First prize.
I saw an interesting article in this morning's Sydney Morning Herald here. My day starts with reading this newspaper, not only because its the newspaper I grew up reading but also because the international dateline is very near Sydney and the day starts there so by the time I wake up in Europe, the news has been well researched and written and I get a good summary of the day before.
What this article talks about is something that I started believing in 1979 or even before, that technical analysis is the only way to get a true idea of what is happening to a particular instrument in the market and that fundamental analysis is flawed due to latency and hidden information.
There can be no better technical analysis than orderflow and price action in context.Using algos to hybrid trade takes this approach to its logical conclusion. Not using algos to action and see the orderflow puts the trader at a significant disadvantage and at the other side of CP traders' trades.
What this article talks about is something that I started believing in 1979 or even before, that technical analysis is the only way to get a true idea of what is happening to a particular instrument in the market and that fundamental analysis is flawed due to latency and hidden information.
There can be no better technical analysis than orderflow and price action in context.Using algos to hybrid trade takes this approach to its logical conclusion. Not using algos to action and see the orderflow puts the trader at a significant disadvantage and at the other side of CP traders' trades.
Friday, January 1, 2016
Monday, December 28, 2015
Beating the HFTs
Its been estimated that the algos are responsible for almost 70% of the trading volume on futures markets. So as the pit locals disappeared as the pits were made redundant, the algos came in to take their place.
I've been writing about this transition for years and I think we have now come to the new normal although there will probably be more evolution due to regulation and technology changes. Most traders are still trading as if the old normal was still in existence.
Seeing the writing on the wall I started using hybrid trading also some years ago as chronicled in this blog. I defined hybrid trading as using an algo and turning it off and on and choosing different logic for a specific context. This is, I believe, what is required to beat the algos. The speed and reactions of an algo are needed to match the speed and reactions of algos if I want to win trading short time frames. Also, my algo can read and react to the order flow much quicker than a human can.
I have hybrid algos in MultiCharts, Ninjatrader and MarketDelta. The Ninjatrader version makes more efficient use of the PC's resources and now, with Ninjatrader 8, there are microsecond time stamps on each tick so that the order flow can be correctly followed. Each platform is better at something than another platform so it is a matter of using the one that best meets your needs in both functionality and ease of programming.
My algos use a button system that allows me to combine a number of conditions and rules with specific orders that will only execute when everything is true. The pic shows the completed Ninjatrader 7 version which requires tick recording. The Ninjatrader 8 version, is still being developed but has better volume capabilities in addition to correctly tracking order flow.
The functionality needed is a series of buttons that allows me to create a triggerable set of rules that shoot an order immediately those rule conditions are true. I watch the context and when I see the right context I arm a number of the buttons so that the algo can trigger the order from my Virtual Private Server that is co located at the exchange so that I get a less than 2ms latency from the moment that the conditions ar met and the time that the order hits the exchange matching engine.
The series of buttons decribe some context as well as order flow conditions. There are also trade management buttons so that I can have the trade management orders shoot out immediately I have a fill.
I've been writing about this transition for years and I think we have now come to the new normal although there will probably be more evolution due to regulation and technology changes. Most traders are still trading as if the old normal was still in existence.
Seeing the writing on the wall I started using hybrid trading also some years ago as chronicled in this blog. I defined hybrid trading as using an algo and turning it off and on and choosing different logic for a specific context. This is, I believe, what is required to beat the algos. The speed and reactions of an algo are needed to match the speed and reactions of algos if I want to win trading short time frames. Also, my algo can read and react to the order flow much quicker than a human can.
I have hybrid algos in MultiCharts, Ninjatrader and MarketDelta. The Ninjatrader version makes more efficient use of the PC's resources and now, with Ninjatrader 8, there are microsecond time stamps on each tick so that the order flow can be correctly followed. Each platform is better at something than another platform so it is a matter of using the one that best meets your needs in both functionality and ease of programming.
My algos use a button system that allows me to combine a number of conditions and rules with specific orders that will only execute when everything is true. The pic shows the completed Ninjatrader 7 version which requires tick recording. The Ninjatrader 8 version, is still being developed but has better volume capabilities in addition to correctly tracking order flow.
The functionality needed is a series of buttons that allows me to create a triggerable set of rules that shoot an order immediately those rule conditions are true. I watch the context and when I see the right context I arm a number of the buttons so that the algo can trigger the order from my Virtual Private Server that is co located at the exchange so that I get a less than 2ms latency from the moment that the conditions ar met and the time that the order hits the exchange matching engine.
The series of buttons decribe some context as well as order flow conditions. There are also trade management buttons so that I can have the trade management orders shoot out immediately I have a fill.
This type of functionality is available in a number of platforms out there now.
