the hare
Senior member
- Messages
- 2,944
- Likes
- 1,282
Perhaps you could think about this slightly differently. If you really must use indicators to define the direction of the trend on the 4 hour chart (which I wouldnt reccommend), then as a starting point, DONT apply the indicator to the 4 hour chart, apply it to the daily, or weekly. Practically any oscillator will work, and you can you can use reasonably fast settings which cuts down the lag. Those little waves on the daily and weekly charts that a fast oscillator picks out so well are the trends in the 4 hour t/f etc. This sounds completely counterintuitive, but its an approach thats worked well for me in the past.
You probably need to use a multi timeframe approach, maybe a fast oscillator in your slowest t/f combined with maybe a simple dow theory approach applied to waves in a faster timeframe.
You have to remember that you are not trying to catch the turning points in the 4 hour t/f, you are just trying to take multiple bites out of the 4hr trend. The indicator is going to be wrong at the start and end of the trend, but 80% of the time its going to be OK (which certainly doesnt mean 80% of trades will be winners).
Maybe add in some oscillator divergence and longer term S&R to give a heads up to potential changes in trend direction, and its probably as close as you'll get from an indicator based approach.
I can reccommend bbmacs thread here at the zoo, IIRC there's some good material in there that maybe makes things a little clearer
You probably need to use a multi timeframe approach, maybe a fast oscillator in your slowest t/f combined with maybe a simple dow theory approach applied to waves in a faster timeframe.
You have to remember that you are not trying to catch the turning points in the 4 hour t/f, you are just trying to take multiple bites out of the 4hr trend. The indicator is going to be wrong at the start and end of the trend, but 80% of the time its going to be OK (which certainly doesnt mean 80% of trades will be winners).
Maybe add in some oscillator divergence and longer term S&R to give a heads up to potential changes in trend direction, and its probably as close as you'll get from an indicator based approach.
I can reccommend bbmacs thread here at the zoo, IIRC there's some good material in there that maybe makes things a little clearer
bbmac
Veteren member
- Messages
- 3,584
- Likes
- 789
Hare - Thanks for the heads up Lol !
-----------------------------------------------------------
Trend can be seen using overall price action - peak/valley analysis ie a classic downtrend is a succession of factal LL's and LH's and a classic uptrend is a successon of fractal HL's and HH's on any given t/f.
Sometimes a trend does not move classically but the overall 'direction' of travel on
that t/f is still apparent, - this is what I call a general trend and I use the 100 and 200 sma's to determine what the general trend or conditions/sentiment is if price is not trending classically or indeed in a range.
Re a General trend: 100sma under 200sma and price under the 200sma/preferably the 100sma - a general downtrend. 100sma above 200sma and price above the 200sma/preferably the 100sma - a general uptrend is present.
Look at 1hr/4hr gbpusd right now - price in a general downtrend on 1hr, whereas on 4hr in a classic downtrend.
This is just one methodology to determine trend - there are many others but something PA based is always preferable.
G/L
-----------------------------------------------------------
Trend can be seen using overall price action - peak/valley analysis ie a classic downtrend is a succession of factal LL's and LH's and a classic uptrend is a successon of fractal HL's and HH's on any given t/f.
Sometimes a trend does not move classically but the overall 'direction' of travel on
that t/f is still apparent, - this is what I call a general trend and I use the 100 and 200 sma's to determine what the general trend or conditions/sentiment is if price is not trending classically or indeed in a range.
Re a General trend: 100sma under 200sma and price under the 200sma/preferably the 100sma - a general downtrend. 100sma above 200sma and price above the 200sma/preferably the 100sma - a general uptrend is present.
Look at 1hr/4hr gbpusd right now - price in a general downtrend on 1hr, whereas on 4hr in a classic downtrend.
This is just one methodology to determine trend - there are many others but something PA based is always preferable.
G/L
the hare
Senior member
- Messages
- 2,944
- Likes
- 1,282
Ive always found it remarkably difficult to determine trend using simple indicators, which perhaps isnt that suprising really. Ive even heard it argued that trends do not actually exist ! although I'd argue that you can prove they do with some fairly basic statistical analysis.
I remember years ago taking a bunch of completely random trades, and of course if you analyse the data retrospectively, you find that if the markets trending up, most of your buy trades tend to do well, and your sell trades dont do so well, and vice versa. On that basis you conclude that if you could just measure "the trend" then you can make money tossing coins (and of course you can).
I remember duckfu once saying that by definition, we are always trying to exploit some trend somwhere, if we buy, we want it to go up, and if we sell we want it to go down, and although he's right, and this seams like a sensible approach, and an approach that certainly works, I'm not 100% sure that its the optimum way of going about things !
