$30.00
Recommend
14 
 Thumb up
 Hide
48 Posts
1 , 2  Next »   | 

Wargames» Forums » General

Subject: How Designer's Calculate Combat Values in Wargames rss

Your Tags: Add tags
Popular Tags: [View All]
Cameron Taylor
New Zealand
Auckland
flag msg tools
Die Ostfront, 1941–1945
badge
So serious...
Avatar
mbmbmbmbmb
I've been designing my own wargame as a bit of a side project. When researching the Order of Battle (OOB) and Table of Organisation and Equipment (TOE), I discovered that some formations had wildly different arms than others, giving them drastically different combat power using the Quantified Judgement Method (QJM). The same types of formations (e.g. Infantry divisions) may have wildly different support assets (e.g. artillery pieces, tanks, engineers) attached to them. (You can calculate precise combat power using QJM and its rigorous quantitative–based analysis, but it requires a complete OOB and TOE, e.g. # of each weapons system.)

Yet when I look at various wargames (e.g. OCS The Blitzkrieg Legend, France '40, Next War: India–Pakistan) the same types of divisions are uniform (e.g. All Infantry Divisions are Strength 4, whilst all Panzer Divisions are Strength 8). The only real differentiation is in the Action Rating (AR) / Troop Quality (TQ) / Efficiency Rating (ER) etc., which may affect surprise and gives die roll modifiers (DRM).

That's when it struck me: I know how board wargames calculate combat power. They either:

(a) Take the support assets of higher echelons and spread them out across formations, so as to make the
(b) As per The Complete Wargames Handbook by James Dunnigan, units are arbitrarily assigned values and through playtesting the values are tweaked to give roughly historical results (i.e. sensitivity analysis and case studies).

Truth be told, wargame designers probably use a combination of the two approaches. For example, in the Operational Combat Series (OCS) battalions are calculated as being uniform according to type (e.g. medium infantry battalions have identical strength) within the same nationality, whilst the multiplier for the nationality is calculated through sensitivity analysis. (The exception are artillery units, which have their strength calculated by the number and size of tubes, with the multiplier also being calculated through sensitivity analysis.) The only difference is in the action rating. The same applies to the Next War series.

Now I understand why it's never explained how designers arrive at particular values. It's a real sausage factory. The realism of any wargame is up to the subjective talent of the designer in producing plausible outcomes, not some quantitative–based analysis. Honestly, I feel a little deflated. It just seems so arbitrary.
8 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Roger Hobden
Canada
Montreal
Quebec
flag msg tools
Avatar

CRT's seem to be another instance of design for effect.

3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mike Hoyt

Durango
Colorado
msg tools
Avatar
mbmbmbmbmb
SeriousCat wrote:
(You can calculate precise combat power using QJM and its rigorous quantitative–based analysis, but it requires a complete OOB and TOE, e.g. # of each weapons system.)


You are seriously deluding yourself.

To give just one simple example, mentioned in the Force on Force rules but generally applicable

30 untrained milita members armed with modern assault rifles are no match for 30 trained men in an organized platoon, even if armed with identical rifles.

Combat power is hardly just the sum of the OOB and TOE. Training, Morale, Supply State, the list of additional factors goes on and on.
18 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Cameron Taylor
New Zealand
Auckland
flag msg tools
Die Ostfront, 1941–1945
badge
So serious...
Avatar
mbmbmbmbmb
For example, the OOB and TOE of the FRG 3rd Panzer Division is wildly different from 1e Divisie within the same I Netherlands Corps of NORTHAG in Northern West Germany, 1989. The methodology for calculating combat power is from the book Numbers, Predictions, and War by Col. Trevor Dupuy, the OLI values are from the monograph Correlation of Forces by Maj. David Hoggs, whilst the OOB and TOE are from the page NORTHAG wartime structure in 1989 on Wikipedia. Where the exact variant of a weapons system was not found, the closest analogue was used instead (e.g. M109A3 Self–Propelled Howitzer substituted for M109A3G Self–Propelled Howitzer).

Format:
(Formation)
(Weapons System), (Number of Systems Present) x (Individual OLI) = (Total OLI)

3rd Panzer Division, I Netherlands Corps, NORTHAG

7th Panzergrenadier Brigade, Hamburg
Staff Company, 1st Panzergrenadier Brigade, Hamburg
Luchs, 8 x 70 = 560
M577, 8 x 2.69 = 22
Total = 582

71st Panzergrenadier Battalion, Hamburg
Leopard 1A5, 13 x 720 = 9,360
Marder, 24 x 194 = 4,656
M113, 12 x 2.69 = 32
Total = 14,048

72nd Panzergrenadier Battalion, Hamburg
Marder, 24 x 194 = 4,656
M113, 23 x 2.69 = 62
Panzermörser, 6 x 97 = 582
Total = 5,300

