The Germans inflicted, I read recently, a 50% greater kill rate on their enemies than was inflicted on them. As Alex has already told me this was probably correct up to 1943, but what was their kill rate after this period?

Can someone also please tell me why, when referring to artillery, self propelled guns, tanks etc. the British refeffed to the calibre sizes in pounds whereas the Germans used the metric system. I was always under the impression the metric system (although developed in the 18th. Century) was not universally adopted until the 1960's.

quote:

---

Originally posted by John Wood:
The Germans inflicted, I read recently, a 50% greater kill rate on their enemies than was inflicted on them. As Alex has already told me this was probably correct up to 1943, but what was their kill rate after this period?

---

The real issue here is not kill rates, but relative combat effectiveness. Kill rates will vary with the situation. What is useful to know is whether, regardless of combat environment, whether one side was performing better than the other.

This issue of German combat effectiveness in WWII has been address basically four times, so I will be making multiple posts on the subject to address the various efforts.

The pioneering work on the subject was done by Trevor N. Dupuy, with his book Numbers, Predictions and War (we will abbreviate it as "NPW"). His subsequent works (Genius for War and Understanding War) further developed this theme. All these books are available from TDI.

His basic approach to measuring human factors was to design a model of combat that he had confidence in, and then look for consistent performance differences between armed forces (and even among different divisions) to measure human factors. As such, the model is used to factor out other influences on the battle, like terrain, posture, weapons, etc....leaving one with the only unexplained difference being human factors, modeling error, and random variations. This approach, of course, assumes that the model is doing a reasonable job overall in measuring historical engagements.

The Trevor N. Dupuy designed Quantified Judgement Model (QJM) was used (although this method could be used with any combat model). The Tactical Numberical Deterministic Model (TNDM) is the current version, and is described elsewhere in our website.

To vaidate or test his model, Trevor used 149 engagements, most from WWII. For each of these engagements, he calculated a "Combat Effectiveness Value". This is a catch all value that measures morale, training, leadership, etc. A number below 1.00 means that Allies were less effective than the Germans. A value greater than 1.00 means that the Allies were more effective than the Germans. For this discussion, I used average values, but they do vary by situation, unit and combat. See pages 235-237 in NPW for more detail.

For the two WWI cases (two Germans attacks versus the British in 1918)....the average CEV is .68

There were 60 cases from the Italian Campaign in WWII from September 1943-June 1944. In these 60 cases, the average Allied CEV was .71. This includes 13 cases when the Germans were attacking. In the case of the 13 German attacks, the average Allied CEV was .58.

In the case of the 60 Italian Campaign engagments, 28 were with British forces (average CEV of .65) and 32 were with US forces (average CEV of .77).

There was only one Italian Campaign engagement after June 1944, it was in September 44, with a CEV of .57. One should not draw a conclusion from one engagement.

There is some French 1944 data. There was 6 engagements from Jul - Sep 1944, all US versus German (US attacking in all cases). The average CEV for these engagements is .49, meaning that the relative perfomance difference between the US and Germans was greater (worse for the US) than it was in Italy. This is only 6 cases, so I would be hesitant to draw to much from it.

Trevor then did 14 engagments from November-Decembe 1944, all US versus German (Germans attacking in one case). The average CEV for these engagements was .76.

Therefore, looking at the QJM data, we are looking at:

WWI --- 1918 France --- UK --- 2 cases -- .68
WWII -- 1940 France --- F ---- 1 case --- .76
WWII -- Italy(9/43-6/44)UK -- 28 cases -- .65
WWII -- Italy(9/43-6/44)US -- 32 cases -- .77
WWII -- Italy (9/44)--- US --- 1 case --- .57
WWII -- France(7-9/44)- US --- 6 cases -- .49
WWII -- France(11-12/44)US -- 14 cases -- .76

More to come.

In 1995, we did the casualty estimate for the Joint Cheifs of Staff for Bosnia. As part of that estimate, we did a number of runs of battalion-level actions lasting a few hours using the TNDM. Furthermore, the model was being used extensively by South Africa (among other nations we have sold it to), and they clearly were using it for battalion-level combat. So, I figured it was time to go through and properly validate the model for battalion level combat.

For this effort we used 23 battalion-level engagements from WWI, 23 from WWII and 30 post-WWII. The results of this validation were written up in The International TNDM Newsletter, Volume I No 5 and No 6. Hopefully, my web people will post these newsletters on our site at some point in the future.

