[Freeciv-Dev] Re: Artillery and sea units (PR#1476)

[Thanasis Kinias <tkinias@xxxxxxxxxxxxx>]

scripsit Ross W. Wetmore:
> At 04:19 PM 02/05/24 -0700, Thanasis Kinias wrote:
 
> >I adapted Dupuy's QJM to handle similar scale combats over the period
> >AD 1000 to 2000 for another wargame project, and I'm working on how it
> >might be used for Freeciv, if we can get stacking.
> 
> Can you give us a quick summary of its highlights or a reference?

The Dupuy Institute has a Web site, and there is a page which describes
the `Tactical, Numerical, Deterministic Model' [1], which is what they
call it now.  The page is more marketing than description, as they sell
software (MS-DOS Turbo Pascal [!]) -- for $93,000 plus $13,000 per year!

I am working, however, from the Quantified Judgement Model, which was
described in Dupuy's book, _Numbers,_ _Predictions_ _&_ _War_ more than
twenty years ago.

I won't go into all the formulae here, but the basic idea is that you
start with a lethality rating for a weapon, which could be a spear or a
main battle tank.  You sum these for various categories (infantry,
cavalry/armour, artillery, air) and apply modifiers based on weather and
terrain.  For example, jungled swamp multiplies infantry by 0.6 but
tanks by 0.2.  Then you calculate relative mobility for the two forces,
which takes into account both the level of mechanization and the mobile
firepower, giving the more mobile force a multiplier to combat power
(which can be quite large).  (Mobility is also influenced by
environmental variables like terrain and weather.)  `Soft' factors like
morale, training, and logistics get multiplied in, too.  In the end, you
have a simple ratio, which gives you the level of victory and allows
calculation of casualties.

If I can get something useable, I'll post it to a Web server and send a
URI here.

> There are some advantages to a multi-round battle. In Freeciv the 
> firepower is insufficient to knock out any unit in one round, so it
> takes (generally some multiple of five) rounds to resolve. This allows
> for graphics effects, optional strategies like retreats, and generally
> smooths the odds to a more strategic level while not completely ruling
> out the freak chances of battle.
> 
> Exploiting the rounds can introduce interesting effects like switching
> the unit in a stack for the next round or other pre-programmed strategies
> that make battles more interesting and subtle. There is a realistic
> element to this in that battles typically were decided in phases over
> a reasonable period of time, and not with a single blow.

A QJM-like system handles modern combat on a continuum.  In other words,
there aren't discrete rounds.  Until the situation changes (like fresh
troops arrive, or the battle moves into different terrain) you assume
the casualty rates and advance rates to be the same, although the
advance will slow due to exhaustion.  You can thus model a week of
combat with one set of calculations.

In Freeciv terms, it would be unnecessary to have more than one round,
but if more rounds were desired (for any of the reasons you listed), you
could just reduce the casualties per round.  I would implement it with a
`rounds' variable, which would be the number of iterations and a divisor
for casualties in each iteration.

1. <http://www.dupuyinstitute.org/tndm.htm>

-- 
Thanasis Kinias
Web Developer, Information Technology
Graduate Student, Department of History
Arizona State University
Tempe, Arizona, U.S.A.

Ash nazg durbatulûk, ash nazg gimbatul,
Ash nazg thrakatulûk agh burzum-ishi krimpatul