73rd Panzergrenadier Battalion, Cuxhaven
Marder, 24 x 194 = 4,656
M113, 23 x 2.69 = 62
Panzermörser, 6 x 97 = 582
Total = 5,300

74th Panzer Battalion, Cuxhaven
Leopard 1A5, 41 x 720 = 29,520
M113, 12 x 2.69 = 32
Total = 29,552

75th Panzer Artillery Battalion, Hamburg
M109A3G, 18 x 223 = 4,014
Total = 4,014

70th Anti-Tank Company, Cuxhaven
Jaguar 2, 12 x 262 = 3,144
Total = 3,144

8th Panzer Brigade, Lüneburg
Staff Company, 8th Panzer Brigade, Lüneburg
Luchs, 8 x 70 = 560
M577, 8 x 2.69 = 22
Total = 582

81st Panzer Battalion, Lüneburg
Leopard 2A4, 28 x 1177 = 32,956
M113, 12 x 2.69 = 32
Marder, 11 x 194 = 2,134
Total = 35,122

82nd Panzergrenadier Battalion, Lüneburg
Marder, 35 x 194 = 6,790
M113, 12 x 2.69 = 32
Panzermörser, 6 x 97 = 582
Total = 7,404

83rd Panzer Battalion, Lüneburg
Leopard 2A4, 41 x 1177 = 48,257
M113, 12 x 2.69 = 32
Total = 48,289

84th Panzer Battalion, Lüneburg
Leopard 2A4, 41 x 1177 = 48,257
M113, 12 x 2.69 = 32
Total 48,289

85th Panzer Artillery Battalion, Lüneburg
M109A3G, 18 x 223 = 4,014
Total = 4,014

80th Anti-Tank Company, Lüneburg
Jaguar 1, 12 x 262 = 3,144
Total = 3,144

9th Panzer (Lehr) Brigade, Munster
Staff Company, 9th Panzerlehrbrigade, Munster
Luchs, 8 x 70 = 560
M577, 8 x 2.69 = 22
Total = 582

91st Panzer (Lehr) Battalion, Munster
Leopard 2A4, 28 x 1177 = 32,956
M113, 12 x 2.69 = 32
Marder, 11 x 194 = 2,134
Total = 35,122

92nd Panzergrenadier (Lehr) Battalion, Munster
Marder, 35 x 194 = 6,790
M113, 12 x 2.69 = 32
Panzermörser, 6 x 97 = 582
Total = 7,404

93rd Panzer (Lehr) Battalion, Munster
Leopard 2A4, 41 x 1177 = 48,257
M113, 12 x 2.69 = 32
Total = 48,289

94th Panzer (Lehr) Battalion, Munster
Leopard 2A4, 41 x 1177 = 48,257
M113, 12 x 2.69 = 32
Total = 48,289

95th Panzer (Lehr) Artillery Battalion, Munster
M109A3G, 18 x 223 = 4,014
Total = 4,014

90th Anti-Tank Training Company, Munster
Jaguar 1, 12 x 262 = 3,144
Total = 3,144

3rd Artillery Regiment, Stade
31st Field Artillery Battalion, Lüneburg
FH-70, 18 x 223 = 4,014
M110A2, 18 x 173 = 3,114
Total = 7,128

32nd Rocket Artillery Battalion, Dörverden
LARS, 18 x 311 = 5,598
MLRS, 16 x 311 = 4,976
Total = 10,574

3rd Armored Reconnaissance Battalion, Lüneburg
Fuchs, 18 x 61 = 1,098
Leopard 1V, 34 x 720 = 24,480
Luchs, 10 x 70 = 700
Total = 26,278

3rd Air Defense Regiment, Hamburg
Gepard, 36 x 280 = 10,080
Total = 10,080

1e Divisie, I Netherlands Corps, NORTHAG

102nd Reconnaissance Battalion "Huzaren van Boreel", Hoogland
Leopard 2A4, 18 x 1177 = 21,186
M113-Command & Reconnaissance, 48 x 0 = 0

11e Pantserinfanteriebrigade, Arnhem, NL
101st Pantser Battalion "Regiment Huzaren Prins Alexander", Soesterberg
Leopard 1V, 61 x 720 = 43,920
YPR-765, 12 x 114 = 1,368
Total = 45,288

12th Pantserinfanterie Battalion "Garde Regiment Jagers", Arnhem
YPR-765, 70 x 114 = 7,980
YPR-765 PRAT, 16 x 221 = 3,536
Total = 11,516

48th Pantserinfanterie Battalion "Regiment van Heutsz", 's-Hertogenbosch
YPR-765, 70 x 114 = 7,980
YPR-765 PRAT, 16 x 221 = 3,536
Total = 11,516

