From Truth to Noise?

Tubbs

Well-known member
Messages
261
Likes
2
Recently I've been looking at different timeframe data for a number of markets. Do the traders on these boards think that the long term timeframes such as monthly data are more 'truthful' that is less noisey than 5 or 1 minute data.

This might sound a really stupid question - I just want to ask if you think it is what is the reasoning. Is the long term data more predictable. If so how would you prove this? All different timeframe data seems to display the same tendencies. This I think is quite interesting in itself. Should all timeframe data look the same? why?
 
Should all timeframe data look the same? why?

Hi Tubbs

I know what you mean - perhaps signals produced from TA on daily charts for example, are more sound from a TA basis than 1-minute charts. But why should this be so?.........
All technical analysis indicators etc. do the same calculations, regardless of the timeframe they are applied to. An indicator does not recognise one day as being any more valuable than one minute, because to the indicator they are both just one period?

I'm not sure if there is a definate answer yes or no.....
 
Last edited:
Every market has its own quirks, and its own "culture" if you like. You might well find some seasonal markets trade more predictably on longer time frames.

I think you will find that in general smaller time frames are more predictable. Consider a 5 minute chart where there is a hit on a daily pivot and a long shadow hammer forms there sitting on the pivot and then price starts moving up. I think a lot of traders will wager the price will continue to move up to the next level of resistence, maybe a few points up.

Now consider a weekly chart. The price does a similar things and hits a 52 week low, then moves sharply up. Now how would you like to wager it hits up at the next resistence level. Take into consideration that might take a week or so, and in that time, there might be some really bad market news.

Do you get my point?
 
jtrader said:
I know what you mean - perhaps signals produced from TA on daily charts for example, are more sound from a TA basis than 1-minute charts. But why should this be so?.........

1min chart trader is vulnerable to contra setups on 5min,15,60,240,Daily,Weekly
5min chart trader is vulnerable to contra setups on 15,60,240,D,W
15min chart trader is vulnerable to contra setups on 60,240,D,W
60min chart trader is vulnerable to contra setups on 240,D,W
240min chart trader is vulnerable to contra setups on D,W
Daily chart trader is only vulnerable to contra setups on Weekly
etc.

So there should be less noise on longer timeframes as the possible "strategy space" is reduced. Also, longer timeframes are more likely to reflect swings/changes in sentiment due to shifts in fundamentals. Longer timeframe traders also have deeper pockets and should therefore produce cleaner/stronger trends and patterns.
 
Last edited:
great replies everyone
I think the arguments are disagreeing somewhat. I agree with all the comments made though. Maybe one way would be to do a statistical analysis. If this period is an up period what is the percentage that the next period is up. If this is the same for all timeframes then there is no more predictability than any other. I believe, however, that is a linear analysis and maybe too simplistic for the market. worth a try though.
 
1min chart trader is vulnerable to contra setups on 5min,15,60,240,Daily,Weekly etc.

Hello C6,

Similarly, those trading longer timeframes will often have shorter timeframe players taking the other side. Even in the strongest of secular trends, it is very unusual - to say the least - for a market to move solely and continuously in one direction in all time frames. The further you zoom out, the more potential traders there will be to push back against your position. Long a weekly breakout and think of all the daily / hourly players who are selling to you. Short a one minute breakdown and the tick scalpers will be buying from you.

It all comes down to perceived value and this will obviously vary enormously between participants. A 10 second ripple or a monthly tide actually look much the same (fractal) if you remove the chart periodicity. This is probably becuase human psychology is consistent regardless of time frame. So I would say there is no intrinsic advantage to playing longer timeframes, especially if you subscribe to the view that a longer term trade demands a wider stop loss.
(For the rare few that can pinpoint an entry on a fast chart, employ a "scalper" size stop and then ride it for a huge payoff on a slower chart there probably is an advantage. But that lies more in their skill than any benefits offered or not by a certain periodicity).

You might claim that the opportunity cost of one's time is a reason not to trade short time frames, but you should be paid more by a shorter time frame as you make yourself available to ride a larger number of smaller waves - increasing size apppropriately - than your less active adversary. The only disadvantage here is commission costs, but if they are low they are easily absorbed.

Noise? No such thing imho. Every trade counts and matters to someone. In fact the devil is often in the detail that many write off as meaningless.

