"Realistic Attrition"

[Franciscus]

Leaving attrition out of the relation, agreed. Factors other than movement rates should discourage deep winter campaigning (i.e., in frozen conditions).

That's an important point. How many important marches can anybody recall were conducted during the ACW in frozen tracks or during snowstorms ?

[Gray_Lensman]

deleted

[Jagger]

--Rates of march are too slow.
--Cohesion loss rates are way too high.
--Attrition rates are also (probably) way too high.

I can agree with all 3 of these conclusions and I think that they ought to be addressed as separate issues.

I agree with all three points as well. However they cannot be treated as separate issues as they are all linked. Rates of march reduce cohesion which then increases attrition. See how they are linked? So impossible to treat as separate issues as a change in one impacts all three. This is a very important point to understand if we are going to reach a workable and satisfactory solution.

I don't think Jagger's numbers are good, until they properly value the effects of Mud in relation to Frozen, Snow. All the numbers can be reduced to effectively speed up all movement to get more historic movement rates, but Frozen, Snow movement (unless you are referring to really deep Snow), is effectively easier to move through than Mud. It could be the attrition factor for Mud is too severe, but to curtail that by making it easier to move thru Mud is an incorrect approach to the problem. Also, Mud may be appearing somewhat too frequently in the game's summer's months and that could be adjusted by lowering the probability percentages.

Have you really thought through your solution to its logical conclusion?

If you make mud marches more difficult than winter marches, then winter campaigning will make more sense than spring/fall campaigning. How would you justify this result historically?

If you make mud marches more difficult than winter marches, armies will suffer higher marching attrition when campaigning in rainy weather than in freezing temperatures below 32 degrees farenheit. Do you have historical examples which would consistently justify armies losing more men while marching in rain vs marching in temperatures below freezing? How would you justify those sort of game results?

In your definition of mud conditions, how many days of the 15 day turn involve substantially reduced marching due to very heavy mud? Is your answer consistent with the probabilities of mud appearing in the weather tables?

In your definition of snow/frozen conditions, how many days of the 15 day turn involve substantially reduced marching due to snow/frozen conditions?

These last two questions are important in determining the total distances marched over a 15 day period when comparing mud and snow/frozen conditions. Does 3 days of slowed mud marching and 12 days of normal ground marching conditions result in more or less distance marched vs 15 days of continuous slowed marching in frozen or snow ground conditions?

I have brought up these points multiple times in this thread and I have yet to see you answer any of them. You just keep stating my numbers are wrong and must be changed to meet your requirements before they are acceptable.

You need to answer these questions before I can accept your requirements as supportable and logical. As of now unfortunately, I don't have the impression you have a solid understanding of the linkage of attrition game processes nor have thought out the consequences of your solution well enough to propose a solid solution...or dismiss other's solutions out of hand. Think about the questions above a little and perhaps your answers will dismiss my concerns. Convince me with an argument instead of just telling me, "your numbers are wrong", just because they don't meet your conventional expectations. It is amazing how often convention is often wrong.

[bigus]

Anyone who lives in the North will know that Frozen ground in fall and spring will eventually through the day turn to mud. For game purposes I think the order should be ....
Clear>Mud>frozen>snow>blizzard.

This falls in line with the seasons as well Since you can have muddy conditions in spring, summer and fall.
Mud and Frozen should be roughly the same.

bigus

[Gray_Lensman]

deleted

[Gray_Lensman]

deleted

[Gray_Lensman]

deleted

[Gray_Lensman]

deleted

[bigus]

Seems a good analysis Jagger. I'll see with PhilThib and if he feels too that there is no problems, this can go official I believe. I'll wait for Michael's return though, so not before 3 weeks.

I think this debate might go on for longer than 3 weeks...... leure:
What we need is several DB's to work from. Then playtest...playtest....playtest. A working solution has to be out there.
It's easy to say the numbers are no good by looking at them. I've platested about 15 scenarios to date with Jaggers terrain numbers. I like them. Although not perfect ,I think they are better than the current ones. IMO

bigus

[Gray_Lensman]

deleted

[runyan99]

If you make mud marches more difficult than winter marches, armies will suffer higher marching attrition when campaigning in rainy weather than in freezing temperatures below 32 degrees farenheit.

I suspect the problem is the way the attrition options functions.

I take it that with the option on, movement causes attrition. Thus, slow movement over many days causes high attrition. Thus, your solution is to make mud faster than frozen, so that attrition is summer mud is lower than in winter.