11th Horse Artillery Battalion "Gele Rijders", Arnhem
M109A3, 20 x 202 = 4,040
Total = 4,040

11th Armored Anti-Tank Company, Ermelo
YPR-765 PRAT, 1 x 114 = 114
Total = 114

12e Pantserinfanteriebrigade, Vierhouten, NL
59th Pantser Battalion "Regiment Huzaren Prins Oranje", 't Harde
Leopard 1V, 61 x 720 = 43,920
YPR-765, 12 x 114 = 1,368
Total = 45,288

11th Pantserinfanterie Battalion "Garde Regiment Grenadiers", Arnhem
YPR-765, 70 x 114 = 7,980
YPR-765 PRAT, 16 x 221 = 3,536
Total = 11,516

13th Pantserinfanterie Battalion "Garde Fusiliers Princess Irene", Schalkhaar
YPR-765, 70 x 114 = 7,980
YPR-765 PRAT, 16 x 221 = 3,536
Total = 11,516

14th Field Artillery Battalion (Reserve), Vierhouten
M109A3, 20 x 223 = 4,460
Total = 4,460

12th Armored Anti-Tank Company, Vierhouten
YPR-765 PRAT, 1 x 221 = 221
Total = 221

13e Pantserbrigade, Oirschot, NL
11th Pantser Battalion "Huzaren van Sytzama", Oirschot
Leopard 1V, 52 x 720 = 37,440
YPR-765, 12 x 114 = 1,368
Total = 38,808

49th Pantser Battalion (Reserve) "Huzaren van Sytzama", Oirschot
Leopard 1V, 52 x 720 = 37,440
YPR-765, 12 x 114 = 1,368
Total = 38,808

17th Pantserinfanterie Battalion "Regiment Infanterie Chasse", Oirschot
YPR-765, 70 x 114 = 7,980
YPR-765 PRAT, 16 x 221 = 3,536
Total = 11,516

12th Field Artillery Battalion, Oirschot
M109A3, 20 x 223 = 4,460
Total = 4,460


I won't bother listing out the OOB and TOE for 4e Divisie, I Netherlands Corps, but the total OLI is given below as well.

Total (3rd Panzer Division) = 409,689 OLI
Total (1e (Mechanised) Divisie) = 93,592 OLI
Total (4e (Mechanised) Divisie) = 427,070 OLI

In the wargame NATO: Operational Combat in Europe in the 1970's by SPI (admittedly ~15 years apart from 1989), I anticipate the only real difference was the upgraded weapons systems (e.g. moving from the Leopard 1 to Leopard 2). The differences in OLIs shouldn't be too pronounced between the periods. (There was a remake with an updated OOB called Group of Soviet Forces Germany by Decision Games.) 1e Divisie and 4e Divisie are rated identically, despite having vastly different equipment, whilst 1e Divisie almost has parity with 3rd Panzer.

Counter Values (3rd Panzer Division) = 6–6–8
Counter Values (1e (Mechanised) Division) = 5–5–8
Counter Values (4e (Mechanised) Division) = 5–5–8

Sources:
Dupuy, T.N. (1979). Numbers, Predictions, and War. Indianapolis, New York: Bobs–Merill.
Hogg, D.R. (1993). Correlation of Forces: The Quest for a Standardized Model. Fort Leavenworth, Kansas: School of Advanced Military Studies, US Army Command and General Staff College. pp. Appendix B 1–28
Wikipedia. (n.d.) NORTHAG wartime structure in 1989. Retrieved from https://en.wikipedia.org/wiki/NORTHAG_wartime_structure_in_1...
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Cameron Taylor
New Zealand
Auckland
flag msg tools
Die Ostfront, 1941–1945
badge
So serious...
Avatar
mbmbmbmbmb
blockhead wrote:
SeriousCat wrote:
(You can calculate precise combat power using QJM and its rigorous quantitative–based analysis, but it requires a complete OOB and TOE, e.g. # of each weapons system.)


You are seriously deluding yourself.

To give just one simple example, mentioned in the Force on Force rules but generally applicable

30 untrained milita members armed with modern assault rifles are no match for 30 trained men in an organized platoon, even if armed with identical rifles.

Combat power is hardly just the sum of the OOB and TOE. Training, Morale, Supply State, the list of additional factors goes on and on.


I'm afraid you're just misinformed. QJM takes into account training, morale, terrain factors, weather, etc. Central to the QJM combat power formula is the concept of relative Combat Effectiveness Value (CEV), which is calculated as casualty causing potential. (You can even calculate it using lower level tactical wargames.)

