Support & Resistance? Why do some work and some not?

[Laissez Faire]

Exactly,

I received in past a lot of charts (metals LME) from Glencore and Natexis. So many lines that they were always right. The problem was that it was only clear in hindsight which line was the correct one.
I was shocked that these "professionals" made that kind of charts. Threw them instantly in the garbage.

There are however some lines that seem to hold some significance and either slow down price or reverses off it.

Prior Day High/Low/Close/50%. Possibly the pivot point.

You could also plot the daily Open and the rolling 50% level of the day. It could help as some kind of guidance in combination with reading price.

The prior day Close (16:00) is actually quite frequently traded and often a reversal. In fact, price reversed of that level to the tick (ES) yesterday.

Not sure if this is what people consider S/R though.

 

[FreakofNature]

There are however some lines that seem to hold some significance and either slow down price or reverses off it.

Prior Day High/Low/Close/50%. Possibly the pivot point.

You could also plot the daily Open and the rolling 50% level of the day. It could help as some kind of guidance in combination with reading price.

The prior day Close (16:00) is actually quite frequently traded and often a reversal. In fact, price reversed of that level to the tick (ES) yesterday.

Not sure if this is what people consider S/R though.

That's a lot of lines already though

 

[Laissez Faire]

That's a lot of lines already though

Sure is.

So, random after all...?

However, most days aren't outside days, so you're left with either the high/low of the prior day. The Close should always be watched though. It is statistically significant.

 

[Dolemite]

The above picture is a great example despite the fact that it is a chart created by tossing a coin you have to be able to read the chart and I think that the resistance level that is drawn on the chart is wrong at least for this chart.

Please tell me you are kidding. Are you saying that based on the lines you have drawn you can predict the next coin flip??

 

[Buy1Sell2]

Support and resistance lines seldom work because folks dont know how to analyze with indicators. These lines fall into the category of scribbles.

 

[Saber]

Please tell me you are kidding. Are you saying that based on the lines you have drawn you can predict the next coin flip??

Please read my post!

Trading is not about predicting but about the probabilities.

 

[qxr1011]

Please read my post!

Trading is not about predicting but about the probabilities.

I would say its about probabilities of making correct predictions.

 

[qxr1011]

Are you saying that based on the lines you have drawn you can predict the next coin flip??

yes,

because it is a coin flip for many, but not for some

 

[wrbtrader]

Support and Resistance Analysis is an analysis...not a trade signal.

Analysis + Trade Signal ----> Trade

I remember someone recently here at the forum asked someone else what's their trade signal that tells them when to enter the trade ?

The guy reply back via saying "trend analysis".

Trend Analysis is an analysis...not a trade signal.

In addition, there are dozens and dozens of different ways to compute s/r levels. I've seen at least 20 different ways here at ET alone that has been shared. There was one thread were someone ask members to post their charts with s/r levels...

Most charts of the same trading instruments had different s/r levels. This is why so many say s/r analysis is subjective because of so many different ways to compute or estimate the levels. Some even conclude its a coin flip and rightfully so until the below occurs.

Therefore, I think its best to just concentrate on the type of trade signals being used with s/r analysis because you can backtest the performance of the trade signal with different types of s/r analysis...not so easy to do with s/r levels considering they are just a type of analysis.

Simply, there should be a discussion about how the s/r analysis was being computed and then a discussion about the specific trade signal being used with the analysis prior to being able to find answers.

P.S. For anyone that didn't notice, there's a few different ways of computing s/r levels mentioned in this thread that would produce different levels (lines) in comparison to someone else using other types of ways to compute s/r levels.

 

[iamnewuser911]

They are 17/18, software I use to show others charts of signals or examples, I can't thicken the lines, so I use 17/18, bet you thought it was something useful, grin. I use multiple timeframes of whatever works well. I trade off 60 minute but risk is $500 but going for $1,000 plus, whereas trading off 2 minute charts it is much smaller risk and targets. I use 18 as when backtesting it works, and any from 14 to 22 sma's work, but 18 gave me smallest draw down of losing trades by 1% over others, I use 18 from monthlies to 30 second charts, I recall I started in 1978 with 21, but when home P/C came out, being tad faster works out better.

I use various ways to define trend, slope, close beyond a pivot, thrust bar completely above the 18sma. I don't like taking an indicator below or above the chart to define trend cause many indicators are either late or too early to the party. I have used other periods of sma depending on strength of some strong trends to have ability to get into them with a tight stop. It is those trends where it ripping at 60 degree or higher.

Crazy backtesting stats you have there.

MA seems to be a large part of your decision making. Have you tried removing it when backtesting? How much did the results diff with and without?

One more question: Trading off 60 mins and looking for entries on lower time frame like 5/1m. Should lower time frames be on the correct side of MA as well? Since I often identify locations on hourly (now added this MA on hourlyfor decision/filtering).