This is the fifth and final part of my extended review of The Stress of Battle by David Rowland. It is such a strong piece of operational research on WW2 heroism that I thought that it would be useful for wargame designers (and players) to understand what the research evidence is for what went on in WW2 battles. This part is on the effects of heroism and combat degradation.
Combat degradation is a measure of how less effective weapon systems and individual soldiers are in actual combat when compared to training exercises and range work. A score of 1.0 is equivalent to not being degraded at all. Degradation to 0.3 would mean that it was operating at 30% of its peacetime range effectiveness.
the analysis by Rowland’s team broadly matches that done by Wigram in 1943, that there are three classes of effectiveness.
About 20% of those involved could be classed as heroes (26% for guns, 9% for tanks).
Of the rest, one third were ineffective (either they didn’t engage, or what they did do didn’t have any significant impact) (27% of the total);
The remaining two-thirds were about 30% effective (53% of the total);
Weapon systems crewed with at least one hero were about five times more effective than those with no heroes;
Overall effectiveness of a unit = 0.2+([Heroes/gun]*0.8)
Leadership improves combat effectiveness (i.e. more officers/SNCOs present leads to greater effectiveness, which is the reason that tanks are less effective than gun crews).
Rowland and his team compared the effectiveness of the most effective and the partly effective groups in both the historical battles for which there was information and also for the field trials conducted by the British Army in the 1970s & 1980s. What they found was that there was the same variability within the two groups, which was attributed to opportunities to engage. However there was a significant difference between the groups, which was attributed to heroes being more effective.
Heroism seems to be a product of genetics, social conditioning and values. Many recipients of gallantry awards had previously been mentioned in despatches, or were decorated again.
Comments on citations for subsequent decorations indicate that a second award always required a stronger case than the first award did.
Heroes maintain their combat effectiveness in future battles, even if not further awarded.
Heroism is more likely at higher ranks (i.e. officers and senior NCOs (Sergeants and above) are more likely to be in the higher performing groups than other ranks).
NB there is a possibility that the awarding of decorations was unfairly skewed by rank, and that those of lower rank that performed heroically weren’t adequately recognised.
Gurkha units were noticably different from British unit, and appear to be 60% more effective in inflicting casualties on the enemy and 60% more likely to be decorated. This comes at the price of higher levels of casualties.
The defintion of Surprise is “the achievement of the unexpected in timing, place or direction such that the enemy cannot react properly”. This is distinct from Shock, where soldiers could react, but didn’t.
Again historical analysis was used and battles where surprise and shock were involved were identified. These were then compared with other battles with similar characteristics so that only either Shock or Surprise were different. The two factors being compared individually with a reference set.
Attack surprise reduces infantry defence effectiveness by 60% at 3:1 attack ratio.
Attack surprise may vary with force ratio (being more marked at low ratios and less effective at higher ratios)
Surprise for tank vs tank reduces casualties Â by a factor of 3 at 1:1 attack ratio for the side achieving surprise.
Attacks below 1:1 ratio were successful 65% of the time when surprise was achieved, where attacks at these ratios were never successful without surprise
At force ratios above 1:1 surprise is less important to success, although there is still higher levels of success with surprise, just not statistically significant.
with surprise force ratio is less important to success (at 1:1 70%, at 3:1 76%)
without surprise the probability of success increases in proportion to the force ratio (at 1:1 40%, at 3:1 54%)
Infantry attacks caused shock in about 15% of cases, rising to 50% when combined with surprise and some of the factors below. Three factors were found to have influenced the ability of infantry to inflict shock:
Charge distance was usually under 100 metres (limited by weight of kit), where it was longer that was found to be because the enemy had already broken.
Visibility was significant, typically shock occurs at night or in poor visibilityincluding where the terrain offers concealment
Defence morale was affected by Battle cries, cheers and yells seemed to put defenders off balance.
Bayonets played a major role (but not to cause casualties, as a psychological weapon inducing the enemy to surrender or run away).
Tank attacks caused shock in about 10% of battles analysed.
‘Invulnerable’ tanks cause shock which can lead to panic, in about 50% of cases
Surprise alone caused shock in 27% of the time
Surprise + invulnerable tanks gave 70% Shock
Surprise + poor visibility gave 85% shock
Surprise + all of the above gave 95% shock
Air attacks cause shock most often when they are a dive/strafe attack where the aircraft is aimed directly at the target.
Typically shock by ground attack reduces defence effectiveness by 65%.
Unlike small arms, the effectiveness of weapons used for anti-tank combat has changed considerably over the course of the mid-20th century. From non-specialist gunfire in WW1, to high velocity armour piercing in WW2 and then to Anti-Tank Guided Weapons in the Cold War period. This makes the operational research on anti-tank combat harder to do because the start point needs to be battles where only one kind of AT weapon is in action. Much of the analysis on anti-tank combat starts with the ‘Snipe’ action during the second battle of El Alamein in North Africa where data on each of the guns individually was available.