It can predict over 95% of combat engagements outside the combat ratio of 0.9 to 1 : 1.1. I'm not saying the methodology is perfect, for example it doesn't take into account combined arms effects, but ultimately any model is only validated by its predictive power. 95% accuracy is hard to argue against, no matter how you feel about it.
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Wendell
United States
Alexandria (temporarily)
Virginia
flag msg tools
All the little chicks with crimson lips, go...
badge
Hey, get your stinking cursor off my face! I got nukes, you know.
Avatar
mbmbmbmbmb
It's pretty easy so long as you remember German armored units in Third Reich are worth 4.
17 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
M St
Australia
Unspecified
flag msg tools
designer
SeriousCat wrote:
Central to the QJM combat power formula is the concept of relative Combat Effectiveness Value (CEV), which is calculated as casualty causing potential. (You can even calculate it using lower level tactical wargames.)

Which means that 95% of the fudging goes into judging the CEV value. You are simply shifting the variability to another level by hiding it in a particular variable.
19 
 Thumb up
1.30
 tip
 Hide
  • [+] Dice rolls
M St
Australia
Unspecified
flag msg tools
designer
SeriousCat wrote:

That's when it struck me: I know how board wargames calculate combat power.

No, I think you know how some do it.

Quote:

Now I understand why it's never explained how designers arrive at particular values. It's a real sausage factory. The realism of any wargame is up to the subjective talent of the designer in producing plausible outcomes, not some quantitative–based analysis. Honestly, I feel a little deflated. It just seems so arbitrary.

Stated like this, it's also wrong.

First, it's not the case that it's never explained. There are examples where it has been explained. Not many, but they exist.

Second, the fact that it's not explained doesn't mean it's completely arbitrary. In fact the claim that because it's by judgment means it's arbitrary is a contradiction in terms. It's not arbitrary exactly because it's arrived at by judgment. It's just not the simplistic style of counting rifles that the uninitiated may think is behind it.

Third, that it's not explained doesn't even mean it's not based on quantitative analysis - it's entirely possible that particular designers treat the analytical steps as a sort of trade secret. It doesn't mean it's based purely on quantitative analysis because that would be an unrealistic expectation.

Last, if you think that any form of quantitative analysis does not require an enormous amount of judgment before the numbers are crunched, in any field, you'd be surprised what it is actually like.

Now, as regards this comment:

Quote:
The realism of any wargame is up to the subjective talent of the designer

That is 100% true of any wargame far beyond the OOB and unit ratings. Minor OOB differences generally pale compared to the impact of rules such as ZOCs, sequence of play, and whatever else is in the game. And there is no way in which these factors can be turned into rules solely by "quantitative analysis". As has been often pointed out, wargames are holistic, and a careless rule can wreck a design. That's why the designer is the single most important factor in design quality. There is no wargame design by numbers.
26 
 Thumb up
0.02
 tip
 Hide
  • [+] Dice rolls
John Middleton
United States
Laramie
Wyoming
flag msg tools
badge
Avatar
mbmbmbmbmb
I've actually found that strict order of battle info is one of the least important parts of a games design.


Granted, the OOB has to be accurate and appropriate to the game. For example, most Bulge games have similar OOBs, that definitely improve over time as more research becomes available, but the thing that makes one Bulge game different from another is the interaction of ZoC, armor exploitation, artillery, weather, air cover, etc.

It isn't which rifles each division had or how the companies were lettered.

Of course games with strict command control and unit coherence rules make every bit of info on a counter critical, most games don't seem to bother.
7 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Darrell Pavitt
United Kingdom
Unspecified
Unspecified
flag msg tools
Avatar
mbmbmbmbmb
Frank Chadwick gives you a whole playbook worth of explanation as to how the values were calculated for The Third World War
7 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Pelle Nilsson
Sweden
Linköping
flag msg tools
designer
badge
Avatar
mbmbmbmbmb
Have a look in Biddle's Military Power. He makes some really good arguments about how the technology and even size of units are not very important numbers in the calculations, compared to training and deployment. Even if you do not fully agree with him, I think all the fuzzy stuff that obviously goes into deciding how strong a unit is makes it extremely difficult to make a good comparison, so nothing wrong with just saying they are all a 4 and hide it in all the other noise from things that can not be perfectly calculated.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Etien
United States
Thibodaux
Louisiana
flag msg tools
General
badge
The War Room
Avatar
mbmbmbmbmb
blockhead wrote:
SeriousCat wrote:
(You can calculate precise combat power using QJM and its rigorous quantitative–based analysis, but it requires a complete OOB and TOE, e.g. # of each weapons system.)


You are seriously deluding yourself.

To give just one simple example, mentioned in the Force on Force rules but generally applicable

30 untrained milita members armed with modern assault rifles are no match for 30 trained men in an organized platoon, even if armed with identical rifles.