For WWI, we had nine engagements of US Marines versus German in 1918. They produced an average CEV of 1.1 in favor of the Marines, with a median CEV of 1.2.

In the same area and time frame, we also had 14 engagements of US Army versus Germans. They produced an average CEV of 1.1 in favor of the US Army, with a median CEV of 0.85.

This tends to point to close parity between the US Army and USMC, although the higher median figure would indicate the USMC may have been more consistently better. The USMC had a higher CEV than the Germans in 5 out of the 9 cases, while the US Army had a higher CEV than the Germans in 5 out of 14 cases. Regardless, it does appear that in 1918 in WWI, the US were the equal of their German opponents.

We only had four US versus German WWII engagements. They had an averge CEV of 1.3, in favor of the US and a median CEV of 1.3. Because of the nature of the engagements selected, and the low number of cases, one should not draw too much from this conclusions. Two engagements were from 42/43 and two were from late 1944. Both pairs of engagements had similar CEVs.

We did have 12 engagements of UK versus Germans in February 1945....all from the same day and same operation. The Average CEV was 0.9 (against the British) with a median of 0.8.

In that book, he has a chapter on "German Combat Effeciency". In there, he comes up with a "German casualty inflicting efficiency superiority of 2.4". Unfortunately he does not provide a set of tables showing the examples or cases that he used or the numbers used. He does not show his math. As such, we have not been able to evaluate this work.

quote:

---

Originally posted by Chris Lawrence:
The third major attempt to measure combat effectiveness (US versus German) was done by Niklas Zetterling in the book Normandy 1944. (I'm probably missing a few here, like Bob McQuie's work in the 1970s, Dean Hartley's analysis of the LWDB in the 1980s, etc.).

In that book, he has a chapter on "German Combat Effeciency". In there, he comes up with a "German casualty inflicting efficiency superiority of 2.4". Unfortunately he does not provide a set of tables showing the examples or cases that he used or the numbers used. He does not show his math. As such, we have not been able to evaluate this work.

It is mathematically derived based upon a formula that multiplies the number of soldiers committed by the number of casualties for one side and divided by the numbers committed and casualties for the other side. It is then multiplied/divided by a factor of 1.4 to account for the defense.

---

In effect I can be said to have used the same methodology as shown by Trevor in Numbers Predictions and War, page 99. I have used the overall casualties for the campaign and assumed an average allied theatre numerical superiority of 3-1. The choice of 1.4 rather then 1.3 to account for the advantage of defence was based on east front data. It must of course be remembered that the result I show is the "Score Effectiveness", whihc means that the "CEV" is the square root, whihc would suggest a German CEV superiority of about 1.5.

In these 11 cases, it does appear like there was rough parity in perfmormance.

quote:

---

Originally posted by Alex H:
Is it 'fair' to convert US vs German combat efficiency ratio to US vs Soviet efficiency ratio using the German vs Soviet efficiency ratio as a benchmark?

---

I cannot find a strong arguement why not. There are lots of little reasons (i.e. different German army on different fronts, nature of fighting was different between two fronts, values are both from mathematical constructs....and so we are comparing two abstract measurements, etc.). One must understands that when one compares two slightly fuzzy numbers one does produce an even fuzzier number. Still, it would be possible to rank and rate the CEV of all the armies in WWII relative to each other, and this would have validity if one had a sufficient number of combat examples.

P.s. no-one has answered my second question yet.

This work was done by Richard Anderson and I, and was published in our Enemy Prisoner of War Capture Rate Study, Phase I & II

In the study we did do a direct comparison between US performance in Italy versus Ardennes. This is summarized below:

-----------------------Ardennes------Italy

US Successful Attack
--Number of Cases------28------------22
--Percent Success------68% ----------59%
--Lowest Ratio---------1.15 ---------1.67
--Highest Ratio--------7.83 ---------4.25
--Average Ratio--------2.24 ---------2.50

US Failed Attack
--Number of Cases------13 -----------15
--Lowest Ratio --------1.23 ---------0.72
--Highest Ratio -------2.24 ---------4.28
--Average Ratio -------1.57 ---------2.52

German Successful Attack
--Number of Cases -----11 -----------5
--Percent Success -----37% ----------29%
--Lowest Ratio --------1.05 ---------1.53
--Highest Ratio -------9.14 ---------5.12
--Average Ratio -------3.92 ---------2.21

German Failed Attack
--Number of Cases------19------------12
--Lowest Ratio--------0.34 ----------0.73
--Highest Ratio ------2.40 ----------5.87
--Average Ratio ------1.17 ----------1.87

This looks only at success....so we then looked at casualty effectiveness.

