"Over the Hump Airlift Simulation Update and Play Test AAR" Topic

Well after some further research, I put together enough data to do a sample play test and ran through a couple of days of a hypothetical scenario. In 1942 when the Hump airlift to China was suggested to keep China in the war after all the land communications were blocked by the Japanese capture of Burma, General Stilwell suggested that to meet the goal of 5000 tons per month by air would require 140 C-47s and 234 crews in good weather and 300 C-47s and 280 crews during the six month monsoon season.

Doing the math on this hypothetical scenario gives about 50% availability per day for each C-47 to make 5000 tons in 30 days at 2.5 tons per plane, and with 70 planes flying per day on average with 234 crews, each crew would fly about every third day. This plan envisioned daylight only and one round trip per day for the 70 planes up each day. The monsoon variation suggests that half the days would be rained out and then with about 50% average daily availability and 300 planes, the 5000 tons could be made in 15 flying days per month with one daylight round trip per plane on the days that 150 planes could get into the air.

There is data from later in the war on aircraft losses per month and per 1000 eastbound trips (the losses are totals which would include the east bound leg as well as the west bound leg, but measured against the eastbound leg as the main cargo hauling leg). Unfortunately the data set from Air Force documents includes losses from December 1943 through August 1945, and only cargo tonnage for January through December 1943. The largest loss per 1000 trips eastbound was 8 in January 1944, and I would expect losses would have been higher in the early days, so research is continuing on that.

With the data point for 8 losses in 1000 trips though, that is enough to get started on a play test. So, for the play test, I set up a spread sheet for 140 C-47s with an east bound and a west bound time track, and divided each round trip into 8 potential loss events, two events for the east bound takeoff with a full cargo load, one event for potential interception by Japanese fighters, one event for "weather" en route which would also include in-flight mechanical failure and mid-air collisions, one event for landing in China, one event for the west bound takeoff with no or little return cargo, one event for "weather," and one event for landing back at base in India. Since each event is a one in a thousand chance, I picked 0,0,0 on three D10 as the magic number. There was no interception event coming back to India because the planes could fly higher with no load and could go farther north, not fully out of reach, but less chance of being intercepted, plus I only had 8 events to work with and these were all used up. I'm not sure how I'll allocate these for scenarios for later months where there were only one or two losses per 1000 trips, but that is a challenge for another day.

The C-47 average flying speed is less than 180 mph, so I more or less arbitrarily decided that the average speed made good to the target airfield, what with flying through valleys to avoid the peaks that were higher than the cargo loaded ceiling, avoiding interceptors by flying into clouds, etc., would be 150 mph, which on a 10 mile per hex grid would be 20 miles or 2 hexes on even turns and 30 miles or 3 hexes on odd turns. I also made the average distance from the Assam bases to the Kunming bases as 550 miles.

So now armed with all that information, I went through my good weather hypothetical scenario of 140 aircraft, one by one, using a 50% availability factor with D10s, and on Day 1 received only 63 aircraft for the day! A poor result for all that wristage! I started cycling the aircraft through the loading stages, with the minimum two 10-minute turns followed by the "Morale" chance that I have mentioned previously, selecting 50% to match the Stilwell availability factor. I began the day at 0600 and by 0620 the first planes were taking off. In General Tunner's book there is a highlight that during the "Big Day" celebration maximum effort a plane flew across the Hump every minute, so I took that to be sort of the best possible optimum spacing, and limited takeoffs to 10 per 10-minute turn,, even though planes would normally fly from several different bases, and also land at several different bases Chinaside.

As expected some planes did not get off the ground right away after the minimum 20 minutes of loading time, but all 63 were in the air by 0750 with no crashes on takeoff after a lot more wristage. All planes proceeded uneventfully through the interception and weather checks and began landing Chinaside at 1000. Planes were landing successfully and then all of a sudden at 1020 Plane 29, the 12th plane of those available flying, crashed on landing with a roll of 0,0,0!

