Basic question on continous contracts

[OpenMind]

Am I correct in thinking that the only use of continuous contract chart is to see the long term trend?

When I try to figure out the support resistance areas that might trigger a trade, the price levels in the near month chart will be different from the continous contract chart. As a result the patterns will also be different. What looks like a clear triangle in the near month chart may look completely different in the continous contract chart.

Does one therefore stick to the near month chart when one is looking for key price levels or patterns to order a trade?

Thanks.

 

[frugi]

A continuous contract is a way of splicing together the front month and a number of back months (of futures) so as to gain a longer term view of an instrument. e.g. Splicing together ESH, ESM and ESU to get nine months of ES data.

There is a problem with doing this in that when the back month expires the front month is likely to be trading at a substantial premium. So there will be a gap in the chart. e.g. YMH rolls over into YMM on the 9th Mar. YMH closed on the 8th at around 11000 while YMM opens on the 9th at around 11080. The premium over cash is much higher because the new contract has 3 months to expiry. So you have to back-adjust the continuous chart, eliminating rollover gaps working backwards from the right hand side of the chart. This should preserve the TA although you should remember after the first adjustment the back month(s) prices will have been moved to comply with the front month, so they won't be their original true value. In this example you would move the whole of YMH up by c80 points to fill the gap.

I'll try and put a chart up to show what I mean as I have explained this really badly.

Too tired ... links instead.

http://www.commodities-trading-systems.com/continuous-contracts-and-rolling.html
http://www.premiumdata.net/support/futurescontinuous.php

To answer your question, I THINK continuous charts are fine for TA and backtesting as the price progression is what is important, not the absolute values, but I may be wrong. Something doesn't feel right. Sandpiper? Helllllp.

 

[OpenMind]

Thanks Frugi. Metastock seems to calculate continuous contract data differently. It apparently takes a weighted average of the next few months and comes up with a price. As a result TA is not preserved.

As an example, take Cotton #2. The May contract chart shows a triangle, which is much less pronounced in the continuous contract chart.

 

[sandpiper]

Frugi,

Can't see anything wrong with your excellent explanation.

Cotton No.2 isn't my gig, however, It looks like your continuous contract is set up to incorporate active months rather than all contract months, i.e. March, May, July, October and December... However, it doesn't look as if the previous data has been adjusted at all on roll-over (as per Frugi's post), i.e. the continuous contract shows that gap up on roll-over of the active December contract that wouldn't be there on a back-adjusted continuous representation. That would seem to me to be the right way to do it if you want to use the chart to gauge support/resistance levels. After all, if the continuous contract was back adjusted, you couldn't say with certainty that supposed S/R actually represented prices that had actually traded.

Obviously, this representation of continuous contracts will corrupt any patterns such as your triangle.

So, in answer to your original question, you could:

Stick to the front month or a non back-adjusted continuous representation when looking for price levels. Use a back-adjusted continuous representation when looking for 'patterns' or back-testing. You could of course use other data for your longer term pattern recognition/back-testing, such as the cotton index.

 

[Silent.Trader]

Am I correct in thinking that the only use of continuous contract chart is to see the long term trend?

I'm not an expert on the matter, but imho continous contracts are mainly used for backtesting. As other contributors already remarked, continuous contracts have their problems. Personally I use the underlying index for analyses. This way most issue's with continuous contracts can be waived. But it's important to keep in mind that future and index are not identical and may behave (slightly) different, obviously the same is valid for the continuous contract and the real contract. This strategy off course won't work for many commodity futures.

Depending on the timeframe you're after you may also choose to take a look at the futures that have existed longest. In many case this will provide you with a history of about 9 months or more. Objection about this strategy is of course that at the begin of the futures life span it tends to be illiquid.

 

[RichardDale]

Another discussion regarding continuous contracts can be found at:
http://www.premiumdata.net/support/futurescontinuous.php

 

[lucky_irish_man]

A piece on continuous and back-adjusted continuous contracts:

http://www.aarontrade.com/html/futures_trading_prices.html