barbarossa-tactical

[sPzAbt653]

'...except that I can clearly see that it is using the default AD setting.'

Is the default setting of '10' designed for a specific turn length? I don't see anywhere in the manual that explains what the 'default' is designed for, but I gather from Bob's comments that it may be. I'm not clear on what it's effects are, either. But if a setting of '10' was intended for say, one day turns, I can see where it would have an effect. The manual describes the AD setting as having an effect on combat, increasing or decreasing losses.

[wmorris]

Original: Curtis Lemay

And I think any Barbarossa scenario that ends with the Axis balked at Smolensk has some serious design problems. It is not demonstrating what the historical results would have been if the Soviets had just chosen to fall back further and faster.

I agree with this assertion. My opinion expressed above was that an organized defensive effort could have been offered starting at Smolensk, not that the Germans would be stopped there.

[wmorris]

Original: sPzAbt653:

Some similar changes were made in 'Buzz's FitE mod'. German supply level was reduced greatly at turn 22 and recovered later. Supply radius was also modified for both sides.

I have heard the changes in this mod discussed, but have never seen the mod itself. Link?

My (somewhat rhetorical) question: Is this mod an attempt to optimize simulation, or was it to balance the first 30 or so turns to offer a 50-50 chance of taking Moscow between skilled "gamey" players?

[sPzAbt653]

ORIGINAL: wmorris

My (somewhat rhetorical) question: Is this mod an attempt to optimize simulation, or was it to balance the first 30 or so turns to offer a 50-50 chance of taking Moscow between skilled "gamey" players?

I would say that it was an effort to improve the simulation. Much discussion is included in the 'Buzz's FitE Mod' thread, located down near the bottom of page 2 (if you haven't seen it yet). I don't think the mod was posted anywhere, only distributed thru e-mails. The version I have is from 11-07, so it may be dated. Try e-mailing Buzz for the latest, or I can send you what I have.

[Curtis Lemay]

ORIGINAL: ColinWright
Your comments about the attrition divider may well be accurate. However, at the same time I'm inclined to be suspicious of relying on a single tool like that. For one, it just seems unreasonable that combat in one place would be bloodier than combat elsewhere. Naturally, if units are vaporizing all the time advances will go faster. But is that really what was happening? I'd prefer to see the actual causes analyzed and then simulated.

Else one winds up with a subtler version of the fallacies inherent in such scenarios as the disc 1940 or the work of those who rely on shock to produce all desired rates of advance. Sure, one can make the front move back and forth more or less as it did historically. But that is not the same as simulating what took place. I'm reminded of SPI's War in the East -- which as I recall, largely obtained the historical rate of advance in 1941 through the wonderfully simple device of simply not having most of the Red Army there at all. It may work, but it's not a simulation of much of anything.

This isn't to say that the attrition divider shouldn't be changed. Merely that it shouldn't be seen as a panacea.

Here's a fact: TOAW combat lethality is completely independent of turn interval.

So combats will incur the same average losses whether they are conducted using 2.5 hour turn intervals or if they are conducted using whole-week turn intervals. All else being equal, that means that scenarios using 2.5 hour turn intervals will be 28 times as bloody as scenarios using whole-week turn intervals. That's clearly absurd, and exactly what the AD setting was designed to address.

The default AD setting is completely arbitrary. And its setting affects both sides equally. This is unlike shock, where the default setting of 100 is inherently neutral, and a change affects only one side. Since it's arbitrary, there is nothing inherently "right" about the default setting, and adjusting it to find the correct value for a given situation isn't "fallacious" - it is an essential design task.

Norm didn't provide any instruction on what the AD values should be - he left that to us. But, the full-day turn interval is the midpoint of the intervals. Coupled with my trial-and-error experiences, that leads me to conclude that for most cases the default will tend to work best with full-day turn intervals and needs to be decreased accordingly for longer intervals and increased accordingly for shorter ones. FITE uses the half-week turn interval.