Planes continued coming in and the first westbound return takeoffs began at 1030. The last incoming flight landed at 1130 and 62 loads at 2.5 tons each for 155 tons were delivered. The last return takeoff went up at 1220, and the first return landings in India began at 1410 with the last of the 62 surviving planes landing at 1600. One plane was lost in 63 east bound flights and that worked out to be about twice the rate of 8 losses per 1000 east bound trips. The longest load time for the east bound trip took 70 minutes and the longest turnaround for the west bound trip took 90 minutes.

It was a pretty interesting day, although as it has turned out so far, the player is pretty much along for the ride, providing the dice rolling and recording the action, and not much in the way of command decisions yet.

The second day was similar, only this time I was able to put 69 planes in the air going through the 139 survivors one by one on the 50% test. The first plane took off at 0620, all were up by 0810, and all planes successfully avoided any interceptions or weather incidents en route east bound. The first landing Chinaside was at 1000, the first return takeoff was at 1030. All 69 planes landed successfully delivering 172.5 tons of cargo, and the last was down at 1150. The last return takeoff was at 1250 with the first landings in India 1410. There were no adverse events en route, and all landings were proceeded uneventfully until Plane Number 89 crashed on landing at 1500. The rest of the planes all landed successfully with the last down at 1630.

As you can tell this method of play takes many dice rolls, and believe me it is a real shocker to see 0,0,0!! "Snake Eyes" has come up twice now, and on average over these two days it should have come up maybe once in all those rolls. With two crashes in two days and 132 eastbound trips, I'm expecting a few days now with no adverse events. Day 3 will start with 138 planes in the pool, and we'll see how it goes for further play tests …

Some observations. First of all it would have been much more fun to actually lay this out on a map and use counters or models to represent the planes rather than recording it all on a spread sheet. Also, since it was a play test for later analysis, I recorded each dice roll. The only thing more tedious than making thousands of dice rolls is actually recording them, and for gaming amusement, I would not normally suggest recording dice rolls or the hour by hour progression of each plane along its tracks, just make the rolls hoping to avoid the dreaded snake eyes and move the planes along the route on the map.

Second, after going along for the ride, the next thing to do is put some player decision making into the game, not actually flying the planes, but perhaps providing a "Command Point" resource that can be expended on things like turnaround time reduction, increased availability, maintenance time reduction when we get to the 50-hour and 200-hour inspection and maintenance, launch search and rescue operations, etc., so that the player can see where the sensitivity is in making a scarce allocation of Command Points to one area or another.

The "game" such as it is, would be to beat the historical tonnages delivered, and flying 140 C-47s could become pretty boring after a few days. With more data and increased granularity among aircraft types, the player would be able to see the value of introducing the C-46s and then the C-54s, plus there are the C-87/109 conversions of B-24s that had a loss rate "five times higher" than the other types.

Playing through a progression of scenarios from the early C-47 days through to the C-87/109, then C-46, then C-54, with an increasing number of Command Points with each successive Commander until reaching the ultimate airlifter, General Tunner, could provide a degree of insight into the Hump operation, and airlift in general, that is a different experience than just reading about it.

I am finding the idea of the Hump simulation intriguing and am still looking for the holy grail of the squadron logs! I'm not so optimistic on finding what I'm looking for, especially since the Air Force's own statistical summary made shortly after the events has a lot of holes in it. The logs may be there somewhere, but apparently no one has ever tried to compile the day by day information if it still even exists. Also, I'm a little surprised that the basic Order of Battle data on overall numbers and types of aircraft is not readily available on a month by month summary basis. There is data for a few dates, such as July 1945 or the first 37 C-47s to arrive in theater, but I haven't found any monthly summaries yet and I thought that would have been easy. But hey, research is fun even if it isn't easy.