The problem is that is a flawed solution to a flawed option. Movement in mud wasn't faster than winter. Instead, lack of fodder and the need for shelter tended to prevent winter campaigns. If the roads were hard, and there wasn't two feet of snow, movement was possible, but the men and horses might start dropping like flies if the conditions were too cold.

Here is how it should work:
There should always be some base rate of attrition to account for desertions and sickness. Call it X per turn. After that, the way it should work is that harsh weather, not movement, causes excess attrittion. Frozen = 2X attrition. Snow = 2X attrition. Blizzard=3x attrition. Then adjust downward for being entrenched, in a region with a city, etc. You can move in frozen conditions, but at a cost of attrition.

Movement on the other hand should cause cohesion loss, but not abnormal attrition. Movment in difficult conditions like mud or snow means movement takes longer, thus more cohesion loss. But, simply marching from point A to point B should not result in permanent casualties, if that is what the game is currently doing.

[Gray_Lensman]

deleted

[berto]

Anyone who lives in the North will know that Frozen ground in fall and spring will eventually through the day turn to mud...

I interpret "Frozen" to mean "very cold but dry." That's actually the typical winter conditions hereabouts (northern Illinois; but definitely not this winter, where it's been one snow after another!).

[Rafiki]

Since this currently is at a very "moddish" stage and in order to not confuse newcomers, I've placed it in the modding forum for now

[lodilefty]

<snip>

Also while putting together these comparisons, I couldn't help but wonder when the movement type of "Ranger" is actually applicable, especially since the specific model "Rangers" is defined as "$LitFoot" in the AACW_DB_Models.xls file

The Ranger type isn't used, maybe because of it's ability to 'walk on water'.
Over in WIA testing, I've proposed it be applied to units that historically had water movement ability and could really scoot in the wilderness: Rangers, Native Warriors, and the leaders thereof. We're also looking at it for other leaders, except Europeans who would have been less likely to ride long distances on a horse or in a canoe... interesting alternatives..

[Pocus]

It seems I overlooked some of the changes proposed, which is, given the amount of posts, not so surprising.

Mud movement cost should be high, higher than a movement cost in snow or frozen. The problem lies more in the chances of having mud, and not in the cost per se.
Mud in the game is really that, heavy rain and wagons stuck in tracks. This is far worse than moving in the others weathers types in my opinion.

I would advocate to mostly check if there is not too much mud weather in the weather matrix, but leaving the move cost in mud significantly higher than snow or frozen in the terrains matrix.

[Pocus]

Normal March & cohesion & hits incurred:

You lose cohesion on move, depending of the CohMove variable of the model + an additional penalty because of presence of enemy control. On average, 20% of the cohesion loss is giving one hit. 2/3 of these hits trickle back into your conscript pool.

=> It is normal to also lose men by fatigue. Exhausted men, frostbites and such make wounded which translate to hits to be removed to the regiment hits.
Cohesion loss can be seen as stragglers or fatigued men (but not enough to not be able to follow the regiment).

Weather impacts the number of days you take to move from point A to B. CohMove is a constant for a given model. But as you take longuer to move in bad weather compared to fair weather, you lose more cohesion. This is coherent.

Normal Attrition by Weather
Attrition by weather is checked at the start of the turn, if you are not in a region with a friendly structure. To get attrited, you have to be in a weather which is tagged as harsh or very harsh in the terrain table. Mud cost many days in travel, but is for example not tagged as harsh in clear terrain, but snow is. So you don't get additional attrition if you don't move in clear terrain, when the weather is mud, but will get some if you are under snow. This is WAD too.

Extended/hardened/Historical attrition
This new option introduce a new concept and is a complement to losses by weather, but extends the concept further.

The basic premise is that if you are in hostile land (ie enemy land or far from home), men can deserts. This is particularly true if under bad weather or if the region is pillaged. With this option, you can suffer cohesion loss even if you are not moving in fair weather! In this case, you would need to be in a pillaged, very disloyal region, but this is theorically possible. In this case, the cohesion loss means that the men are losing morale and some can deserts.
if you are in a disloyal region, and the opposing faction has the benefit of the Scorched earth rule, things are even more magnified.

This is another set of variables, and IMHO, you should discuss that in another thread as the mechanics are the same than the March Attrition.

[Gray_Lensman]

deleted

[Pocus]

Touché Actually you are quoting me on the weak part of the system... While I was typing that the system was coherent, I pondered further and found myself thinking of that:

Say you travel 10kms (that's 6.2150404 miles for you ) in 3 days under mud conditions but 25 kms under clear weather. The men under the current system will spend the same amount of fatigue, because they marched 3 days . But it is fair to assume that in reality, it asked for more efforts to march 3 days under mud, even if it led to a shorter travel distance in the end.