Saturday, December 19, 2015
Nothing Has Changed Except Everything
If you go back to the beginning posts of this blog you will see that I emphasize that the math of your trading plan has to work or you can't be CP.
That has not changed nor can it ever. The three legs of the stool ate win rate, average profit and average loss. Those three metrics have to work to enable CP. However, there are many permutations and computations of these three metrics. basically, a high win rate allows a lower average profit. That's how scalpers make a living, keeping losses small too.
Remember, you can still be marginally CP if your win rate is just 35% as long as the win to loss ratio is 2:1 or greater.
That being said, if one randomly puts on a trade either long or short there should be a 50% chance of profit. If that is the case then just putting on a trade and having your stop loss marginally smaller than your profit target should make you CP as long as commissions are covered too.
So why do most traders find it difficult to get to CP? The answer has to be that they don't follow the trading plan. As someone who is also guilty of that sin more often than I'd like to admit, I understand the problem. If this is the issue then putting the majority of your effort into becoming more disciplined must be at the top of every trader's To Do list.
My focus for the last year has been to use technology to help me better execute my trading plan. Its not the latest indicator. Its not the latest methodology. Its just following order flow and managing my trade so the math can work.
Happy Holidays to everyone and I wish you all a profitable 2016.
PS. The everything that has changed is the markets. No Locals except electronic locals and lots of algo action. But its still just order flow.
That has not changed nor can it ever. The three legs of the stool ate win rate, average profit and average loss. Those three metrics have to work to enable CP. However, there are many permutations and computations of these three metrics. basically, a high win rate allows a lower average profit. That's how scalpers make a living, keeping losses small too.
Remember, you can still be marginally CP if your win rate is just 35% as long as the win to loss ratio is 2:1 or greater.
That being said, if one randomly puts on a trade either long or short there should be a 50% chance of profit. If that is the case then just putting on a trade and having your stop loss marginally smaller than your profit target should make you CP as long as commissions are covered too.
So why do most traders find it difficult to get to CP? The answer has to be that they don't follow the trading plan. As someone who is also guilty of that sin more often than I'd like to admit, I understand the problem. If this is the issue then putting the majority of your effort into becoming more disciplined must be at the top of every trader's To Do list.
My focus for the last year has been to use technology to help me better execute my trading plan. Its not the latest indicator. Its not the latest methodology. Its just following order flow and managing my trade so the math can work.
Happy Holidays to everyone and I wish you all a profitable 2016.
PS. The everything that has changed is the markets. No Locals except electronic locals and lots of algo action. But its still just order flow.
Wednesday, December 16, 2015
Voltage And Current Regulation !!
Voltage And Current Regulation !!
Bonsoir, aujourd'hui j'ai choisi de vous parler un peu de domaine de r�gulation de tension et de courant, Vu a l'importance de ce dernier domaine surtout a l'alimentation de diff�rents composants �lectroniques,
Commen�ons tout d'abord par la notion de r�gulation de tension.
Un r�gulateur de tension en faite un �l�ment capable d'assurer la stabilisation d'une tension � une tension bien fixe, et qui est obligatoire pour les montages �lectroniques qui ont besoin d'une valeur de tension qui est pas directement disponible. Un r�gulateur de voltage est parfois compos� d'une collection de composants �lectroniques simple de base (r�sistances, diodes zener et transistor par exemple), mais il peut aussi �tre de type "int�gr�" ou "avanc�e" et contenir tout ce qu'il faut dans un seul et m�me boitier, pour rendre pratiquant son fonctionnement veuillez bien lire le reste de cet article.
LM = pr�fixe utilis� par le fabricant.
78 = signifie qu'il s'agit d'un r�gulateur positif.
79 = signifie qu'il s'agit d'un r�gulateur n�gatif.
xx = tension de sortie fixe et je dis bien fixe (valeur enti�re sur deux chiffres, par exemple "05" pour 5 V).
Valeurs courantes disponibles : 5 V, 6 V, 9 V, 10 V, 12 V, 15 V, 18 V, 24 V.
Les 1 A ou 1,5 A du bon vieux LM7812 ne suffisent pas ? Plusieurs solutions pour passer un ou deux crans au-dessus, vous pouvez :- utiliser un r�gulateur plus puissant;- ajouter un gros transistor � un r�gulateur;
Bonsoir, aujourd'hui j'ai choisi de vous parler un peu de domaine de r�gulation de tension et de courant, Vu a l'importance de ce dernier domaine surtout a l'alimentation de diff�rents composants �lectroniques,
Commen�ons tout d'abord par la notion de r�gulation de tension.