Combat power is hardly just the sum of the OOB and TOE. Training, Morale, Supply State, the list of additional factors goes on and on.


Well unless the militia group is cloaked in civilian garb hiding within a domestic population and protected by the rules of engagement. (OIF II and III)
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Holman
United States
Philadelphia
Pennsylvania
flag msg tools
badge
Avatar
mbmbmbmbmb
I remember this discussion WRT to a computer game, possibly an early version of Gary Grigsby's The Operational Art of War.

The database allowed the creation of new units, and strengths were calculated by combining component weapon systems. The numbers were granular enough that a force of 24 Sherman M4A3's would actually rate slightly differently from a force of 24 Sherman M4A1's. (In practice it would make no difference, but the numbers were that specific.)

People played around and quickly learned the limits of plausibility. For instance, it was learned either that a force of 50 jeeps *would* or *would not* defeat a single Tiger tank. I can't remember which it was, but it doesn't matter, because the important question was what these numbers were actually modeling.

After all, if one jeep can't affect a Tiger in normal combat then 50 can do no better. On the other hand, it's completely plausible that a force of 50 jeeps and their drivers (armed with pistols and a couple of grenades each) almost certainly could end a Tiger's career if it had no support.

I forget how the long design debate ended, but at the time it was both funny and seemed desperately important.
13 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Paul Kreutzet
United States
flag msg tools
mbmbmbmb
I have an alternative take on combat values for a moderate-complexity ancients tactical game: not having any. That is to say, weighing other factors more significantly, based on a few assumptions. The first assumption is that in an ancient battle line, the numbers of individual fighters clashing along a front line that we simulate with units and spaces will be about the same. If a space has a 50 meter front then about equal numbers of ancient era troops will be lined up opposing each other along that distance that we simulate with a unit; opposing troops will equally optimize their coverage of a given frontage.

So how to develop a combat factor, and what are significant variances to account for? First, units would be rated for morale, a rating of their overall status, equipment, readiness, responsiveness, professionalism, and prestige/training within a given army. Second, unit type has to be represented and this affects combat calculations and results- light infantry, heavy, cav, archers, etc. Additionally and to account for different unit densities, incorporate unit step levels (which is somewhat of a quantity-based combat factor). That, and dice rolls to provide randomization of combat results.

Anyway, I offer this up as thought and playtesting experiment.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Cameron Taylor
New Zealand
Auckland
flag msg tools
Die Ostfront, 1941–1945
badge
So serious...
Avatar
mbmbmbmbmb
M St wrote:
Third, that it's not explained doesn't even mean it's not based on quantitative analysis - it's entirely possible that particular designers treat the analytical steps as a sort of trade secret. It doesn't mean it's based purely on quantitative analysis because that would be an unrealistic expectation.


So you're syaing, "Trust me, it's accurate, just don't look at my figures." Umm... that's real shady. If a definite quantitative value has been assigned to something (e.g. Combat Strength, Movement Allowance, etc.), I'd like to know where it's from. Call me pedantic, but I like to know the reasons behind how things work.

M St wrote:
Which means that 95% of the fudging goes into judging the CEV value. You are simply shifting the variability to another level by hiding it in a particular variable.


Again, you are misinformed. If there were 'fudging', the CEV wouldn't be a precise calculable figure based on a purely quantitative approach. (The P Ratio is the prediction, whilst the R Ratio are the actual results. You can precisely calculate CEV after a battle using the P and R ratios, or you can predict CEV before a battle by using the formulae below.) For reference, the CEV is calculated as follows:

Ka = ( CEVa² x Kd ) / u
Kd = ( u x Ka ) / CEVa²

Where:
a = Attacker Identifier
d = Defender Identifier
CEV = Relative Combat Effectiveness Value
K = Casualties Inflicted per Day
u = Posture factor

Furthermore, without simulations determining the K value, you can approximate K using this alternative process:

CEVa = SQRT( La / Ld )

La = K / ( u x r x h x z x SQRT(sz) )

Where:
L = Lethality
a = Attacker Identifier
d = Defender Identifier
K = Historical Casualty Inflicting Rate (Lookup Table)
u = Posture Factor (Lookup Table)
r = Terrain (Lookup Table)
h = Weather (Lookup Table)
z = Season (Lookup Table)
sz = Size (Lookup Table)

If both CEVs are recipricols, use as is
If CEVa > 1 and CEVb < 1: Use "( CEVa + (1 / CEVb) ) / 2", with CEVb as recipricol of CEVa
If both CEVa and CEVb > 0: Divide the larger CEV by the smaller CEV

