[ok] cargo supply NOT growing as it should

[groundbum2]

this is GW3, November 2005 so fairly mature with a LOT of cargo flights that have been flying for a year or more. I understand CBD and the ?e=1 to look at overlap between airports etc.

I've tracked 5 airports around LAX, and 7 around NYC. LAX-NYC city-pair has about 890,000 kilograms of potential cargo according to the route planner, but there is only 250,000Kg available to fly, ie less than 1/3rd of potential. Where is the other 66%? Between us we are supplying 530,000kg, so twice the current supply, so we should be pulling up the numbers to closer to potential.

If this was an isolated city-pair, say KMEM-KSLC then providing double the supply would have pulled up the amount available. Yet in a complex CBD environment, the algorithm doesn't seem smart enough to increase cargo despite our flights being there.

Spreadsheet available at

https://1drv.ms/x/s!AmanmSq8XxG0jhr-9x5zbNH3wg7v cargo splits folder

Simon

[Sami]

There was a previous thread/report where this was extensively investigated, and no issues were found. It's a tad complicated to look/check from player's perspective as all airports do not catch the same demand and areas, so it can't be directly summed and compared like you've done there.

I've saved the GW#3 data snapshot for later analysis and will take a look on the details later on.

[JumboShrimp]

NYC-LAX is the hardest to figure out (debug).  There are 40 combinations just from the airports you listed.  And there are more on the LAX side that you did not list: OXR, SBA,

[wilian.souza2]

So it seems that the real problem lies on the display of potential cargo figures - they should reflect the total amount of cargo that would be available in that airport AFTER discounting the marginal cargo that will remain in all the other airports of both ends's regions, as well as the effect of flights on competing route pairs. This is not what happens today, so we are left shooting in the dark in these complicated areas, guided by information which turns out to be false.

[JumboShrimp]

So it seems that the real problem lies on the display of potential cargo figures - they should reflect the total amount of cargo that would be available in that airport AFTER discounting the marginal cargo that will remain in all the other airports of both ends's regions, as well as the effect of flights on competing route pairs. This is not what happens today, so we are left shooting in the dark in these complicated areas, guided by information which turns out to be false.

I don't think the information is wrong or false.  It just does not say the whole story (as one single number really can't).  There should perhaps be some additional info.  I came up with a couple of ideas that would help the player here:

https://www.airwaysim.com/forum/index.php/topic,74722.msg437937.html#msg437937

[groundbum2]

NYC-LAX is the hardest to figure out (debug).  There are 40 combinations just from the airports you listed.  And there are more on the LAX side that you did not list: OXR, SBA,

just checked, no flights to NYC from those two. Well 3 Airbus's from KSBA to NYC. 

LAX-NYC is perfect for debugging, if the algorithm works for this pair then all the easier pairs should work as well! there's still 66% of cargo AWOL according to the CBD spec..

Simon

[JumboShrimp]

just checked, no flights to NYC from those two. Well 3 Airbus's from KSBA to NYC. 

LAX-NYC is perfect for debugging, if the algorithm works for this pair then all the easier pairs should work as well! there's still 66% of cargo AWOL according to the CBD spec..

Simon

That's a good point.  I will take a look at your spreadsheet again when I have some time.

[schro]

I've done some debugging on a multiple airport metro (LAX) to a single airport metro (OKC) in GW1 and was successful in finding all potential demand appearing as actual somewhere in the combination of route pairs (or at least close enough I didn't bother to open a bug report).

However, I have not taken on a multiple airport metro to another multiple airport metro like LAX-NYC or NYC-LON, however, in my general observations, something feels tiggered with it.. so hopefully we'll get some good investigation on it...  If I remember right, I've got 5k of demand from LHR to various large Northeast US cities that's not worth flying at all even when there's potential demand at 900k+.

[JumboShrimp]

BTW, I have seen an issue of Heavy Cargo "hiding".  For example, I fly a route with a passenger aircraft.  Naturally, some standard and light cargo actual demand drifts toward my route.  But not heavy - which seems to make sense since I am not supplying any heavy cargo.

But then I start looking around for this heavy cargo, looking for airport pair where it might be appearing, and I can't find it / it's not there.

[groundbum2]

BTW, I have seen an issue of Heavy Cargo "hiding".  For example, I fly a route with a passenger aircraft.  Naturally, some standard and light cargo actual demand drifts toward my route.  But not heavy - which seems to make sense since I am not supplying any heavy cargo.

But then I start looking around for this heavy cargo, looking for airport pair where it might be appearing, and I can't find it / it's not there.

