Optimal or Useful?

[Neal 10]

What is the appropriate sized bar or candle to use? I want to specifically address the threads on ‘VSA’ and ‘Volume Based Candles’. Thanks to TinGull for his excellent facilitation and for all the fine contributions to these thoughtful threads.

 

On 3/17, page four of the ‘Volume’ thread, Soultrader post a comment that speaks to the issue I would enjoy discussing:

“I am still trying to find the ideal count (of volume based candles) for the YM chart…â€Â

 

There seems to be an ongoing question for every trader as to the optimal time (for minute charts) or level of volume (for volume charts) to use in order to analyze the charts.

 

My question is: should we seek for what is optimal or useful?

 

We have heard of the dangers of optimizing an indicator to perform well over a past period. It likely will not have a similar performance going forward. We have also heard that a trader should not seek to be right, but to make money. And of course we all know there is no such thing ‘out there’ as the Holy Grail. But still we seek and tweak.

 

But do we continue to waste our time with pursuits of perfection and optimization, wanting to be right, and wanting to find the Holy Grail’s distant cousin?

 

Would it be better to identify what is useful in order to create a structure from which to perceive the market?

 

For instance, if VSA is the basis of the way you wish to look at the market, what time chart should you use? You can’t use volume charts for VSA as they eliminate the volume variation you need in order to make your analysis. Do you use a 5-minute chart to eliminate noise, but to move quickly enough to trade intraday? But could one 15-minute bar give you a different perspective than three 5-minute bars? Two large up 5m bars on high volume, followed by a strong down 5m bar could be a very nice 15m Up-Thrust (ultra high volume, wide spread up bar closing in the middle).

 

Surely you could use multiple charts, but which would you defer to, or which would hold the greater weight? Some would offer that the higher time frame is more reliable, but for intraday trading, decision-making often has to be swift and decisive.

 

Would it be best to consider what is useful, rather than what is optimal?

 

Using VSA again, would it be best to find a reasonable time frame (say, 3-minutes) that you feel isn’t too noisy, but that will give you enough timely entries, and then create a reasonable trade management plan to follow?

We have to apply some type of structure to the market from which to analyze. We have to accept that there are days or times when a 2-minute chart would have identified better setups, and other days that a 5-minute chart would have provided the better setups.

 

We can chase ’optimal’ for a lifetime if we aren’t careful.

 

Does it make sense to:

1) identify what is useful,

2) accept that there are times when other charts will better identify the setups we use,

3) consistently apply our strategy each day, and

4) be content that over time it will yield as many positive and negative results as another time frame or volume level, but will provide us with the structure we need to be able to analyze the market in order to make money.

 

As always, our goal is to make money, not be right.

 

Am I right?

[dalby 10]

the question(s) you are asking is unanswerable.

 

you have to adopt a methodology that suits your trading personality, and stick to it. there is no "right" timeframe. I personally look at a 5 minute, a 15 minute, and a 233 tick charts along with Market Profile (I use volume based value areas and POC not time based btw)...

 

but that's me.

 

charts are merely models of what the market is doing. there is no right and wrong timeframes, tickframes, volumeframes, etc. you don't even need to use a chart at all if you don't want. i've seen a trader trade merely using tick, tape, and pit noise. all models necessarily filter out information. you have to decide what is and isn't important in your methodology.

 

simply put... it depends.

 

there is no perfection in trading. even the best trader is frequently WRONG. that's cause market analysis is probabilistic. there is no determinism here. it's a chaotic dynamic feedback system. it's not a swiss watch.

 

the market necessarily adapts to what traders are doing, since it IS the sum total of all trades. what may work best NOW may not work as well 6 months from now, and almost certainly will not be as useful 10 years from now. there are many consistencies over time, but also many changes.

 

i'm not trying to get all mystical

 

it's far more important to have a solid business plan, solid risk management, solid discipline and a workable system with an edge, than to seek out optimum candles, etc.

 

the mere fact that one is trying to do the latter distracts from the real aspects of trading which are managing risk while applying one's edge.

[walterw 10]

Neal , I believe it must be usefull and adjusted to your trading style plus eventual updates as market conditions migrate.... cheers Walter.

[Soultrader 12]

True enough... perhaps I should of cleared it up. I was looking for an optimal volume chart timeframe suitable for my risk tolerance. With candles, this is more of a confirmation > enter type strategy. Therefore any entry requires a slightly wider stop based on the previous bar high, HOD, LOD, etc... With a wider timeframe the confirmation comes later. Hence a wide stop. With a shorter timeframe the stops can be slightly less. My risk tolerance is a maximum of 20 YM pts per trade. Therefore finding the right interval with volume charts is crucial for me to decide whether I will take the trade or pass on it.