-------------------------Ardennes-------Italy

US Successful Attack
--Number of Cases--------28--------------22
--Average US losses------207-------------463
--Average German losses--541-------------538
--Times US losses lower--14--------------10

--No. of Type 5+ --------14--------------11
--Av. US losses, type 5+-212-------------456
--Av. G losses, type 5+--606-------------727
--Av. US losses, type 4--203-------------470
--Av. G losses, type 4---476-------------350

US Failed Attack
--Number of Cases--------13--------------15
--Average US losses------223-------------413
--Average German losses--502-------------427
--Times US losses lower-- 8-------------- 7

German Successful Attack
--Number of Cases--------11-------------- 5
--Average German losses--428-------------851
--Average US losses------1185------------727
--Times G. losses lower--8----------------3

German Failed Attack
--Number of Cases--------19--------------12
--Average German losses--253-------------419
--Average US losses------222-------------482
--Times G. losses lower--7---------------6

Italian Campaign
----------------------Total-------Total
----------------------Force Ratio--Loss Ratio
All US Attacks (37)-----2.18 to 1--- .89 to 1
US low odds attacks(3)--1.15 to 1--- .27 to 1

All UK Attacks (21)-----2.07 to 1---1.33 to 1
UK low odds attacks(4)--1.30 to 1---2.31 to 1

All German Attacks (17)-1.59 to 1--- .99 to 1
G low odds attacks(7)--- .85 to 1--- .57 to 1

Ardennes Campaign Data
All US Attacks(41)------1.69 to 1--- .40 to 1
US low odds attacks(12)-1.29 to 1--- .69 to 1
All German Attacks(30)--1.52 to 1--- .55 to 1
G low odds attacks(13)-- .85 to 1--- .38 to 1

1. The Germans and the US were roughly equivilent in combat effectiveness, with the US being within 20 to 30 percent of the Germans (possibly lower). This appears to have been especially true in Italy, although they may have had the same combat effectiveness in the Ardennes....

2. The Germans and the UK were within the same order of magnitude of combat effectiveness, with the UK perhaps being somewhat inferior (by 20 to 50 percent).....

etc.

quote:

---

Originally posted by John Wood:
Can someone also please tell me why, when referring to artillery, self propelled guns, tanks etc. the British refeffed to the calibre sizes in pounds whereas the Germans used the metric system. I was always under the impression the metric system (although developed in the 18th. Century) was not universally adopted until the 1960's.

---

The reference to gun sizes in pounds dates back from the Napoleonic Wars and before. It is the actual weight of the shot thrown by the gun and was pretty much universally used during the Napoleonic Wars. Needless to say, as shot (and shells) became more sophisticated, this became a somewhat bizarre standard. I do not know when measuring gun by bore diameter came into being, but by the time of WWII, all the major nations, except England, had switched to that.
The change to metric first occurred in France with the French Revolution (1789). As the French was the leading culture influence in the world during the 1800s, the rest of Europe soon followed suit. I do not know when Germany switched to metric. I do believe the Russians switched as a result of their revolution (1917). I believe the British made the switch in the 1960s. The US still has not switched to metric, although during WII, the US Army used metric measurements for its bore sizes (above .50 caliber), while the Navy still used inches, which is why we have the US Army with a 3" gun.

quote:

---

US Failed Attack
--Average Ratio -------1.57 ---------2.52

---

Hmm...

The average ratio for a US Failed Attack in Italy is higher than that for a Successful Attack. I guess this is a 'raw' number and is not adjusted for all factors included in a CEV calculation?

quote:

---

Originally posted by Alex H:
I guess this is a 'raw' number and is not adjusted for all factors included in a CEV calculation?

---

Yes. There is no QJM/TNDM methodology in the data from the EPW reports. As such, none of the conditions of combat (terrain, weather, force mix, air support, etc.) are accounted for. One still notes definate patterns, though.

In the past, people have chosen to ignore Trevor's conclusions by simply dismissing the model. These produce similar conclusions using no model. This would tend to not only suppport the conclusions, but also the model.

There has also been a tendency in the OR community to claim that you cannot measure human factors. This also clearly shows that one can, and does not need an archane model to do so (although it does help). What is always needed is more examples (cases).