Algos watch Patterns, Trendlines, Channels, Support and Resistance

[Laissez Faire]

I think this is a really bad question to understand if a strategy is profitable.

Let's not get lost in semantics. I'd be happy to rephrase the question as consistently profitable.

 

[tsfx]

Let's not get lost in semantics. I'd be happy to rephrase the question as consistently profitable.

You're right ofc, i jumped the gun a little, my bad

 

[themickey]

My algos are based on this "better pattern" logic.

How did you arrive at making a better algo, was it from years of retail training in the school of hard knocks or do you happen to have trading pro background?

 

[tsfx]

How did you arrive at making a better algo, was it from years of retail training in the school of hard knocks or do you happen to have trading pro background?

school of hard knocks + a succesful trader kinda "guided" me in the right direction. No pro backround. Just lots of lost money and screen time and a realization what is possible and what is not.

 

[M.W.]

As you already included algorithms that assess probabilities of direction I would probably also add an additional class, those that attempt to predict magnitude, volatility algorithms.

Then there is another class, those algos that base logic on news sentiment.

In fact there are many more classes of algorithms, each one based on distinctly different features.

There are essentially only 4 types of algo. Everything else is largely a variant (and subvariant) of one of these classes. Now whether any of these algos watch Patterns, Trendlines, Channels, and/or Support and Resistance (or not) is debatable.

• Trend-following algorithms: These algorithms identify and exploit trends in the market by buying when prices are rising and selling when prices are falling.
• Mean-reversion algorithms: These algorithms seek to take advantage of the tendency of asset prices to revert to their long-term mean.
• Statistical arbitrage algorithms: These algorithms exploit short-term price discrepancies or mean-reverting behavior in assets.
• High-frequency trading (HFT) algorithms: HFT algorithms execute a large number of orders within extremely short time frames, often milliseconds or microseconds.

 

[murray t turtle]

What most people fail to realize is that in a market such as ES which is extremely competitive there are hundreds if not thousands of strategies (manual and algorithmic both) being employed simultaneously with trading decisions being made for a wide variety of reasons/triggers/signals. This, partially explains why liquid index futures move the way they do with plenty of 'noisy' movement.

For example, you'll have arbitrage players/algorithms working simultaneously with mean reversion algorithms or trend following algorithms where a mean reversion algorithm will sell a pop and a trending algorithm / BO algorithm may buy it. There's always a tug of war between buyers and sellers.

And this is with no mention of hedging activity (which is the reason these markets exist to begin with) where a hedger may dump a larger buy/sell order at market in any given moment at a price they deem favorable. This was also..........

All this makes for a complex market where only the few are making any money consistently.

If anyone can do it with KISS, be my guest. But I have not seen any evidence of anyone doing so.

%%
Good points, ES context, even if it was not clear, if you liked the idea of ''chase breakouts'' LOL??
Sometimes when all the technical+ fundy+ funds work is getting done, math can help.
ES front mo DEC= 16% sell/84% buy;
SPY, liquidity leader=8% sell/92% buy. For those that are not good with math, look up liquidity leader. Another clue, for the many that dont make money in Es, keep seeking.....
I used barchart.com for that, sure not limited to that website.

 

[M.W.]

A 100k investment turns to 3.83 million in 20 years. Someone serious would, however, invest at least 500k. 20% annual compound grows to almost 20 million dollars. That's getting rich in my book, sufficiently funded to live a life that affords most ordinary conveniences and many additional enjoyments.

Those investments are meant to be saved and reinvested, not consumed within the operating period.

I think this is a really bad question to understand if a strategy is profitable.

Let me give you an example:

Say i have 100k. That's about 20x the accont size an avg retailer has.

Now, say i make 20% return on that capital every year consistantly. Also, let's say my trailing drawdown is 1:5 doing it meaning that for a 20% return i don't have an open loss bigger than 4% at any given time from the highs of my PnL.

Now, that strategy would put me on top of every hedge fund out there. I'd literally be amongst the best in the world.

And now let's see if i'm getting rich with these numbers. 20% of 100k is 20k per year. That's about 1600 usd / month. Gross.

Would you consider that "rich"?

 

[zghorner]

If this is true, why does everyone argue about watching them (Patterns, TLs, Channels, S and R)?

Even if Price cuts through them like a hot knife through Butter at times, it still makes sense to be aware of them?

So IMO...the longer the trend, and the more times price has bounced from S&R, the more important it is. People have made good money simply buying support and selling resistance. As long as your size is appropriate and you accept they are zones and not specific prices then its possible to do well.

If for nothing else, they give you a directional bias...get positioned and if its wrong, get out or reverse.

I find long term S&R to be the most useful...tendlines are meh, patterns like cup & handle and head & shoulders to be even less reliable.

I've finally joined the PA bros and since have seen my trading do a 180...I look primarily for intraday liquidity and obviously try to be on the right side of it. My research gives me an idea on how for a move typically lasts in one single direction, look to exit as soon as price approaches that level and stops acting right.

also, something i call "volume popping" has been very useful as well comparing intraday volume to MA of volume...little things to give directional bias...enter in areas of indecision/exit/enter/exit/enter/exit until market direction is decided (my position is good and green) then sit tight until exit warning -> then triggers take place.

 

[Darc]

Since when is Al Brooks an authority

So IMO...the longer the trend, and the more times price has bounced from S&R, the more important it is. People have made good money simply buying support and selling resistance. As long as your size is appropriate and you accept they are zones and not specific prices then its possible to do well.

If for nothing else, they give you a directional bias...get positioned and if its wrong, get out or reverse.

I find long term S&R to be the most useful...tendlines are meh, patterns like cup & handle and head & shoulders to be even less reliable.

I've finally joined the PA bros and since have seen my trading do a 180...I look primarily for intraday liquidity and obviously try to be on the right side of it. My research gives me an idea on how for a move typically lasts in one single direction, look to exit as soon as price approaches that level and stops acting right.

also, something i call "volume popping" has been very useful as well comparing intraday volume to MA of volume...little things to give directional bias...enter in areas of indecision/exit/enter/exit/enter/exit until market direction is decided (my position is good and green) then sit tight until exit warning -> then triggers take place.

View attachment 326591

Agree, Support and Resistance, Volume and Al Brooks (even though Al doesn't use Volume) are under rated around here.



I don't know what goes through @Laissez Faire s head sometimes.

 

[schizo]

Pattern recognition algos? Don't really fit to any of those categories.

As you already included algorithms that assess probabilities of direction I would probably also add an additional class, those that attempt to predict magnitude, volatility algorithms.

Then there is another class, those algos that base logic on news sentiment.

In fact there are many more classes of algorithms, each one based on distinctly different features.

As I've said in my earlier post, they are all "largely a variant (and subvariant) of one of these (four) classes."

Pattern recognition algos? Depending on what pattern you're talking about, it can be neatly fit into either trend or mean-reversion algo (or a variant of those two).

Probabilities of direction? Magnitude, volatility? Ditto.