Survivabilty – Lancaster vs Halifax

There is an interesting comparison in ‘The Berlin Raids’ by Martin Middlebrook between the survival rates of Lancaster crews from number 1 Group and number 5 Group during the Battle of Berlin.

———- Sorties — Shot-Down —- Aborted —– Killed — Baled-Out

1 Group — 2598 — 143 (5.5%) — 207 (8.0%) — 780 — 242 (23.7%)
5 Group — 3002 — 128 (4.3%) — 166 (5.5%) — 750 — 141 (15.8%)

These two groups operated the same aircraft type, in the same operations, but throughout the Battle of Berlin number 1 Group consistently loaded their Lancasters with the maximum possible bomb-load; a sample averaged 9317lb for 1 Group compared to 8986lb for 5 Group during the battle.

Some Lancaster crews took to jettisoning some of their bomb-load (including 4000lb ‘cookie’ bombs which could not be dropped ‘safe’ and therefore could be seen exploding) over the North sea in an effort to regain some of the performance and height of their aircraft.

It also seems that the chances of crew survival were affected by the morale of the crews and their willingness to ‘press-on’.

My Dad flew in both Halifax’s (51/578 Sqdns) and Lancaster’s (35 PFF Sqdn). His favourite aircraft of the two was the Halifax, he says it was a much sturdier aircraft and could take more punishment. Dad was a midupper gunner and has said that during the corkscrew manoeuvre in the Lancaster the tail planes moved so much he was sure they were going to snap off whereas in the Halifax there was hardly any movment doing the same manouvre.
He was on the Nuremburg raid in March 1944 and Berlin the week before. On an earlier Op whilst on Halifax’s the aircraft had a large hole blown in the fuselage between the midupper and rear turret. My Dad is sure that if he’d had been in a Lancaster at the time he wouldn’t be here today!
Type683

`

I’ve heard & spoken to other crew who feel the same general thoughts !
(over a 35-40 year period, sporadically !)

My ONLY thing I’d add to the ‘debate’ is that if it were MY life on the line, I’d rather get into a Late Mk Halifax (B.III, B.VI/BVII) rather than a Lanc’ B.I/B.III with the knowledge that IF the crate I’m in were to get hit, I’m far better off in a Radial-engined bird that can afford to suffer a few knocks in the engine dep’t, rather than a B.I/B.III Lanc’ with it’s (easier to knock-out/disable) liquid-cooled Merlins that might take just “coin-sized-shrapnel” to seize (or cease !) working & increase the risk of bailing

Haven’t seen that mentioned till now

Most folks know about the Spar/limited movement/squeeze in the Lanc’

(I love BOTH types BTW :D)

My Dad flew in both Halifax’s (51/578 Sqdns) and Lancaster’s (35 PFF Sqdn). His favourite aircraft of the two was the Halifax, he says it was a much sturdier aircraft and could take more punishment. Dad was a midupper gunner and has said that during the corkscrew manoeuvre in the Lancaster the tail planes moved so much he was sure they were going to snap off whereas in the Halifax there was hardly any movment doing the same manouvre.
He was on the Nuremburg raid in March 1944 and Berlin the week before. On an earlier Op whilst on Halifax’s the aircraft had a large hole blown in the fuselage between the midupper and rear turret. My Dad is sure that if he’d had been in a Lancaster at the time he wouldn’t be here today!
Type683

All good stuff chaps. What’s good here is that while we are setting out a stall to test a hypothesis, no one is falling into the poor analysis of selecting stats just to ‘prove’ their preference, but exploring the options from different angles, and accepting that these data sets can show different – valid – answers.

My earlier remark wasn’t that it was a futile exercise – clearly we all appreciate the higher level issues of the loss rate and the real costs – but that the other variable factors (IMHO) than an aircraft type make putting too much emphasis on the data risky.

That said, I think there’s a good chance of getting some meaningful analysis on the topic and questions, and the group work is more likely to result in original ‘tested’ results than any one person’s efforts.

Regards

Good to see your input James. By the way, I agree – the statistics can be made to prove anything, and a lot of the extrapolation I did is essentially meaningless. It also looks like we were essentially on the same track about why the Halifax was safer.