And one other observation for now, I have enjoyed a long career in naval ship design and shipbuilding program management and policy formulation, and I am a fervent student of convoy operations in WWII, and especially enjoy Mal Wright's Convoy games. So when I looked at the early airlift goal of 5000 tons per month and worked through a couple of days of many moving parts to get 327.5 tons over the Hump, and then considered that a single Liberty ship could supply about two months worth all by itself, I am both amused and amazed. The entire three year Hump operation delivered about 650,000 tons of supplies to China, the amount that just two moderate sized convoys could have delivered out of the thousands of convoys that occurred in WWII. Of course ships couldn't deliver to China, and once the Burma road was cut, the only option was air, and the Hump operation was magnificent in its conception and execution when you consider everything that was done to move 650,000 tons of supplies over the Himalayas! It is also a commentary on what a shoe string the whole China theater operated on for all those years, keeping the China front going on a total of only 650,000 tons of supplies. Amazing achievements!

Sounds interesting.

I ran into a guy who flew the Hump back in the day, in C-46's. I was getting the car smogged when he started talking about his career.

Seems they would sometimes have to lower their airspeed, flaps, gear in order to try to get the enemy fighters to stall out at altitude, in order to avoid them.

He lost a few friends to mountain peaks in the fog.

Don't recall much else that might be applicable to your game.

If desired, for playtesting, you can set it up on a spreadsheet for faster review. Probably not as satisfying as rolling the dice on your own though.

Mako,

Many thanks for the comment! I had thought about using random number pages from some of my old Ops Research graduate school texts, but for some reason, I like the tactile feel of the good old dice, probably the 50+ years of wargaming experience! If I could generate the random numbers in the spread sheet, I might give that a shot, but none of my ten thumbs have a clue about how to do that …

My friend in the Friday Breakfast Club that I mentioned in an earlier post also flew C-46s and was on the Hump for two years even though there was a rotation policy after one year, that Tunner instituted when he arrived, to replace the previous 650 hours and you're home that caused a lot of overflying and crashes from pilot fatigue. My buddy Paul, now aged 90+ and still flying, but without me as a passenger, flew for a commercial airline before "enlisting" in the Army Air Corps, and because of his flying experience he was appointed as the "Night Pathfinder" by his base Colonel. Paul liked flying the C-46 even though others called it "Old Dumbo" and other less complimentary names, and especially liked flying at night because there was no risk of interception by Japanese fighters.

Paul told me about his timed turn method as in: ok we been flying about this long on this course and there should be a big mountain dead ahead so we'll turn left now and fly for five minutes then turn right then four more minutes and left again … all on total instrument flight rules, when instruments were fairly primitive and didn't include radar … Paul must have been pretty good at night pathfinding because he is 90+ and still talks about it!

Paul also mentioned how the fuel tanks would be measured with yardsticks before departure from China, and the tanks drained down to only 500 gallons for the return trip, and then he recounted many instances when he had to land in India short of his base and beg a few gallons of fuel from the local airstrip and set off again to make it all the way home.

One of the sources I found mentioned about a huge storm that blew up from west to east, the high winds added an hour of time to the west bound flights, and 13 planes went down that day running out of fuel short of even the emergency strips … dangerous flying over the Hump, and General Tunner said that it there was more danger of going down over the Hump than there was in flying bombers over Germany! Not sure if I have enough stats in hand to prove that, but he was there and I wasn't!

I haven't factored in a severe weather event yet, but I'm working on a few chrome features to layer onto the basic game. It is a pretty interesting project so far, and as I said before, I would expect the commercial value of the completed game to be pretty close to zero, so I'll likely post it somewhere as a free download if it ends up looking worthy of more than my own spare time amusement.

Sounds very interesting. I, too, knew an Over-the-Hump pilot when I was a kid. Unfortunately, I was too young to understand the significance at the time and when I was old enough, he had moved. I'm sure by now he is gone.

Sadly, the ranks of the Old Guard are thinning day by day, and anytime anyone can get their story down on "paper" before they're gone it is an important legacy for those who are following along. With the power of information technology, more of those stories can be recorded now than in the past, and if only there were a few laptops around in Assam in 1942-1945, I wouldn't be having so much difficulty in nailing down my orders of battle and daily "counts."

Are you sure that most flights dead-headed back to India? I have read that many Chinese soldiers flew out over the Hump, some to join Stillwell's Chinese army in India, others to pick up aircraft in India, still others on the first leg of their journey to flight training in the US.