There are legitimate methodological issues with QJM, which I'm happy to discuss. For example, its successor TNDM involves a precise logistics calculation, whereas QJM only includes a general logistics calculation which doesn't involve any source data. The biggest issue is that at the division and battalion level (where QJM is based) combined arms are assumed to be in effect, instead of being calculated independently. But criticising something out of hand without knowing anything about it is unhelpful.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Cameron Taylor
New Zealand
Auckland
flag msg tools
Die Ostfront, 1941–1945
badge
So serious...
Avatar
mbmbmbmbmb
DegenerateElite wrote:
I've actually found that strict order of battle info is one of the least important parts of a games design... Granted, the OOB has to be accurate and appropriate to the game. For example, most Bulge games have similar OOBs, that definitely improve over time as more research becomes available, but the thing that makes one Bulge game different from another is the interaction of ZoC, armor exploitation, artillery, weather, air cover, etc.

It isn't which rifles each division had or how the companies were lettered... Of course games with strict command control and unit coherence rules make every bit of info on a counter critical, most games don't seem to bother.


Whilst OOB, TOE, and troop quality are crucial, I absolutely agree that C2 is definitely one of the weakests parts of wargames. The exception for board wargames seems to be the Tactical Combat Series, which features planning and restrictions on combat. For computer wargames the exceptions seem to be Command Ops series, which features command delay and command confusion as one goes down to lower level echelons, and Flashpoint Campaigns series, which features simultaneous movement, C2, C2 disruption, and electronic warfare.

pelni wrote:
Have a look in Biddle's Military Power. He makes some really good arguments about how the technology and even size of units are not very important numbers in the calculations, compared to training and deployment. Even if you do not fully agree with him, I think all the fuzzy stuff that obviously goes into deciding how strong a unit is makes it extremely difficult to make a good comparison, so nothing wrong with just saying they are all a 4 and hide it in all the other noise from things that can not be perfectly calculated.


Biddle's research is very interesting, but his methodology is significantly flawed. Environmental and tactical factors are assumed to be ordinal (e.g. rated on a 1–5 scale), so tactical surprise may have the same effect as air superiority, regardless of how much of an advantage each has. Most egregiously, Biddle doesn't consider force multipliers in his force ratio projections, instead using raw figures for numbers of types of weapons systems (e.g. REDFOR has 1,000 artillery pieces vs. BLUEFOR 1,200 artillery pieces).

Moreover, what's frquently lost in the discussion is that even Biddle writes in Military Power that this research is exploratory, not confirmatory, so whilst generalisations can be drawn from the research it has no predictive power. Biddle's research is not completely without value. His analysis on defensive depth and offensive containment is very enlightening and perhaps is the first (at least publically available) method of quantitatively explaining defence in depth.

PaulWRoberts wrote:
I remember this discussion WRT to a computer game, possibly an early version of Gary Grigsby's The Operational Art of War.

The database allowed the creation of new units, and strengths were calculated by combining component weapon systems. The numbers were granular enough that a force of 24 Sherman M4A3's would actually rate slightly differently from a force of 24 Sherman M4A1's. (In practice it would make no difference, but the numbers were that specific.)

People played around and quickly learned the limits of plausibility. For instance, it was learned either that a force of 50 jeeps *would* or *would not* defeat a single Tiger tank. I can't remember which it was, but it doesn't matter, because the important question was what these numbers were actually modeling.

After all, if one jeep can't affect a Tiger in normal combat then 50 can do no better. On the other hand, it's completely plausible that a force of 50 jeeps and their drivers (armed with pistols and a couple of grenades each) almost certainly could end a Tiger's career if it had no support.


The game you're referring to is Norm Kroger's The Operational Art of War I. (Gary Grigsby is the one behind Steel Panthers, which has impeccable combat modelling.) It was a bug in improper combat modelling. This has been fixed in TOAW II. We are now up to TOAW III and IV is currently under development. Now if one weapons system can't penetrate or even damage another, no matter how many of them you have you still won't do any damage.
4 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Cameron Taylor
New Zealand
Auckland
flag msg tools
Die Ostfront, 1941–1945
badge
So serious...
Avatar
mbmbmbmbmb
nyhotep wrote:
Frank Chadwick gives you a whole playbook worth of explanation as to how the values were calculated for The Third World War


Thank you for pointing me towards that. I'll have a look in the PDFs now.

EDIT:


It seems that in the Third World War series (The Third World War: Briefing Booklet, pp.4–7) the combat values are arbitrarily determined as well. They have no source data for their strength values. For example, an MBT 120mm has a strength of 0.15, though how this number is arrived at isn't explained. Even though the numbers are arbitrary, his reasoning makes logical sense, such as having higher defense strength for infantry with ATGM or attack helicopters being 6 times as powerful as MBTs 105mm.This leads me to assuming that the designer used the method described by James Dunnigan. The research done for TTWW is clearly substantial, but it is qualitative in nature (e.g. reading doctrinal manuals, policy anaylsis briefings, military theory) instead of quantiative.