Oh, that’s easy … But that wasn’t what you were after was it?

Haha LOL, nope!

This perception stems from me reading over time that the Halifax in it’s fully developed form (the MkIII) had more or less overcome all its earlier limitations and become a viable front line offensive aircraft.

The hitch of course is that (as Air Ministry alluded to above) it wasn’t as hard hitting as the Lancaster, so it wasn’t preferred in that role. I guess you could say, it wasn’t as efficient a weapon [of MD – duck for cover].

Therefore as you say, it got diverted to other tasks.

But when it lined up with the Lanc for good ‘ole bombing duty, I “think” it had a not dissimilar loss rate and a better survivability rate.

Not that it matters much in the scheme of things … I don’t want to make a BoB Hurricane vs Spitfire debate of all this. Appropos which, I do appreciate Clive James’ remarks about my fellow countryman.

Eddie’s on the right track for what I’m trying to do.

Cheers James, Don

Clearly, figuring out what I want to know (how “safe” was the ultimate form of Halifax [MkIII] vs its stablemate the Lancaster) will require a bit more digging on my part.

Oh, that’s easy. The Halifax was ‘lots’ safer as the loss rate on Coastal wasn’t as appalling as Bomber Command; Halifaxes were used in (wartime) Coastal Command, Lancasters weren’t, so Halifaxes were safer.

But that wasn’t what you were after was it?

It all depends on how long a period you want to analyse for – on a “one op basis”, clearly the Lanc comes out ahead – for the direct comparison of the Lanc vs Halifax III on the night of 24/25 March 1944, 93.9% of Lanc crewmembers survived, 91.6% of Halifax crewmembers survived, and at those loss rates, after 16 ops, the survival rate is even, with 28.2% of Lanc crews still flying vs 10.8% of Halifax crews, with 13.8% of the Lanc crews having safely baled out and 31% of the Halifax crews having safely bailed out.

The statistics are horribly grim though, either way.

In a sense the Halifax was far more survivable, because as an aircraft that on virtually every op was lost at a higher rate than Lancs, whether in the Hercules or Merlin engined form, it was far less likely to be sent on the most dangerous raids (they only went to Berlin on nine of nineteen occasions).

Very interesting thread. I think top level data might be meaningful, but IMHO there are too many variables for more detailed data to have much relevance, even to ‘which type was more survivable’.

One other thing I seem to see about losses is a lot of the time crews delayed in bailing out – the aircraft would be attacked, and either by the account of the nightfighter crew or isolated survivors, there was a significant amount of time that often elapsed between the aircraft being initially hit and breaking up. I assume this was because crews did not realise how serious the damage to their aircraft was and that every second spent trying to save the aircraft was a second less that would be available to bail out.

This excellent point is predicated on the assumption that each of these attacks was going to be 100% successful in knocking down the bomber; secondly that the human reaction would be to bale out as rapidly as possible after receiving damage, whereas the normal reaction would be to try and ‘save’ themselves by saving the bomber in many cases; as indeed bombers were saved on occasion, which is another factor in trying to arrive at a comprehensive statistic for these losses.

So firstly one needs to asses how many aircraft got back after an attack of this kind – by definition, those attacks were all ‘unsuccessful’, and thus the data set includes the ‘glancing blow’ or uncompleted attacks right through to the fully completed attack where the bomber crew were incredibly lucky and their aircraft hung together afterwards – those being from a significant number to what is actually a statistical anomaly.

I’m also curious as to what point a crew ‘should’ bale out, as implied above. Some cases are obvious, perhaps – major fire and major structural failure for instance. But on the other hand the history is filled with stories of crews bringing back the aircraft against the odds with significant damage, fire and injury. These were widely disseminated at squadron level and in award citations, reinforcing the (IMHO) natural human reaction of trying to save the aircraft so it would save the crew – I’d suggest, even without data, contributing to greater losses from crews trying and failing to save their machine.