Un r�gulateur de tension en faite un �l�ment capable d'assurer la stabilisation d'une tension � une tension bien fixe, et qui est obligatoire pour les montages �lectroniques qui ont besoin d'une valeur de tension qui est pas directement disponible. Un r�gulateur de voltage est parfois compos� d'une collection de composants �lectroniques simple de base (r�sistances, diodes zener et transistor par exemple), mais il peut aussi �tre de type "int�gr�" ou "avanc�e" et contenir tout ce qu'il faut dans un seul et m�me boitier, pour rendre pratiquant son fonctionnement veuillez bien lire le reste de cet article.
ce qui est demand� en faite du r�gulateur de tension est de fournir une tension constante pour n�importe quel courant de output quelque soit la charge. ceci est vrai seulement si votre r�gulateur est id�ale pr�sent� au figure si dessus. Alors quoi faire ?
Ceci c'est le d�composition interne d'un r�gulateur form� sur un simple circuit int�gr�.
La valeur output Vout sera autmatiquement compar�e � une tension de r�f�rence fixe qui jouera le role de consigne Vref, par la boucle de contre r�action (feedback).
La sortie de comparateur touchera automatiquement la source de courant qui va polariser le transistor 'BALLAST" de fa�on de r�gler Vout.
Le moment ou la sortie de comparateur sera completement nul c'est le moment ou on a eu la tension de r�ference qu'on cherche.
La valeur output Vout sera autmatiquement compar�e � une tension de r�f�rence fixe qui jouera le role de consigne Vref, par la boucle de contre r�action (feedback).
La sortie de comparateur touchera automatiquement la source de courant qui va polariser le transistor 'BALLAST" de fa�on de r�gler Vout.
Le moment ou la sortie de comparateur sera completement nul c'est le moment ou on a eu la tension de r�ference qu'on cherche.
Les r�gulateurs fixes sont appel�s ainsi parce qu'ils ont �t� con�us pour fournir une tension continue d'une valeur donn�e, qui est non modifiable. Il en existe de multiples sortes sous plein de cat�gories, la famille la plus c�l�bre ici c'est celle qui inclut la s�rie LM78xx (ou uA78xx) et LM79xx (ou uA79xx).
Ils sont tr�s simple � utiliser, et il suffit de peu de connaissances pour savoir le choix correct a utiliser,
leur nom indiquant de lui-m�me de quoi s'agit et de quel voltage j'aurai.
leur nom indiquant de lui-m�me de quoi s'agit et de quel voltage j'aurai.
LM = pr�fixe utilis� par le fabricant.
78 = signifie qu'il s'agit d'un r�gulateur positif.
79 = signifie qu'il s'agit d'un r�gulateur n�gatif.
xx = tension de sortie fixe et je dis bien fixe (valeur enti�re sur deux chiffres, par exemple "05" pour 5 V).
Valeurs courantes disponibles : 5 V, 6 V, 9 V, 10 V, 12 V, 15 V, 18 V, 24 V.
Alors le r�gulateur choisit va d�pendre automatiquement de valeur demand� par la charge de circuit.
Exemple d'utilisation de ces derniers sur ISIS:
Exemple d'utilisation de ces derniers sur ISIS:
Comme �a se voit ici depuis le 15 V d'entr�e j'ai fait sortir le 5v et 6v et le 8v et le 9v... .
Ce r�gulateur inclus une tension d'entr� sur 2 pins et une autre de sortie sur deux pins et puisque une c'est la masse elle est commune ce r�gulateur poss�de 3 pins (IN,OUT,GND). ceci c'est pour le brochage de r�gulateur 78xx(tension postive) j'indique qu'il faut bien faire attention a propos du brochage de r�gulateur de tension n�gative vu que l'ordre des pins est different.Voila l'exemple de 8V positive et 9V n�gative.
Posons que je cherche a avoir une tension stable de 10V et je poss�de que de r�gulateur de 5V comment je fait ?
Ici j'ai besoin de rappeler l'utilit� de la diode zener voici cette exemple
Ici j'ai besoin de rappeler l'utilit� de la diode zener voici cette exemple
Pour avoir une tension de sortie de 10V, il est pr�f�rable d'ins�rer une diode zener de 5,1V entre borne "masse" d'un r�gulateur 5V et masse "r�elle", comme le montre le sch�ma ci-dessus du coup la tension de sortie sera 5V venue de r�gulateur avec le 5.1V de diode zener donc 10.1v.
Les condensateurs que vous voyez de part et d'autre a cot� de r�gulateur, Le choix de leur valeurs comment s'effectue ?
Condensateur C1 : c'est le condensateur responsable au filtrage,et qui permet de "lisser" les arches de sinuso�des pour en obtenir une tension � peu pr�s stable.