Of course there are situational exceptions. That's why the setting determination needs to be made experimentally. And, of course, other factors need to be considered as well (I listed a bunch of them). But, thus far, the experiments done on FITE suggest that it is a candidate for a reduction in AD setting.

[sPzAbt653]

ORIGINAL: Curtis Lemay


... for most cases the default will tend to work best with full-day turn intervals and needs to be decreased accordingly for longer intervals and increased accordingly for shorter ones.

I understand that this is not arbitrary, and I think my silly brain gets it. Generally, one day turns calculate 1x losses, so 2 day turns would calculate 2x losses, and half day turns would calculate .5x losses (if the attrition divider is set accordingly). If a half-week turn has the AD set to a one-day turn, the calculated losses would be undesireably low, resulting in a possible stagnant front, for the game scale. But I think I found something contradictory. In 'Plan Martin', with one week turns, D. McBride says :

3.91: Attrition Divider: 35 (losses are around 25% of the default setting, often accepted as the norm for one week turns)

I'm not asking anyone to explain someone elses statement, but this does seem opposite to my interpretation. Although as Bob stated, there are other factors to be taken into consideration. It seems for a setting that is so critical, there should be some guidelines.

[sPzAbt653]

Umm .. I got that pretty wrong. 'Plan Martin' is Half Day turns. So I think the designer was saying that the AD is set to 25% of the accepted norm for one week turns, I think.

There should still be some guidelines for silly fools like me.

[wmorris]

Original: sPzAbt653

I would say that it was an effort to improve the simulation. Much discussion is included in the 'Buzz's FitE Mod' thread, located down near the bottom of page 2 (if you haven't seen it yet). I don't think the mod was posted anywhere, only distributed thru e-mails. The version I have is from 11-07, so it may be dated. Try e-mailing Buzz for the latest, or I can send you what I have.

I mus be really sleepy or exceptionally dim, I cant find the thread...

[sPzAbt653]

tm.asp?m=1401109

[Ian R]

ah, thanks

[ColinWright]

ORIGINAL: sPzAbt653

ORIGINAL: Curtis Lemay


... for most cases the default will tend to work best with full-day turn intervals and needs to be decreased accordingly for longer intervals and increased accordingly for shorter ones.

I understand that this is not arbitrary, and I think my silly brain gets it. Generally, one day turns calculate 1x losses, so 2 day turns would calculate 2x losses, and half day turns would calculate .5x losses (if the attrition divider is set accordingly). If a half-week turn has the AD set to a one-day turn, the calculated losses would be undesireably low, resulting in a possible stagnant front, for the game scale. But I think I found something contradictory. In 'Plan Martin', with one week turns, D. McBride says :

3.91: Attrition Divider: 35 (losses are around 25% of the default setting, often accepted as the norm for one week turns)

I'm not asking anyone to explain someone elses statement, but this does seem opposite to my interpretation. Although as Bob stated, there are other factors to be taken into consideration. It seems for a setting that is so critical, there should be some guidelines.

It'd be good to resolve the discrepancy between Bob/Curtis' interpretation that one day is the point the AD is set for and DMB's statement that it is a week.

Here it's worth noting that I've suspected that Norm basically designed his engine using his original Korea scenario as the test bed. In other words, I picture Young Edison up there in the garret, tinkering away with his program and deciding what was right by seeing what results it delivered for the Korea scenario.

As I recall, Korea used one-week turns. So while it pains me to say so, DMB may be right: the AD is designed for one-week turns, not one day.

Note, though, that I wouldn't assume a linear model for lethality. You can pump out a real fire-storm for a day -- but then you start to slow down. In OPART terms, over seven one-day turns, you won't be putting out as much on turn 7 as you did on turn 1. So if we assume one week is the break point, it might be best to multiply the attrition divider by less than seven for one day turns. The same point would apply no matter what the break point is -- in other words, even if Bob/Curtis is right.