IIRC, the best example of a scientific analysis of this kind of distortion of data is in Stephen Jay Gould’s ‘The Median is Not the Message’ essay. The key point here is that he talks about ‘skewed’ and ‘extended tails’ in data.

http://www.phoenix5.org/articles/GouldMessage.html

Regards,

Thank you both

Creaking Door … I don’t have Chorley, I’ll see if my library has it. Sounds fascinating but as you say, sobering.

Eddie … sounds fine to me, but on reflection this tells us the Battle of Berlin was a hard campaign. Maybe I should side with Kev on this one?

Clearly, figuring out what I want to know (how “safe” was the ultimate form of Halifax [MkIII] vs its stablemate the Lancaster) will require a bit more digging on my part.

cheers D

Don,

If you have it, look at “The Berlin Raids” by Middlebrook. It’s where my raw data came from. It’s the overall Battle of Berlin loss rates, including both Merlin and Hercules Halifaxes.

The way I calculated the odds of being shot down on 30 op tour were the chances of survival for a single op (for a Lanc 100%-5.2%, so 94.8%), and then I did that number to the power of 30 (because to survive those odds, you would have to be in the 94.8% each time). Thusly – 0.948^30=0.201=20.1% chance of returning home safely.
Then, for the remaining 79.9% of aircrew, who were shot down, I said that 19.2% of those survived. 79.9% x .192 = 15.3%

From those two numbers added together, we get the Lancaster survival rate of 34.4%.

So, for the 17 Berlin ops, which was the highest number known for any crew, .948^17, the Lanc crew’s safe return rate would be 40.3%. Of the remaining 59.7%, 19.2% survived, which is an additional 11.5% of the total, making an overall survival rate of 51.8%.

The Halifax safe return rate over the same 17 ops would be a meagre 20.5%, and an additional 35.0% of the remaining 79.5% would have survived by escaping shot down aircraft. This would have been 27.8% of the total, making an overall survival rate of 48.3%

Those data in Chorley’s Bomber Command Losses, are they perhaps in tabular form and/or in some other way relatively easily worked with?

I have Middlebrook’s BCWD, and whilst I agree that will supply some aggregates, it doesn’t include a breakdown of aircraft by mark or sqdn (thinking here of Halifax II, III, V etc.).

BTW I don’t share your memory of BCWD re. “most likely to die”, I thought Middlebrook said it was the pilot (typically last out after trying to hold the aircraft stable for as long as possible)…

Chorley records day-by-day, each bomber lost (including serial number) and, most poignantly, the names of all the crew and whether they became a POW, evaded capture, were injured or, in the vast majority of cases, were killed.

It is not really suited for easy analysis but it makes for very sobering reading as it is common to find a page with seven bombers listed from which not a single crew member survived. I think it is a very powerful format; listing the bombers and crews lost on the Nuremberg raid 30/31 March 1944 takes sixteen pages.

You could well be right about the most dangerous crew position…..it has been a while since I looked into it.

Huh – the other way round?

Eddie … that’s not at all what I expected.

I’m interested in your maths, any chance you could send me your spreadsheet or whatever so I can get my head around what you’ve done?

I assume you’ve used Halifax loss rates from the BoB? IIRC this is when they peaked leading to the MkII (and V?) being pulled off front-line duties (along with the Stirling).

I remain unmoved (so far) in my view the Halifax MkIII was the aircraft of choice for the survivor.

Best, thanks, Don
[PM for email address]

Just out of curiosity, I wonder how those rates compare with American types?

Did night AC losses have a greater aircrew loss rate than daytime losses?

I think I am correct in saying that comparing the Lancaster and the B-17 the chances of crew survival were more or less reversed; only about one in seven Lancaster crew survived being shot-down, about one in seven B-17 crew were killed.

There would be many reasons for this difference but to my mind probably the most important would be the time available to the crew between the aircraft being hit, the bale-out decision being taken and the point where the crew could no longer get out of the aircraft.

A typical nightfighter attack on a Lancaster, Halifax or Stirling would come from very close, underneath in the case of a Schräge Musik attack, and a lot of concentrated cannon-fire would hit the bomber. This sort of attack would typically target the fuel tanks between the engines rapidly leading to catastrophic fire, loss of engines and possibly main-spar damage. Loss of control or break-up of the aircraft would occur very quickly giving the crew little chance for escape.

