random vs deterministic model!

[JMM]

The system is deterministic but it seems it is not because there is a lot of parameters you can't know...

For 1 vs 1, without other units, you always get the same result.. euh.. that said, there is a first parameter modifying this result.. the time scale.
The geometry of unit position is important, so changing the time scale can modify the distance between the units, so the result of casualties.

BTW, in a real battle, it's not 1 vs 1, but 100 vs 100 with a lot of interactions.. each calculus is deterministic, but you can't predict the result because you don't know the events modifying the calculus...

I find it's not the best to have an elite regiment losing a fight against a regiment of line with a dice... because the elite always wins except if the external conditions infers a problem about its behaviour.

I hope it's clear.

JMM

[Hook]

In statistics, we use "standard deviation" to describe the variation in expected results from the extremes at each end of the "bell curve". I think it is vital that the game take the same approach - i.e. in the example above the average may be 23 casualties but on rare occasions it might be only 16 or 17, or as many as 29 or 30, and all the results between with varying degrees of probability.

This is a good use for a normally distibuted random number.  In your example, the mean would be 23, the standard deviation would be 2.  95% of all results would fall between 19 and 27, 99% would be from 17 to 29.

Similarly, if some historical source says a battalion will break after taking 7% casualties, there should be a bell curve of results possible such that sometimes a battalion might break after only 5 or 6% casualties whilst sometimes a battalion might take 8 or 9% casualties before breaking.

This is not such a good example.  It's not like you could go through at the beginning of the game and assign a breaking point with mean 7% and standard deviation of 0.5%.  That 7% is actually an emergent behavior resulting from a lot of variables.  As JMM said above, it's too many for us to predict the outcome.

Hook

[Cpl Steiner]

For 1 vs 1, without other units, you always get the same result.. euh.. that said, there is a first parameter modifying this result.. the time scale.
The geometry of unit position is important, so changing the time scale can modify the distance between the units, so the result of casualties.

Forgive me JMM but this sounds a bit weird.

So, in other words, if my time-scale is 20/60, the moment the volley is computed could be when my unit is 80 yards from the target whereas if it is 30/60 it might be 100 yards from the target?

If this is how the game works players might be able to manipulate results be adjusting the time-scale at critical moments so that greater or fewer computations are performed.

[stupiddk]

I am more interested on how the losses are implemented on the field of battle. But if I were to say, I would say that some losses should be mathematical, like missing, and others like death should be determined by a variety of factors like height, range, experience, morale, and many more that have been mentioned before. I think the game has a VERY good 3d engine capable of a lot more than people expect. I believe it should be possible to represent mathematical and dice roll deaths on the battlefield.

[quartermaster]

One of the advantages of current levels of computer power is that it allows deterministic methods to be applied so it makes sense to exploit this approach.

Warfare is however not a simple case of numbers nor is everything linear or represented by a "normal" distribution.

Defensive fire against an attack is not a continuum when officers can control the point at which a volley is fired.  Once independent firing has started then you can consider a more statistical outcome. 

The concept of randomness which is used in many conventional wargames should not in any case lead to the example - Elite unit is defeated on a 1 in 6 basis.  What you might find is that the non-elite unit needs a score above 6 to defeat the Elite unit.  Other factors reduce the required score to 6 or below so that the possibility now exists that the Elite unit may be defeated but this is allowed some randomness once the result is a possibility.

[JMM]

Forgive me JMM but this sounds a bit weird.

So, in other words, if my time-scale is 20/60, the moment the volley is computed could be when my unit is 80 yards from the target whereas if it is 30/60 it might be 100 yards from the target?

If this is how the game works players might be able to manipulate results be adjusting the time-scale at critical moments so that greater or fewer computations are performed.

No... maybe one meter, no more! you can try! Just play with different time scale.
I talk about that to say it's possible you have a very light difference with this deterministic system.. no more..

Internal engine works with 1 mn historical step.. and it controls the speed in order to place the units at the right distance at the end of the step.
No really time to speak about that (sorry) but like all numerical systems, there is some errors (Lorenz..)
So, the model is robust and gamers can't play with the time scale to modify the results.

JMM

[Cpl Steiner]

Hi JMM,

Fair enough, it sounds like the game is pretty hard to manipulate, like you say.

My feeling on all of this though is that "determinism" has become some sort of dogma to be incorporated into the game at all costs just because some general here or there said that battles are won or lost by strategy and tactics, not luck.

If the game plays well anyway then all well and good but from a purely logical point of view we should recognise that, yes, sometimes even elite soldiers miss their target, and sometimes even the greenest recruit stands his ground. In war, nothing can be said to be certain.

[JMM]

If the game plays well anyway then all well and good but from a purely logical point of view we should recognise that, yes, sometimes even elite soldiers miss their target, and sometimes even the greenest recruit stands his ground. In war, nothing can be said to be certain.

Yes, right.. That said, we don't know the reason but maybe there is a reason... in this case, the system is deterministic but we think it's a random process...
That said, with HW:LG, I am sure it will be difficult to say it's a deterministic process because it's impossible to know all events.
However, I don't say other system are wrong... just it's an other model 
We'll talk about this system if you want again... after the release 

Enjoy !

JMM