[Curtis Lemay]

ORIGINAL: sPzAbt653
But I think I found something contradictory. In 'Plan Martin', with one week turns, D. McBride says :

3.91: Attrition Divider: 35 (losses are around 25% of the default setting, often accepted as the norm for one week turns)

I'm not asking anyone to explain someone elses statement, but this does seem opposite to my interpretation. Although as Bob stated, there are other factors to be taken into consideration. It seems for a setting that is so critical, there should be some guidelines.

No telling when Daniel wrote that. At one point I also thought that the default setting was for whole-week turns - and said as much, so he might even have been referencing me. For a while, I even had the AD set to 20 for CFNA, under that assumption. I only changed my mind after lots of trial-and-error. The more important point than what he said would be what he did, however. He set the AD to 35 for a half-day turn-interval scenario. That's close to in-line with assigning the default to full-day turn intervals.

As to guidelines, assuming the values are linear, then here would be the values for the setting, along with some values I've actually used for it:

Whole-week: 2 (Examples: Germany 1945 & Soviet Union 1941)
Half-week: 4 (Examples: CFNA, France 1944, The Next War 1979)
Full-day: 14 (Examples: Killer Angels 1863, Cambrai 1917, Kaiserschlacht 1918; Exception: Okinawa 1945)
Half-day: 28
6-hour: 56 (Example: Waterloo 1815, actually set to 40)

Okinawa is the exception to the rule due to the need to model the Jap cave defenses. It uses an AD of 50. And, at 40, Waterloo must have needed to be bloodier than normal WWII.

[ColinWright]

ORIGINAL: Curtis Lemay

ORIGINAL: sPzAbt653
But I think I found something contradictory. In 'Plan Martin', with one week turns, D. McBride says :

3.91: Attrition Divider: 35 (losses are around 25% of the default setting, often accepted as the norm for one week turns)

I'm not asking anyone to explain someone elses statement, but this does seem opposite to my interpretation. Although as Bob stated, there are other factors to be taken into consideration. It seems for a setting that is so critical, there should be some guidelines.

No telling when Daniel wrote that. At one point I also thought that the default setting was for whole-week turns - and said as much, so he might even have been referencing me. For a while, I even had the AD set to 20 for CFNA, under that assumption. I only changed my mind after lots of trial-and-error...

We may be crediting Norm with too much wisdom and diligent research.

The default setting might just be what worked for his Korea scenario. This, on the one hand, involved a lot of human-wave attacks by the Chinese, and on the other, involved the usual lavish American application of munitions. There's no particular reason to think the value he settled on represents some sort of universally valid constant.

[wmorris]

Original: Curtis Lemay

Whole-week: 2 (Examples: Germany 1945 & Soviet Union 1941)
Half-week: 4 (Examples: CFNA, France 1944, The Next War 1979)
Full-day: 14 (Examples: Killer Angels 1863, Cambrai 1917, Kaiserschlacht 1918; Exception: Okinawa 1945)
Half-day: 28
6-hour: 56 (Example: Waterloo 1815, actually set to 40)

I had a lightbulb moment about this and am gonna play around with it in my Fite mod. The mental picture I had was how a sim of US Civil War combat outcomes would look with the divider moving up and down as tactics evolved...

[ColinWright]

ORIGINAL: wmorris

Original: Curtis Lemay

Whole-week: 2 (Examples: Germany 1945 & Soviet Union 1941)
Half-week: 4 (Examples: CFNA, France 1944, The Next War 1979)
Full-day: 14 (Examples: Killer Angels 1863, Cambrai 1917, Kaiserschlacht 1918; Exception: Okinawa 1945)
Half-day: 28
6-hour: 56 (Example: Waterloo 1815, actually set to 40)

I had a lightbulb moment about this and am gonna play around with it in my Fite mod. The mental picture I had was how a sim of US Civil War combat outcomes would look with the divider moving up and down as tactics evolved...

This might be quick. You can't alter the attrition divider in the course of play.

[wmorris]

This might be quick. You can't alter the attrition divider in the course of play

Understood. I just meant the mental picture. Still, it would be nice.