Does trendline work?

[tommcginnis]

I doubt anyone can really code an objective trendline system with programming. Give the same chart to 10 "pro charters", very likely all of them would draw different trendlines on the same chart.

Again, a question: Just what do you think algorithmic programs are programmed to hunt? Thousands and millions of them out there, all tuned to hunt for conditions where some MA (no matter how computed) moves above-or-below some other MA (no matter how computed) to trigger an entry (whether long or short), and then another set of criteria -- perhaps entirely unrelated!! -- to trigger an exit.

Take Volentee's S&P Peaks graph:
What if you dropped that peak capture down to the troughs, and drew a duplicating line -- you'd have a channel, right? Draw a third line right in the middle. Does that look familiar? Would that middle line look like a simple linear regression of the mass of candles?

With that, you've just unlocked
1) a short cut to channel creation --
2) an immediately available data pool for hypothesis testing
3) a condition-set whereby you might trigger entry, warning, or exit conditions.

("Not Rocket Science, guys!" Shoosh, you! They might get it, if you keep your trap shut.)

Of course they work....on the left side of the chart.

Absolutely correct! You can even say that they work by construction. And as soon as they don't (according to precepts and conditions set within the algorithm), the algo is triggered to exit, or piece out, or reduce rate-of-entry, or whatever. Depending upon the construction.

 

[ET180]

one can trade with naked chart - it will work by itself...

When one is looking at a "naked chart", they are still interpreting it. They are still looking for some characteristic of the price action that tells them when a good reward : risk trade setup has arrived.

Trends are real.

I think that's the key to this whole discussion. If you believe that trends are real, then it would be legitimate to attempt to label them and determine when they are broken. If trends are not real, then it makes no sense to consider trendlines or really anything in technical analysis because even to believe / attempt to anticipate a reversal, that would require you to believe that price was trending in the first place.

I guess the follow-up question would be, why do markets trend? (For very short time frames, tick-level, it's hard to see a trend. But over the longer time frames, trends seem to emerge.)

 

[CharlesS]

so the answer is still NO - NOTHING WORKS... you have to be able to win with naked chart, there lies the fundamental EDGE that you read the market better than others

"WORKS" doesn't work

 

[GloriaBrown]

When one is looking at a "naked chart", they are still interpreting it. They are still looking for some characteristic of the price action that tells them when a good reward : risk trade setup has arrived.



I think that's the key to this whole discussion. If you believe that trends are real, then it would be legitimate to attempt to label them and determine when they are broken. If trends are not real, then it makes no sense to consider trendlines or really anything in technical analysis because even to believe / attempt to anticipate a reversal, that would require you to believe that price was trending in the first place.

I guess the follow-up question would be, why do markets trend? (For very short time frames, tick-level, it's hard to see a trend. But over the longer time frames, trends seem to emerge.)

Trend is different from trendline. Trendline is a straight line and look like a great marketing tool.

 

[quant1]

I don't trade with technical analysis nor do I know much of the lingo. But in thinking about what a "trend line" is I've arrived at the following observations:

1. A trend line is some function which takes in historical data and returns some positive value for each time in the period. An example is a simple moving average over a given period.

2. There are two obvious ways (to me) of how to trade an instrument given some trend line: either we can assume that the direction of the trend line indicates future behaviour (momentum) or we might trade mean reversion to the trend line.

3. In the momentum case, we might look at a derivative or finite difference in the trend line over some lagged period. We might trade such that we expect the future price to converge towards the extrapolation of the trend along this "tangent line".

4. In the mean reversion case, we expect the future price to converge towards towards the trend line itself and thus trade accorsingly.

5. In both cases it seems that given a working definition of trend line, we should be able to see if one of these hypotheses are true.

So in summary, if we can be specific about a definition for trend line and can show some strategy related to trading with or around that trend line is of positive expectation, then the answer to the question is yes. However, I'm not sure that these observations match the way trend lines are used in practice.

 

[easymon1]

Just draw some straight lines on chart and you will win money. Does that make any sense? Can you prove it works or it is a tale?

It's just a tale.

The best deal that I've heard of is to privatize your wins, and socialize your losses.

Not sure how that deal evolves for us regular folks, so . . . good luck with that.

 

[comagnum]

Here's a nice little trade from Friday with a reward-to-risk of nearly 10:1, at no time was it in the red. It really does not matter what entry signal you choose, the real trick is cutting losses and leaning into your winners - riding them as far as possible.

Here is a sample below from SBC Capital of how one of their 7 digit trader selects equity day trades, for those that think T.A. is silly & bash it. I would bet this trader would attribute most of his success to risk/trade mgmt and having the right mind set - not so much how he chooses his entries.

 

[GrowleyMonster]

Depends on where you draw your lines.

First of all, it is human nature to expect good things to happen and hope for everything to go according to plan. I think a lot of trend lines are drawn a bit overly optimistic.

Also, other people are drawing the same lines. Not neccessarily to use directly, but to predict what other traders, especially retail traders, will do. Those guys, thinking on the higher strategic level, will sell before you do. Or buy before you do. Because they know at least statistically where you and the rest of the thundering herd will act.

If you use trend lines you draw on your own charts, it could be a good play to not wait til the price actually touches the line, to act. The other guy wants to beat you, and a thousand others, to the punch. So you beat HIM to the punch, if you are smart. Particularly if you are playing with a large position relative to the float and volume.

Trend lines are also good for spotting and verifying a breakout. Same precaution applies, of course. If you see a nice volume spike, for instance, as the price penetrates the trend line, and the next probable level of resistance is quite a bit higher, don't wait too long to buy back in. Otherwise just let it go, sit back and count your money. Don't chase it.

This advice is given for entertainment value only. YMMV.

 

[GloriaBrown]

View attachment 207370

It's just a tale.

The best deal that I've heard of is to privatize your wins, and socialize your losses.

Not sure how that deal evolves for us regular folks, so . . . good luck with that.
View attachment 207371

Your trendlines graph look funny xd

 

[GloriaBrown]

That would be a good way about going about to see if they work or not. It's more difficult to code up automated trendlines. I have done it, but there's still subjectivity as to where to place the lines. Support and resistance levels (trendlines with 0 slope) are a bit more straight-forward and less subjective to implement. That would be the place to start.



You can kind of do this...Tim Knight over at TastyTrade has a show every week day where he goes over some of his most promising charts. If you subscribe to his charting service (think it's like $40 / month with a free trial), you can also see his charts and past charts. Unfortunately, he does not (publicly) keep track of how his predictions have worked out. Since he's mostly been bearish during the entire bull market run over the past 10 years, if he's still able to make money, then his approach must be working as he's going against the trend. Either that or he makes his money off of subscriptions.

Either way, if anyone here has found an edge with chart patterns, unless they are selling a book or service, why would the disclose their edge?

mostly been bearish during the entire bull market
Means his trendlines are BS already