One approach in wargames I particularly like is John Tiller's Panzer Campaigns and Modern Campaigns series, which splits attack into Soft, Hard, and AA. For each type of attack there are two variables, the first for a penetrating attack (i.e. it's strong enough to overcome the defensive strength of the target) and the second for a non–penetrating attack. The simple difference this makes in games like Panzer Campaigns: France '40 is dramatic, because the legions of Panzer I and II tanks are highly effective against infantry, but do little damage to Allied medium armour.

ftstevens wrote:
I have an alternative take on combat values for a moderate-complexity ancients tactical game: not having any. That is to say, weighing other factors more significantly, based on a few assumptions. The first assumption is that in an ancient battle line, the numbers of individual fighters clashing along a front line that we simulate with units and spaces will be about the same. If a space has a 50 meter front then about equal numbers of ancient era troops will be lined up opposing each other along that distance that we simulate with a unit; opposing troops will equally optimize their coverage of a given frontage.

So how to develop a combat factor, and what are significant variances to account for? First, units would be rated for morale, a rating of their overall status, equipment, readiness, responsiveness, professionalism, and prestige/training within a given army. Second, unit type has to be represented and this affects combat calculations and results- light infantry, heavy, cav, archers, etc. Additionally and to account for different unit densities, incorporate unit step levels (which is somewhat of a quantity-based combat factor). That, and dice rolls to provide randomization of combat results.

Anyway, I offer this up as thought and playtesting experiment.


Now that is truly radical.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
roger miller
United States
California
flag msg tools
publisher
Weapons and numbers give a rough idea of a units potential combat strength but not much else. As Marcus said it is how all the rules interact with the units that make it a model worth anything. I prefer to set combat values within the model. Meaning you should be able to set up historical engagements and get a range of results that make sense when compared to history. Then broaden it out to multiple combats and then the whole battle. If your strengths hold up to produce good results at both micro and macro level you have good strength numbers.
6 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
John Middleton
United States
Laramie
Wyoming
flag msg tools
badge
Avatar
mbmbmbmbmb
As I recall, there was lots of interest in Dupuy's book and calculations when it came out years ago, but most designers and players felt it made pretty dull games.


The numbers have to be adjusted for all the other aspects of the games design and his formula does not reflect this.


So, unfortunately, or fortunately on POV, you have to make guesses based on experience and playtesting.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Cameron Taylor
New Zealand
Auckland
flag msg tools
Die Ostfront, 1941–1945
badge
So serious...
Avatar
mbmbmbmbmb
rmiller1093 wrote:
Weapons and numbers give a rough idea of a units potential combat strength but not much else. As Marcus said it is how all the rules interact with the units that make it a model worth anything. I prefer to set combat values within the model. Meaning you should be able to set up historical engagements and get a range of results that make sense when compared to history. Then broaden it out to multiple combats and then the whole battle. If your strengths hold up to produce good results at both micro and macro level you have good strength numbers.


I guess as long as you're consistent, even if the actual values are arbitrary and the relationships not entirely accurate principles of warfare can be represented in spirit (e.g. tanks fair poorly in rough terrain).
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Lawrence Hung
Hong Kong
Wan Chai
Hong Kong
flag msg tools
badge
Avatar
mbmbmbmbmb
Number crunching is important. What makes wargames distinguishing from other boardgames are that they are based on hard numbers as a science. The art of designing the system and sub-systems by rules is also important of course. And that what makes wargaming a fun exercise than otherwise, especially commercial wargaming. The Army doesn't believe wargaming should be fun and so they rely on hard data simulations which are nothing but dull boring exercises of drill nature. Flashpoint Campaigns: Red Storm is really a cool wargame combining both. Don't miss this wargame.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jur dj
Netherlands
Leiden
flag msg tools
Avatar
mbmbmb
I remember first being very taken by the QJM but it has some inherent assumptions which in practice limit its use. The most important one is that it that it takes inflicting casualties as the only way to measure combat effectiveness.

Thus, while the QJM might prove a reasonable predictor of inflicted casualties in combat, it has very little to say about the rest of combat. It cannot really predict when troops break or retire. It cannot predict tactical or operational outcomes.

If game designers would work within this limitation, that would be fine, but using combat factors only based on QJM, you reduce combat to the killing bit and your overall model becomes flawed to some degree.

Next problem is that the CEV is the residual of the other factors that presumably can be measured. That makes it weak as an analytical tool and is very hard to apply as a game designer. Because whereas you can derive CEV from historical examples, you cannot predict them independently, or you end up with using the CEVs that Dupuy calculated.