I've spent most of this year working on strategies that dont incorporate any analysis of trend at all, and it certainly seamed to be a much easier thing to do than developing a trend based method.
Just out of interest, do you tend to get a 50/50 distribution of long and short trades over a year, with similar returns from each, or do you get asymetrical results with one side taking more trades or doing far better ?
bbmac
Veteren member
- Messages
- 3,584
- Likes
- 789
Interesting stuff. Duckfu - wonder what happened to him - last I heard he was running some site called 'stocktwits' like a twitter for traders - dunno what happened with it though ?
I would class myself more as an intrady swing trader - ie looking for potential swing points at pre-identified potential supp/res/sbr/rbs, but like the methodology involved for any part of trading I think you need to have a methodology to ascertain what a market is doing, particularly on the t/f's of interest, and this methodology is incorporated into my trading edge. Ie generally speaking - a strongly trending market is more likely to blow thru potential supp/res than bounce meaningfully at it, and in a ranging market vice versa..
I mentioned cable in post above and that is interesting today beacuse although 4hr and daily are in classic downtrends and the 1hr is in a general downtrend, the intraday sentiment sinbce the 5690 Lo early in the day has been biased to the upside, so here we have another consideration. - This intraday bias has played out with a HH now on 1hr and the upside bias more visible on the t/f's below it.
I tend to try and get with the trend where possible via repeating set-ups to do so, so via my 3 t/f analysis. Ie if the middle (intermediate) t/f is trending and this is in the same direction as the longest - (trend) t/f and particularly if these overall price action conditions are replicated on the t/f (s) above that, then my bias will be 'with trend.' This does not mean to say that there are not hi-probability opportunities to go counter trend at potential supp/res but just that the trend direction remains highest in probability given/whilst those opa/general circumstances/conditions persist.
The distribution of long/short trades tends to come in clusters depending on the market conditions on the t/f's of interest but over 1yr it is probably pretty much even although I havn't looked particularly closely at this stat to verify this feeling.
G/L
I would class myself more as an intrady swing trader - ie looking for potential swing points at pre-identified potential supp/res/sbr/rbs, but like the methodology involved for any part of trading I think you need to have a methodology to ascertain what a market is doing, particularly on the t/f's of interest, and this methodology is incorporated into my trading edge. Ie generally speaking - a strongly trending market is more likely to blow thru potential supp/res than bounce meaningfully at it, and in a ranging market vice versa..
I mentioned cable in post above and that is interesting today beacuse although 4hr and daily are in classic downtrends and the 1hr is in a general downtrend, the intraday sentiment sinbce the 5690 Lo early in the day has been biased to the upside, so here we have another consideration. - This intraday bias has played out with a HH now on 1hr and the upside bias more visible on the t/f's below it.
I tend to try and get with the trend where possible via repeating set-ups to do so, so via my 3 t/f analysis. Ie if the middle (intermediate) t/f is trending and this is in the same direction as the longest - (trend) t/f and particularly if these overall price action conditions are replicated on the t/f (s) above that, then my bias will be 'with trend.' This does not mean to say that there are not hi-probability opportunities to go counter trend at potential supp/res but just that the trend direction remains highest in probability given/whilst those opa/general circumstances/conditions persist.
The distribution of long/short trades tends to come in clusters depending on the market conditions on the t/f's of interest but over 1yr it is probably pretty much even although I havn't looked particularly closely at this stat to verify this feeling.
G/L
Just use (close of current bar - close of bar x bars ago)!
It's misleading to think that you can do any better when looking at the chart of an underlying - proper trend analysis requires converting the underlying process into % or log returns and then hoping to spot a drift term in those returns.
I guess the hare already answered. There are some details i dont agree with him, but he know what hes talking about.
1) I do not recomend oscillators, personally i dont like it.
2) A trend at the 4 hr chart will ajdust to short periods of time.
I think you can use a multi time frame scheme, if you do it just take a look at the daily chart to see the real trend.
I follow the 4 hr a nd 1 hr charts with a regression line at a range from 60 to 100 periods depending on volatility. I use the daily chart to see if the short time trend is alligned with the daily chart.
1) I do not recomend oscillators, personally i dont like it.
2) A trend at the 4 hr chart will ajdust to short periods of time.
I think you can use a multi time frame scheme, if you do it just take a look at the daily chart to see the real trend.
I follow the 4 hr a nd 1 hr charts with a regression line at a range from 60 to 100 periods depending on volatility. I use the daily chart to see if the short time trend is alligned with the daily chart.
nunrgguy
Established member
- Messages
- 682
- Likes
- 126
Here, hope this helps. Whether you use a MA, an Oscillator or weekly opens, it all amounts to the same thing if you use the same logic to set them up (opens and MA only shown):