My observation of cargo is that heavy will hide, ie be lets say 5% of potential cargo, UNTIL somebody puts a cargo plane on the CBD city pair. Then after a few months heavy cargo appears. I've certainly this in the domestic USA, where routes don't have much heavy on them, put a cargo plane on and 6 months later boom loads of CH! So yes potential CL and CS seem realistic, potential CH is very low but increases when it starts being flown, and appears in the available column within months.

Simon

[groundbum2]

if I was debugging this I guess I would at, say look at LAX-NYC in todays GW3 and run the algorithm. If the algorithm said "no changes made" then I would backtrack from there, as only 33% of potential is available, and there is 200% capacity being flown daily and has been for a year or more. Perhaps 200% oversupply isn't enough, and it needs extra freight flights added to tip it over the edge to increasing supply?

Simon (20 years datacenters as Oracle DBA so well used to debugging malcontent code...sorry sami...) ;-)

[JumboShrimp]

if I was debugging this I guess I would at, say look at LAX-NYC in todays GW3 and run the algorithm. If the algorithm said "no changes made" then I would backtrack from there, as only 33% of potential is available, and there is 200% capacity being flown daily and has been for a year or more. Perhaps 200% oversupply isn't enough, and it needs extra freight flights added to tip it over the edge to increasing supply?

Simon (20 years datacenters as Oracle DBA so well used to debugging malcontent code...sorry sami...) ;-)

At first, we need to find that (or if) the data is wrong.  Your suspicion is that the actual demand does not add up.

We can get a ballpark figure, and if it is below potential, then we know there is a problem.

But, on JFK-LAX, for example PHL gets into the equation, and PHL also picks up a lot of demand that is not part of JFK demand.  Same with BDL, and even BOS has some overlap in Connecticut....

[groundbum2]

PHL could pull about 35% of NYC demand, though has a lot of it's own supply, plus has 200% cargo flights so still no upward pull towards potential..

Simon

[groundbum2]

has anything been found yet in terms of debugging, still 60% of the potential cargo shown on the charts is AWOL, unless Air America/Mel Gibson stealth flights are going to/from a grass airstrip.. say central park to LAX beach...

Simon

[groundbum2]

any news on a fix? I know this isn't a "sexy" code problem per se, and is more data related, but clearly CBD in complex city-pairs is broken. Look at any complex city pair (LAX/SFO <> NYC/WashDC/London) and there is sky high potential but no more than 30% of this potential is available when all the possible route pairs are examined. I did the analysis back in the original post.

CBD *is* working for simple stuff, like KLAX-KMSY, boom put a cargo flight on and supply goes up to potential within months.

Simon

[Sami]

Currently not on highest prio to-do list, but the "city based demand v.1.0" (or whatever one could that) is planned to have a general update during spring or summer. Depends a bit on how some testing of future concepts goes.

[groundbum2]

just checked and LAX-JFK shows 1.7MILLION available Kg of freight daily and 160K is being flown, so 10%. These figures are identical to those I reported back in April. So it looks like the cargo CBD fix of the last few days did fix simple route pairs, but has worked no magic on complex route pairs where the cargo suggested by the figures is just not there in real life.

I've not looked at all 40 route pairs but will do in the next days and update the spreadsheet. Until this is fixed I don't see that CBD is fit for purpose, and certainly not fit for pax ops.

spreadsheet updated,  there's no other flights at secondary airports at either end stealing the cargo ;-)

https://onedrive.live.com/edit.aspx?resid=B4115FBC2A99A766!1818&ithint=file%2cxlsx&authkey=!AP73HnNs0ffCDu8

 
Simon

[Cornishman]

I intend no offence to the game developers nor anyone that somehow likes the CBD cargo set-up, but I have to fully agree with Simon. CBD whilst a neat idea as a concept is proving a real nightmare to plough through and establish what would be a good route and what would not be good. I agree it is currently not fit for purpose, I could not see myself staying with AWS if this gets implemented for passenger routes, only on the basis I would not be prepared to invest the dozens and dozens of RL hours I had to put in and create a fairly complex spreadsheet of data just to work out what viable cargo routes exist in GW4 from France to the USA under CBD conditions.  That was just to one country! Yes it might be a realistic concept to match real life situations, but we either need a lot better tools to manage this or we don't have it at all, imo.

[groundbum2]

I differ slightly with Jack as I like the CBD concept and spec, but don't feel the code is working correctly for complex city-pairs. Simple pairs,works fine again now the fix is in. My GW3 KLAX-LIPZ that I use as a litmus test has grown nicely post-fix.

Simon

[Verto]

Are we sure that the cargo bug's fix is still working as intended?

Looking around at some GW2 routes it looks like there was cargo growth, but haven't seen any improvement in the past week.