Consider the case in 1941 where Germans Pz 38t clash with an equal number of BT7, tanks which are theoretically about equal in technical performance. The difference in combat performance is huge because of the training, morale, doctrine, logistic support and command & communications. All very important factors that because they can't be measured independently, are lumped into one residual factor.

The multipliers in the model are as arbitrary as any, because you have no real way of determining how effective one weapon system is at killing as opposed to the other.

So while QJM poses as hard science, may of it's assumptions are based on weak foundations or fitting the model to the outcome.

Interest for Dupuy's method seems to have faded in the 1990s, after his predictions for the Gulf War proved quite far off and interest in manoeuvre warfare became more prominent.

Lawrence Hung wrote:
Number crunching is important. What makes wargames distinguishing from other boardgames are that they are based on hard numbers as a science. The art of designing the system and sub-systems by rules is also important of course. And that what makes wargaming a fun exercise than otherwise, especially commercial wargaming. The Army doesn't believe wargaming should be fun and so they rely on hard data simulations which are nothing but dull boring exercises of drill nature. Flashpoint Campaigns: Red Storm is really a cool wargame combining both. Don't miss this wargame.
7 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
John Middleton
United States
Laramie
Wyoming
flag msg tools
badge
Avatar
mbmbmbmbmb
To be fair, almost all the "experts" estimates on the Gulf War were wrong.


Especially with regard to how long it would take.




1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Cameron Taylor
New Zealand
Auckland
flag msg tools
Die Ostfront, 1941–1945
badge
So serious...
Avatar
mbmbmbmbmb
jurdj wrote:
I remember first being very taken by the QJM but it has some inherent assumptions which in practice limit its use. The most important one is that it that it takes inflicting casualties as the only way to measure combat effectiveness.

Thus, while the QJM might prove a reasonable predictor of inflicted casualties in combat, it has very little to say about the rest of combat. It cannot really predict when troops break or retire. It cannot predict tactical or operational outcomes.


That is a legitimate shortcoming of QJM.

jurdj wrote:
Consider the case in 1941 where Germans Pz 38t clash with an equal number of BT7, tanks which are theoretically about equal in technical performance. The difference in combat performance is huge because of the training, morale, doctrine, logistic support and command & communications. All very important factors that because they can't be measured independently, are lumped into one residual factor.


By the central limit theorem, aggregation can reduce average variance. Whilst it's possible to go into ever more detail, there's no guarantee it will be more accurate. In fact, overmodelling of particular variables may give undue influence to certain factors, causing more inaccuracy, not less.

jurdj wrote:
The multipliers in the model are as arbitrary as any, because you have no real way of determining how effective one weapon system is at killing as opposed to the other.

So while QJM poses as hard science, may of it's assumptions are based on weak foundations or fitting the model to the outcome.


This is a common misconception. The multipliers are gained through multiple regression, a proven and powerful statistical technique for uncovering strong correlations, not just arrived at by subjective sensitivity analysis. Granted, the values are only as good as the database, which is why QJM (or more correctly QJMA) is actually a research model for applying statistical analysis to a database of engagements. It's not static.

jurdj wrote:
Interest for Dupuy's method seems to have faded in the 1990s, after his predictions for the Gulf War proved quite far off and interest in manoeuvre warfare became more prominent.


You mean how he predicted the strategies for both sides and had the closest casualty figure estimates in the world? Dupuy got more right than wrong. The estimates were held in high regard by the military even after the fact, precisely because they were so accurate. It's in public memory through newspapers and laymen that the model wasn't 100% accurate and therefore entirely wrong.

I actually have both books by Trevor Dupuy published before the Gulf War, If War Comes... How to Defeat Saddam Hussein and Attrition. The predictions contained therein came very close to the truth.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
roger miller
United States
California
flag msg tools
publisher
Cameron have you ever tested QJM? When I first saw it I thought wow this is great but I was skeptical of the 95% predictive claim. So I set up a bunch of tests of researching historical WW2 battles where the weapons, numbers, terrain, etc. were as near to identical as possible. Now the model says that identical situations should produce nearly identical casualties but the model failed completely. The Russian front in 1941 especially was a total mess. Identical Russian units with to all outside views had the same equipment, training, etc. when hit by identical German units produced results all over the place. Some units held up well and other collapsed completely. The desert in North Africa also produced results all over the place. Try modeling the Italian armies performance with this model and it fails real quick.

Now Dupuy based his original model on a years worth of combat in NW Europe and within that limited scope it produced decent results but not 95%. They got to 95% by cooking the books. If a combat fell outside the models results they simply adjusted the values for that units morale, training, etc. Clearly these values had to be off since the model had failed. My view is that it was the model that was off and it completely underestimates the variability in outcomes possible in combat.
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
1 , 2  Next »   | 
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.