Condensateur C2 : ce condensateur sert � am�liorer la stabilit� du r�gulateur et permet une meilleur r�ponse aux transitoires.
Condensateur C4 r�gulateur fixe : ce condensateur additionnel et facultatif doit aussi �tre plac� au plus pr�s du r�gulateur de tension. Il sert � am�liorer la stabilit� du r�gulateur.
Condensateur C2 : ce condensateur sert � am�liorer la stabilit� du r�gulateur et permet une meilleur r�ponse aux transitoires.
Condensateur C4 r�gulateur fixe : ce condensateur additionnel et facultatif doit aussi �tre plac� au plus pr�s du r�gulateur de tension. Il sert � am�liorer la stabilit� du r�gulateur.
Condensateur C3 : ce condensateur le r�le de r�servoir d'�nergie pour la charge.Tout r�gulateur est capable de supporter une certaine tension sur ses poles d'entr�es jusqu'� une valeur fixe bien determin�e . De m�me, tout r�gulateur est capable de d�livrer un courant maximal.
Les valeurs maximales sp�cifi�es par les fabricants ne doivent pas vous laisser penser que l'on peut atteindre ces limites en toutes circonstances. Les limites d'utilisation sont li�es aux contraintes thermiques surtout puisque l'energie dissip�e par effet joule ici est proportionnelle au difference de tension entre la valeur d'entr�e et de sortie.
posins que j'ai des forts courants au circuit mises et qu'il existe une grande s�paration entre l'alimentation et la charge, la chute de tension sera si importante . Cette chute de tension peut �tre compens�e par une m�thode dite "4 pins",
Une autre probl�me s'effectue a l'utilisation de ce genre de r�gulateur c'est l'intervalle de courant �quip�e avec la tension de sortie alors
Les 1 A ou 1,5 A du bon vieux LM7812 ne suffisent pas ? Plusieurs solutions pour passer un ou deux crans au-dessus, vous pouvez :- utiliser un r�gulateur plus puissant;- ajouter un gros transistor � un r�gulateur;- mettre plusieurs r�gulateurs en parall�le.
Au cas ou on va se baser sur la premi�re solution il faut bien choisir le r�gulateur en prenant en compte la valeur de courant ou la capacit� de courante qui est capable a fournir.Au cas ou on va gonfler le courant de sortie il faut utiliser un transistor qui va d�pendre de chemins demand�e par exemple ce circuit l�
L'application simple de loi de noeud ici va vous expliquer comment simplement on peut ajouter un courant au courant de sortie du coup j'ai gonfler le courant de sortie.Exemple alimentation 12V 5A exellente avec les moteurs pas � pas si dessous:
L'application simple de loi de noeud ici va vous expliquer comment simplement on peut ajouter un courant au courant de sortie du coup j'ai gonfler le courant de sortie.
Exemple alimentation 12V 5A exellente avec les moteurs pas � pas si dessous:
Sunday, November 22, 2015
Getting to CP just by tweaking your stops and targets.
I want to revisit an oft mentioned topic that I wrote about back in the early days of this blog such as this post.
A critical part of a trading plan is the trade management. There are three legs to the stool of trade management that are important:
I have often said that a high win rate is important to me. That's a psychological dependency. So my trading style takes that into account.
For example, if I have, say, a 75% win rate then I can have a stop to target ratio of 1: 0.34 and be above break even. If I have a win rate of 34% then I can have a stop to target ratio of 1:2 and still be just above break even.
The math of this is critical for being CP. The style of trading doesn't matter but it needs to be created around these numbers.
So the starting point to your methodology is finding your trading pictures that will make you CP. As, say, a 34% win rate trader you can create a stat of average winners much more than twice your losses then you can make a good living.
If you are not CP then maybe its possible to get to CP by tweaking just this trade management portions of your trading plan. You may have the right entry strategy but are managing your exits incorrectly. But it requires discipline.
A critical part of a trading plan is the trade management. There are three legs to the stool of trade management that are important:
- the stop
- the target
- the win rate
I have often said that a high win rate is important to me. That's a psychological dependency. So my trading style takes that into account.
For example, if I have, say, a 75% win rate then I can have a stop to target ratio of 1: 0.34 and be above break even. If I have a win rate of 34% then I can have a stop to target ratio of 1:2 and still be just above break even.
The math of this is critical for being CP. The style of trading doesn't matter but it needs to be created around these numbers.
So the starting point to your methodology is finding your trading pictures that will make you CP. As, say, a 34% win rate trader you can create a stat of average winners much more than twice your losses then you can make a good living.
If you are not CP then maybe its possible to get to CP by tweaking just this trade management portions of your trading plan. You may have the right entry strategy but are managing your exits incorrectly. But it requires discipline.
Subscribe to:
Posts (Atom)