It is true that fundamentals may drive longer time frames ... in this way I suppose they are more truthful, more reflective of the true value of an instrument, but truth could be said to be insignificant. What matters is finding something cheap / expensive by your criteria and then buying / selling it for a profit. Does it matter if the price progression "smoothed" by time is more honestly reflective of the instrument's true value? I would say not. What is last traded price, after all, other than a momentary consensus of value between those participating (or choosing not to - because they see no under- or overvaluation) at that time? The perceived value of the fundamental buy and holder will differ from that of the scalper but each has a fair and equal opportunity to profit from their view, regardless of what the instrument is actually "worth". Compared to similar properties that have been bought or sold in the last few weeks, my flat may be said to be worth 100k but if someone's only prepared to pay 90k and I want to sell it immediately, then the intrinsic value is somewhat irrelevant.

But ....does a particular time frame (or zone, at least) hold temporary dominance and offer the best R:R at any given time? I would say yes it does. For instance the 15 min is strongly trending with clean waves so you trade that until it starts to congest; in the congestion you go for 2 or 5 minute and trade the range; but suddenly there is an FOMC decision so you switch down to 10 seconds for a few minutes until things settle. In each case you can vary your size to keep the capital at risk reasonably constant. Flexibility is helpful as it allows you to adapt to the size of the current waves and conditions ... until they change, which they invariably do to keep us on our toes. :)

All imho.

Interesting topic Tubbs, cheers for bringing it up ... and is your trading floor for locals only? :D
 
Last edited:
For products like the Bund there is definitely a volatility/activity smile throughout the day and if you're using indicators on a short horizon, say a 10 period CCI on 5 min charts then during the midday lull you're going to be getting signals all over the place. You could get arround this by using a longer indicator TF, say 100, 200, 300, etc...
I found that sub 10 min bars in the Bund to be quite 'noisy' and so for swing/longer-term trading I've moved to higher timeframes. For scalping I use a longterm (D,60m) so I can see the very big picture for levels and then short-term (10,5,1) for 30 second scalps.
Basically what I'm saying is that 'noise' really depends on your time horizon and your indicator look back period.
Be wary of having your bar ATRs less than the normal bid/offer spread plus a couple of ticks, eg bund might be 4-5 tick bars min whereas EuroFX might be min of 6-10 tick bars.
 
Frugi, minx good posts. I agree that the term noise is probably not fair. As you say, Frugi, each bit of information if really as valuable as the next. That really is my main problem with trend following. Basically you are looking at a trend already and saying that you are following that. Does a trend actually exist at all? It surely just a past performance and has no real bearing as to what may happen now and in the future. My original question postulated whether a certain timeframe would show to have more predictable trends than any other, i.e. is continuation of a trend more likely on a daily graph than a minute graph. Well I think some good answers have been put forward. I certainly have always favoured short term timeframes over longer ones because I never really saw any better attributes with the longer ones - just more time to be wrong. I think it is a difficult thing to resolve and one that certainly creates the market paradox of risk and reward. The only sensible thing I can say is that the higher the timeframe the more your overhead costs decrease. At the same time of course your stops have to increase and therefore your leverage has to decrease. So after all of that is there anything to be gained? It seems which ever way you look at the market you get the same answers, a bit like a picture where the eyes follow you around the room.
 
I've been trading off 2 minute charts. I do not accept the concept of "noise". If you follow price action there is usually a "reason" for each move. When you trade smaller timeframes, you would be an idiot to trade against the bigger timeframe trades. But in any case, the idea is that a chart reflects all market inputs, so that is taken into account if you can read the chart.
 
And as for my trading floor - it's definitely only for LOCALS!!! Do you know I hadn't even thought of that joke!
 
Tubbs said:
The only sensible thing I can say is that the higher the timeframe the more your overhead costs decrease. At the same time of course your stops have to increase and therefore your leverage has to decrease. So after all of that is there anything to be gained? It seems which ever way you look at the market you get the same answers, a bit like a picture where the eyes follow you around the room.

Not always, an indicator getting you in on a 5 min chart at 10 and out at 5, then back in at 10 and back out at 5 and again and again will bank you -20 ticks. On a longer timeframe, say 30 min, you might get an entry at 12 and stay in the trade with no choppy exits. In one you're -20, the other you could be b/e. Shorter TFs arent always better.
 
minx said:
Not always, an indicator getting you in on a 5 min chart at 10 and out at 5, then back in at 10 and back out at 5 and again and again will bank you -20 ticks. On a longer timeframe, say 30 min, you might get an entry at 12 and stay in the trade with no choppy exits. In one you're -20, the other you could be b/e. Shorter TFs arent always better.

One should not trade on 1 indicator alone. You should be looking for several pieces of evidence before you enter. You should definately consider larger time frames and not trade against them, instead one is looking to get tighter entries into the larger timeframe trades.

On smaller scales, you should look for something faster like candlesticks as most indicators lag will make them useless for the smaller bars.

I trade 2 minute DAX, and have little bother with such chops. A choppy market is never good to trade.
 
Top