Sorry for my non standard nomenclature - by C1 contract I mean the nearest/most liquid contract i.e. the one you trade. That would be the march contract now and has a volume in the 300s: https://www.barchart.com/futures/quotes/LUH25/price-history/historical From the link above, for the last few days (except the 6th Feb) I have the exact same volume data from CSI for this contract :
Thanks, looks good. Maybe I found the wrong symbol on IB, I was looking at ATW that also says it's Rotterdam Coal but had much lower volume for March. Any idea which symbol on IB matches these volumes?
IB means IBKR, the broker Interactive brokers. Our CSI data matches, it's the broker that's showing low volume. What symbol at your broker do you link for execution?
This makes sense - just because you have 10x the equities relative to bonds doesn't mean you should have 10x the risk in that group. But.... while multiple semi-correlated positions at smaller size is definitely more optimal than a larger positions in a smaller set of proxy assets, there could be an issue where in the extreme case all these multiple positions get rounded down to zero as you simply cant make the minimum lot size. Then you have a massive gap in your portfolio. Rob's dynamic optimisation is one way of finding the best proxy portfolio. This is a tricky one.... In a different life time I did attempt to solve this problem use mixed integer quadratic programming but that caused a whole lot of other issues!
Sorry - yes IBKR shows me lower volumes as well. I dont have a proper data subscription on IB though, but by the looks of it the avg volume for the last year as it looks on the charts is about 50 units only.
For those that aren't aware here is a list of all the instruments I collect data for: https://github.com/robcarver17/reports/blob/master/Instrument_list I don't trade them all. Rob
Thanks this is where I got a lot of my ideas on things to add, sorting by your liquidity check page numbers. Also copied your liquidity check By coincidence equities are the group that have a lot of micros or mini contracts. In the other instance where I can only take one contract, I just take the first one whichever signals first basically. However since I have 3 speeds/strategies, I could still end up with short term in one, medium term in another, and long term in a 3rd correlated instrument so there's still a benefit to mixing them (for me). But like you said, there are much smarter ways to do it by adding correlation to the signal in some way, or dynamic optimisation to select. Just looked at my notes - I did try to trade it at one stage. But there was one side of the bid/ask totally empty even though some volume had been reported (I did buy data that month). That worried me enough to just totally remove it, since I use live stop losses held at the broker. I'm still stumped why IBKR volumes don't match barchart and CSI though. CSI does split up the monthly and quarterly contracts, however BOTH of those have higher volume than IBs "ATW" Rotterdam Coal.
The reason for your ICE Rotterdam Coal volume discrepancies will be that IBKR are only displaying outright trading volumes, whereas CSI and Barchart are including volumes generated via calendar spreads. Spread trading dominates Rotterdam Coal volumes. You can see this clearly by looking at the Barchart trading volumes which most often show identical volumes for different contract months indicating that the only trades occurring were spreads
Thanks! I had no idea IB doesn't show spread trades in volume figures. Makes sense now. Still not super confident in trading Coal based on my last experience seeing nothing on one side of the bid/ask at some times of day. If a stop executed a market order into that, the slippage could be extreme.. Risk$ of $2.33m looks OK on Rob's risk report but the 4 week average volume (78) is below 100 https://github.com/robcarver17/reports/blob/master/Liquidity_report @globalarbtrader do you get your volume from the IBKR API? or EOD data like CSI? If from IBKRs API I assume it's also missing spread volume like TWS is?