Traders' Roundtable

Did Propane get split? A few days ago, it was trading at 160, and now it's at 80. In checking NYMEX, it looks like it's always been around 80-90. Did CSI re-adjust?

The attached chart came from CSI, it doesn't show 160.

Strange. This came from CSI also, and it does show 160.

corn v slug

Look at the table entry for 01 June. A rollover from the June contract to the December contract (see "Delivery" column), followed by three days of zero volume and semi-fictitious quotes (Open = High = Low = Close). The spread between December propane and June propane on rollover day (01 June) was huge, and this caused the continuous back-adjuster algorithm to alter past prices rather dramatically. One possible interpretation is that it may be unwise to trade the December contract of propane during early June, because nobody is around to take the other side of your trades.

Although I don't personally know anybody who trades actual propane futures contracts, I do know lots of people who watch the propane chart and who run various technical studies on continuous contracts of propane as part of market-sector analysis. All of these researchers roll over their propane contract at least 6 times per year; in fact the most popular choice is to roll it over 12 times per year. It appears that you've got CSI set up to roll propane only 2 times per year, which may not be the best choice.

Huum. I have the same roll scenario as I do with the other commodities.

This isn't a roll over issue and unless they've changed the contract specs for Propane its not likely to be a "split." The price is in dollars and cents per gallon but it looks like CSI quotes cents. A halving of the price of propane would be surprising.

To resolve the problem force CSI to reload the contract data (delete the contract and reinitialize it would do it) and Corn's data will probably look the same as Sluggos.

I did. Same problem. Sluggo, are you manually rolling over every month?

Corn,

Check the raw data on those two contracts ... what does it do around the roll over time ... perhaps u could share that with us.

Also, u might email Rudy at CSI ... with that file information from here. He used to be quite responsive to issues like this.

I see it all now. My mind has opened up and I have glimpsed the truth. Thanks guys. I still wonder how Sluggo got the correct data for his chart, even though we both have back-adjusted data?

I'll phone CSI anyway on Monday to see how to correct this little anomaly.

Thanks again.

These two messages from the Customer area, might be helpful:
viewtopic.php?p=9891&highlight=rol%2A#9891
viewtopic.php?p=8498&highlight=rol%2A#8498

It ended up being the C++ back-adjuster. CSI recommened I use the older Fortran back-adjuster, and it works. The fact that July has 0 volume messes up the C++ adjuster.

Weird.