‘heroic performance’ plays a large factor in the effectiveness of anti-tank guns
about a quarter of guns (at most) performed heroically (including those where platoon, company or battalion level officers assisted with firing guns)
Â rate of fire is proportionate to target availability (i.e. when there are multiple targets crews fire faster)
the median point for heroes was 0.3 tank casualties per gun, where for non-heroes it was 0.03 tank casualties per gun
tanks are less effective in defence than AT guns alone, or tanks supported by AT Guns
AT Guns with tanks apparently kill three times more tanks than the tanks would on their own
AT Gun performance is attributed to having a higher concentration of SNCOs and Officers with deployed ATG compared to tanks (about three times as many)
heroes were disproportionately represented by SNCOs and Officers (at least in terms of who got the medals), in 75% of cases an SNCO or Officer senior to the gun crew commander was involved
Paddy Griffith is quoted on tank casualties that â€œrelatively few appeared to have been caused by enemy tanksâ€
Overall it shows that the biggest single effect in anti-tank combat was down to leadership. Where gun crews are well lead then they are significantly more effective in battle. This is assuming that the guns in question can have some effect on the tanks that they are shooting at, which was the case in all of the battles examined (including a mix where the guns defended successfully with those where the gun lines were overrun by tanks).
This is the third part of my extended review of The Stress of Battle by David Rowland. It is such a strong piece of operational research that I thought that it would be useful for wargame designers (and players) to understand what the research evidence is for what went on in WW2 battles.
Fighting in Woods
The data comes from an analysis of 120 battles that took place in woods or forests from the US Civil War to the Korean War. It also applied all the things from the previous research and tried to see how woods differed from combat in other types of terrain.
For this part I thought that I would focus on the lessons on urban battles. Rowland and his team used historical analysis on lots of WW2 urban battles and then compared this to a series of field trials using laser attachments to small arms and tank main armaments in the late 1970s and early 1980s. Â The approach was toÂ find battles where single variables could be controlled, and then use them to work out what the effect of that variable was on outcomes.
Here’s an interesting table on how attacker casualties vary by odds and the density of defending machine guns. Interestingly, in successful assaults the defender casualties are constant.
The interesting thing for me is that training/experience counts for a lot, halving casualties. Also attacking with the conventional 3:1 odds for success increases the casualties that you suffer, without having any appreciable difference in those inflicted on the enemy (although it does make it more likely for succesful attacksÂ with untrained/inexperienced troops).
Adding armour support makes a huge difference too. Although tanks in urban areas are more vulnerable if they lose their infantry support. However with infantry they significantly reduce attacker casualties.
Defence experience gave no detectable benefit to causing casualties, but attack experience does (in urban combat)
typically three times as many defenders will surrender (some wounded) as are killed or withdraw, the only sensitivity on this is being completely surrounded (so 20% dead, 60% captured (incl wounded) and 20% withdraw);
attack casualties are less affected by force ratio in urban attacks than in open counrtyside;
successful defence of urban areas is best achieved by light defence with counter attacks supported by armour
Rubble & Prepared Defences
This another area covered. There is a general increase in attacker casualties by about 50% when defenders are in rubble or prepared defences. The primary effect of rubble though is to slow down rates of advance.
Rubble halved the rate of advance compared to undamaged urban areas
maximum unopposed advance rates were about 800 metres per hour in urban areas (400m/hr for rubble)
Opposition slowed the advance by a factor of 7
An interesting aside on this was the relative effectiveness of different types of German Infantry. Parachute troops and Panzergrenadiers were reckoned to be tougher opponents than normal infantry. However the analysis showed that the extra stubbornness was a factor of the higher than normal allocation of MGs to those troops. The rate of attacker casualties per defence MG wasn’t significantly different.
Not exactly a book review, more of a synopsis of a great work of Operational Research by David Rowland. The Stress of Battle: Quantifying Human Performance in Combat is the end result of years of work by David Rowland and his team at the Ministry of Defence. Rowland was the father of historical analysis as a branch of Operational Research.
This particular work looks at a combination of field analysis experiments in the 1980s using lasers, well documented WW2 engagements and a handful of battles from other wars. Almost every page in it is packed with evidence or explanations of the complex methodology used to ensure that you could get controlled results from an otherwise messy and chaotic environment. If you are playing or designing wargames then this is one of the books that you absolutely must have on your book shelves (and have read too).
When I was reading the book I was often underlining or marking sections with post-it flags. In particular I drew the following interesting snippets from the book:
Tanks suppress defenders, but you need at least two tanks per defending MG to have any effect;
Combat degradation is about a factor of 10 compared to performance on firing ranges
Anti-tank guns focus the attention of tanks from suppressing MGs, and the bigger the anti-tank gun the more attention it diverts (unsurprisingly);
Fortifications & obstacles (i.e. properly prepared defensive positions) increase defence effectiveness by a factor of 1.65;
In defending against a 3:1 attack, the average rifleman will inflict 0.5 casualties on the attackers whereas a MG will inflict 4 casualties;
1 in 8 riflemen will cause 4 casualties, and the other 7 none;
MG equivalents for casualty causing are: 9 rifles = 1 MG; 1 medium mortar (81mm) = 3 MG;
Combat effectiveness grows with experience, improving the casualty exchange ratio;
This is just a taster of what the book contains. Really worth reading. Not only that it is fantastically well illustrated with loads of graphs, diagrams and pictures from the field exercises to illustrate the points in the text.