A typical fighter attack on a B-17 would be more fleeting, possibly head-on, from greater range and probably with considerable ‘deflection’ shooting required. This sort of attack would almost certainly lead to fewer cannon hits, less concentrated, about the bomber. The B-17 also enjoyed a reputation for its ability to absorb battle-damage.

There is a further bit of data in “The Berlin Raids” that I found. The short answer is:
Lancaster – 19.2% of crew survived
Halifax – 35.0% of crew survived.

Obviously this is around twice as survivable. So, this leads us to…
The Lancaster crew’s survival rate per op – 94.8% of crew survive by returning home, and a further 1.0% survive by escaping from aircraft lost on ops. Overall, 95.8%.
The Halifaxes crew’s survival rate per op – 91.1% of crew survive by returning home, and a further 3.1% survive by escaping from aircraft lost on ops. Overall, 94.2%.

So, on any given day, a Lancaster crew was 1.6% more likely to live at the end of the day.

It gets more complicated when you consider a full tour, however. Given that at the above loss rates (which I think are atypically high, given that they are solely based on attacking Berlin – the highest number of Berlin raids known to have been flown by a crew was 17), a tour of 30 ops would have around an 80.0% chance of being shot down for a Lanc, or for a Halifax, a whopping 93.9%, the odds end up looking different:
The Lancaster crew’s survival rate per tour – 20.1% of crew survive by returning home, and a further 15.3% survive by escaping from aircraft lost on ops. Overall, 35.4%.
The Halifaxes crew’s survival rate per op – 6.1% of crew survive by returning home, and a further 32.9% survive by escaping from aircraft lost on ops. Overall, 39.0%.

So, I think at more typical loss rates, the survival rate for a full tour would work out to be significantly different. If someone can come up with numbers for, say, 1943 overall, I’m sure we can analyse it! The overriding thing that I see is just how grim the prospects of completing a tour were in the winter of 43/44.

Thank you everyone

Wow everybody, thank you! I love this forum.

Creaking Door
Those data in Chorley’s Bomber Command Losses, are they perhaps in tabular form and/or in some other way relatively easily worked with?

I have Middlebrook’s BCWD, and whilst I agree that will supply some aggregates, it doesn’t include a breakdown of aircraft by mark or sqdn (thinking here of Halifax II, III, V etc.).

BTW I don’t share your memory of BCWD re. “most likely to die”, I thought Middlebrook said it was the pilot (typically last out after trying to hold the aircraft stable for as long as possible). Hmmm … or it may have been Hinchcliffe’s “The Other Battle” that makes that observation.

Paul
Thanks for those “survey” data. Yes, maybe unscientific, but then again the surveyed population was certainly a sample of the survivors, only issue is how many in the sample.

At face value that survey indicates the Halifax was nearly 3 times as survivable as the Lancaster.

Eddie
You’re noting Halifax losses in the order of 1.5 to 2 times the Lancaster loss rate. If the Halifax was 3 times as survivable, I know which one I’d rather be in!

J Boyle
I did look into US aircraft in similar vein some years ago, the data were more readily accessible online. IIRC the survivability rate was many times better than the UK bomber counterparts.

Again IIRC, the B17 was “built strong”, would take a lot of punishment, and when damaged was often a stable platform to exit from. Bear in mind also 12x 50cal but only 4,000 pound bomb-load. Contrast this to the UK aircraft.

And the B24 was, I think, faster? For example, it didn’t/couldn’t formate with the B17’s.

I’ll see if I can track down my prior analysis.

Air Ministry
Thanks for that. RAF Bomber Command’s primary criterion seems to have been the relative offensive capability of its weapons. Hence the natural preference for the Lancaster, well supported by your data observations.

In my view it would be safe to assert that BC relied on darkness (and to lesser extent electronic warfare) for defence/survival rather than demand more rugged aircraft.

Kev
Yes … BUT
(a) sudden loss of an entire airframe would be egually lethal no matter what therefore I feel can be taken out of the picture (it will have a levelling effect on the findings)
(b) agree on losses over the sea etc., but I still contend that an analysis can be conducted in the aggregate and it will tell us something.