 

So basically, I was just testing timeframes that give setups with no more than 20pt risk and that signals valid candle patterns. By valid, I mean with a volume chart I need to see a clear fight between the bull and the bear.

 

With business plan, methodology, money management, etc.... nothing is being changed here with my trading. Im taking my core trading style/methodology and adding a little twist to it by incorporating volume to a greater degree.

[momentom 10]

Soul, you have hit the nail right on the head.

 

The timeframe you use MUST, in my unhumble opinion, relate to the stop placement strategy you use. For example, many of us use the previous high swing on a short. If that is too far away, we must pass on the trade so we need to choose a number that gives us logical exits that are comfortable to trade.

 

One approach is to look at the "average" volume of the market you trade. Just an eyeball calculation. Then, divide the volume into the approximate number of bars you want to see in a day.

 

I am a scalper and mostly trade the DAX and Russel. Both now trade about 250,000 cars a day. 0.25% of 250,000 is 625 and gives me over 400 bars a day if I use a Volume Bar of between 315 and 625 (triand error eyeballing the charts for my setups) and get about 20 setups a day and a stop loss that fits my trading plan. If you do not want so many trades you can either have tighter criteria to maintain tight stops and a larger % of winners or you can slow things down by looking at 0.5% of 250,000. If you are a swing trader look at 1%.

 

Hope this helps. Great threads on VSA and Volume candles.

[TheBramble 10]

We have heard of the dangers of optimizing an indicator to perform well over a past period.

Yes often, and I did subscribe to that piece of advice myself for a long time. But I am now convinced it simply isn't useful.

 

Even if you just allow for being in a Bull or Bear phase for whatever instrument you're trading and adjust your upper and lower levels for indicating overbought/oversold this will find this yields a higher number of successful trade entries. And fewer trade entries too.

 

You can then take this one step further and positively optimise for the recent history of the instrument under analysis. All instruments go through cycles of expansion and contraction and the periodicity of those cycles WILL change over time. You need (IMHO) to let the instrument tell you what beat it's marching too rather than take a position that all instruments will bend to the might of your POWER setups! :)

 

Having said that, I use few indicators and the work involved in tailoring even those few to my traded instruments requires significantly more time and more effort. And it makes it doubly intensive when you have a scanner running on indicator-based formula and parameters too. But, in my view, the payoff is well worth the additional energy and time.

[brownsfan019 17]

Neal - in regards to VBC's, I can tell you that the setting depends on what YOU want the chart to do.

 

In other words, if you want a chart that prints candles 'quicker' than you should start at 1000 or less (on the YM). If you want a 'longer' term chart, go above 1000. The exact number is not terribly important in VBC's.

 

To see what works for you, simply open 3 YM charts and start one at 500 VBC's, 2500 VBC's and 5000 VBC's. From there, you'll 'see' what you like and then you can finetune it even more. If you like the 2500 chart, then run a 1000, 1750, and 2500 chart. And from there, you can narrow it down even more.

 

Just keep in mind that whatever setting you choose, you need to use that setting each and every day. It's very easy to see that today a 1500 VBC chart was best and on Tue it was 500 and last week was 1000. You are never going to know exactly what setting is best for the day. The other option is to have 2-3 VBC charts open at the same time.

[Anonymous 11]

From a VSA perspective, Tom Williams would recommend nothing lower than a 3 min chart.

 

Remember, VSA does not use volume based candles or bars. For VSA to be used one needs to be able to COMPARE volume relatively as well as actually. So one must look at the volume histogram for bar x and make comparisons to the volume histogram of bar x-2 for example.

 

Todd Krueger recommends using multiple time frames.

 

More generally, many people have already said it: chart interval is in some ways methodology dependant.

 

You would not use a 250 tick chart to day trade. Nor would you scalp off a weekly chart.

 

Yes, the above last two statements are pretty obvious, but they do hint at a the larger truth.

 

Currently, I am learning to trade via a certain strategy that "requires" looking a multiple time charts and multiple instruments. I use a 5,10,15 min Euro AND a 30 min Euro AND a 5 and 10 min Swiss Franc. I am transitioning, however, to the YM. Once there I will use a 3,5,10,15 min YM AND a 3 min ES AND a 3 min DIA. These last two are for "sister trade set-ups".

 

This is because I am looking for certain specific patterns that while not showing up on one time frame may be present in the ones higher or lower.

 

Now If you were just looking to take signals off of a Pivot line , then there would really be no need to look at 4 or 5 chart intervals. If you are simply going long on a 21 bar breakout, the issue of "noise" comes in to play. Larger charts would tend to not give as many false signals.

 

How many trades do you want to take a day? One does not want to over trade, but at the same time picking a larger timeframe may mean long periods of non-trading. Is this what YOU want?

 

Simply, the optimal timeframe is method dependant, personality dependant, and market dependant.