According to some of the partial data I have seen there was two way traffic early on with the tonnage going back to India about half the tonnage coming to China, and the lighter loads allowed flights to India to fly higher and farther north. One list I saw for west bound cargo included tungsten and troops. Later on after Myitkina had been retaken by US and Chinese forces, the Chinese troops were transported east bound for the Salween campaign and other operations in China, and it seems that the west bound cargo diminished, although I'm still looking for more complete data on this topic. If you know of good sources for this cargo traffic information, please let me know.

I believe he mentioned the Japanese fighters flew on moonlit nights, but not 100% sure about that. Most nations' pilots did.

Chances of interception with the Mk.I eyeball would be very low, especially without radar direction from the ground.

For Zippy,

This morning I looked up the partial data I have in the 10th Air Force Historical Study for 1942 from a letter by Brigadier General Haynes covering the two month period from 8 April to 14 June 1942. During this two month period, 700 tons was carried east bound and 380 carried west bound. The predominant east bound cargo was gasoline and oil at 422 tons and included 208 passengers at 17 tons and don't anyone even mention the 2 tons of cigarettes, while the predominant west bound cargo was 4,303 passengers at 324 tons followed by tin at 27 tons and tungsten ore at 24 tons.

The China National Aviation Corporation based in China also flew over the Hump in DC-3s (the civilian airliner that was the C-47 in Army Air Force configurations) and I believe that they were the main hauler of Chinese troops west bound for training. The Army CBI History "Green Book" volume 1 mentions that the troop lift to the Ramgarh training base in India began in October 1942 and 4000 troops were carried in that month. There are no tallies by month, but at one point the number of troops flown in per day was limited to 650 by congestion on Assam transportation facilities.

The Air Transport Command data summary only includes tonnage east bound, and with ATC assuming command of the Hump operation in December 1942 the ATC data summary includes east bound tonnage beginning in January 1943 through August 1945, and as mentioned before, aggregate aircraft loss data begins in December 1943 and goes through August 1945.

So while there was some west bound traffic, especially during the build-up by CNAC of Chinese forces training for the offensive into Burma launched from India in 1944, it seems that this west bound traffic was lighter than east bound, and diminished over time, especially with the increasing tempo of Chinese operations by the main Chinese forces operating in China in 1944. Among these scanty shreds of data there is no mention of west bound aircraft lost to Japanese fighters. The airlift squadron logs might tell the whole story on the east-west traffic and aircraft losses, or closer to the whole story, but these may now have been lost to history if they ever were accurately recorded at the time.

For Mako,

I agree with your assessment of the chances of interception with the Mk. 1 eyeball, especially flying over remote terrain like the Himalayas. I have recently been doodling around with Lee Brimmicombe-Wood's new "Nightfighter" game and the introductory scenario is a day-fighter Hurricane trying to intercept a bomber raid over southern England on a moonlit night employing the Cat's Eye sensor. The difficulty level is listed as "impossible" and to date I haven't spotted even one bomber before they have all escaped off the map. The background for this scenario says:

"Early nightfighting relied on single-seat day fighters cruising the skies in the hope that they might find the enemy. Pilots needed "cat's eyes" to pick out bombers in the dark. In practice the technique resulted in few kills and more defending aircraft were lost due to night-flying accidents than enemy aircraft were shot down."

Defending one's home capital city might be worth expending a few aircraft on low probability of success moonlight flying, but I'm thinking that chasing Gooney Birds over the Himalayas on moonlit nights was probably a pretty rare thing, and most likely not very successful if it occurred. If I dig deep enough, I'll probably find an instance or two of a transport being shot down at night, but so far I haven't found any evidence of it, but I've got my eyes peeled!

Also, I can highly recommend Lee's "Nightfighter" for those interested in exploring the progression of nightfighting from cat's eyes up through all the different aspects of radar measures and countermeasures. Lee's "Bomber Command" game is coming along too, and the night raids over Europe can be linked back to "Nightfighter" for combat resolution. Ok, now I know that won't be as interesting as the Hump airlift, but I'm buying it anyway!!