I won’t be at all surprised if we arrive at something along the lines of Chuck Adams’ survey (via Paul above)

Creaking Door and Air Ministry
Is there any chance you can post here, or email to me [PM for address] relevant/accessible data you may have? I’m more than happy to analyse the data and post the results here.

ALL
Thanks again, this (topic) has intrigued me for years and I’m chuffed to think we may see an answer.

D

First of all, regarding post 11, how many Wellingtons had mid upper gunners?

I don’t see how you can ever come to an accurate conclusion on this. For instance, take a look at all those aircraft lost without trace, most over the North Sea. No way of establishing the cause for the loss of the aircraft or even whether the crew managed to take to their parachutes if that’s what they decided to do. Taking into account the aircraft that ditched with the full crew aboard, how many might have suucessfully exited the aircraft by parachute had they been over land?

How about aircraft that were instantaneously destroyed by flak or fighter attack? How do you take them into account?

Far too many variables.

Regards,

kev35

Nothing scientific from me to back things up I’m afraid, however, over the past ten years chatting to veteran Bomber Command aircrew I have been told quite a few times that it was easier to get out of a Halifax in a hurry than a Lancaster. Something about the amount of space inside and the positions of the escape hatches.

I’m sure others will have more to say on the matter.

Interesting question though so will enjoy reading the replies.

Regards
archieraf

Having recently seen the replica Halifax at Elvington its certainly allot fatter than the Lanc, i guess that also makes it a bigger target.

Number of aircraft written off per 100 tons of bombs dropped (Halifax compared to Lancaster):-

1943
April: 2.95 Halifax, 1.46 Lancaster
May: 2.81 Halifax, 0.93 Lancaster
June: 2.28 Halifax, 1.17 Lancaster
July: 1.73 Halifax, 0.73 Lancaster
August: 2.49 Halifax, 0.8 Lancaster
September: 2.45 Halifax, 0.98 Lancaster
October: 3.31 Halifax, 0.84 Lancaster
——————————————–
Average/7 mths: 2.46 Halifax 0.95 Lancaster

Hi folks,

Sorry this is all a bit rushed (long day and I’m pooped) but these stats were compiled by a man from the Ministry who was trying to defend against Sir Arthur Harris’ figures supporting his plea for Halifax production to be switched over to Lancasters. In other words, they are more favourable to the H.P. than Sir Arthur’s own statistics.

And the brutal truth is that this is how the merits of these two aeroplanes were judged by Bomber Command and the M.A.P.; not so much on crew survivability as on how many bombs they hauled before being lost.

It’s an interesting topic, though, and I’ll be back when I’ve caught up with the day. 🙂

What I want to know is that the number of Halifax lost in total never exeeds 2.000 examples but the general total loss rate for the Lancaster is 3.500 (roughly taken) while the total number of Halifaxes built was 6.178 and over 7.000 Lancasters.

Cees – my statement that the Halifaxes were always lost at a higher rate than Lancs is borne out by the overall statistics from the Berlin raids (which I think just about any scholar of the subject would class as being as tough as targets came):
Sorties despatched – Lancaster: 8,291, Halifax: 2068.
Failed to return – Lancaster: 601 (5.2%), Halifax: 285 (8.9%)

And just to make the point about even the late Halifaxes being lost at higher rates than Lancs, on the March 24/25 raid on Berlin, (chosen at random from when the only Halifaxes operating where the Hercules engined ones), the statistics are as follows:
Sorties despatched – Lancaster: 577, Halifax: 216.
Failed to return – Lancaster: 44 (7.6%), Halifax: 28 (13.0%)

The differential between the losses of each type was repeated on almost every occasion that they operated together. I also think it’s fair to say that generally both types would be relatively evenly distributed through the stream. Both were quality aircraft, but I don’t think it’s snobbery to say the Lanc was a more effective bomber, especially considering that, in the essential arithmetic, there were 122 tons of bombs dropped per Halifax lost, against 187 tons dropped per Lanc lost – over 50% greater tonnage.