This is why I agree there is something amiss, that a set of new variables can correct. You can thus use them for the modifications you foresee.

[Gray_Lensman]

deleted

[runyan99]

Normal March & cohesion & hits incurred:

You lose cohesion on move, depending of the CohMove variable of the model + an additional penalty because of presence of enemy control. On average, 20% of the cohesion loss is giving one hit. 2/3 of these hits trickle back into your conscript pool.

=> It is normal to also lose men by fatigue. Exhausted men, frostbites and such make wounded which translate to hits to be removed to the regiment hits.
Cohesion loss can be seen as stragglers or fatigued men (but not enough to not be able to follow the regiment).

Pocus thanks for the explanation. I'll address this bit first. I agree that movement can generate fatigue and stragglers, but I would not actually simulate stragglers with hits, even if 2/3 return to the conscript pool. It is enough to model this with cohesion loss, and recovery, in my opinion.

Extended/hardened/Historical attrition
This new option introduce a new concept and is a complement to losses by weather, but extends the concept further.

The basic premise is that if you are in hostile land (ie enemy land or far from home), men can deserts.

I would challenge this assertion. I don't think that men tended to desert while in enemy territory.

It would be more realistic I think for the attrition option simply to apply a base rate of loss (or chance of loss) for all units each turn, to simulate a constant percentage of men who get sick or desert, have expired enlistments or whatever. This would apply at all times, wheter moving or stationary, but not be increased while moving or in enemy territory.

All things considered, the current system seems to do far too much damage to units simply for moving into enemy territory. This is making movements like Lee's advance to Gettysburg or Sherman's march to the sea prohibitive in AACW.

[Gray_Lensman]

deleted

[lodilefty]

Once we have these new variables to work with, we should probably start over with the movement rates to bring them in line with actual historical march rates. (allowing for the new Divisional slowest rate effects).

After adjusting the movement rates, we can adjust the new variables for the differing weather related movement attrition rates.

Finally, (as needed), we can adjust the "CohMove" column entries as mentioned above in the "AACW_DB_Models.xls" file for specific models cohesion differences.

I took a quick look at the weather matrix. Mud is almost horrifyingly high occurence in Spring and Fall months [over 50% in some terrains]. I think this needs to be part of our integrated solution, too, maybe more viable than making radical reductions in the base movement values.

...and as always, I advocate use of leader abilities and forced march to deal with 'long fast' marches seen historically.....

[Gray_Lensman]

deleted

[berto]

...and as always, I advocate use of leader abilities and forced march to deal with 'long fast' marches seen historically.....

Okay, but please keep in mind the 1862 Kentucky Campaign, where three different commands--Bragg, Smith, Buell--all achieved 'long fast' marches, albeit entirely under fair weather, and all without undue cohesion or attrition loss.

In other words, unless we are to believe that the 1862 Kentucky Campaign was an improbable fluke, you would either have to

(a) give special fast mover attributes to all three commanders (to Buell???)

or

(b) greatly reduce the ill effects of forced marching (since it doesn't seem that the commands at Perryville were especially "strung out" or decimated)

[Gray_Lensman]

deleted

[Jagger]

For everyone's information here:

I requested 5 new weather related attrition movement variables to be exported into one of the (.opt) files.

ClearAttritMult......(should be extremely small on a per day basis when moving)
MudAttritMult.......(even though the cohesion loss rate should be higher, the attrition rate due to Mud should not be much higher than Clear when moving)
SnowAttritMult.....(similar cohesion effects to Mud, but somewhat more severe attrition rate when moving)
FrozenAttritMult....(lower cohesion effects than Mud, but more severe attrition rate when moving)
BlizzardAttritMult...(high cohesion effects and the most severe attrition rate when moving)

Of course the cohesion loss rates will still be tied directly to the movement costs which is reasonable, but with these new weather related attrition movement variables, we can effectively separate the individual attrition effects from the movement rates by adjusting the above variables.

Also, if necessary there is actually a "CohMove" column in the "AACW_DB_Models.xls" file that can be adjusted for separate models if needed to fine tune Cohesion Losses due to Movement for the individual models. This should probably be adjusted last and only if necessary.

To summarize:

Once we have these new variables to work with, we should probably start over with the movement rates to bring them in line with actual historical march rates. (allowing for the new Divisional slowest rate effects).

After adjusting the movement rates, we can adjust the new variables for the differing weather related movement attrition rates.

Finally, (as needed), we can adjust the "CohMove" column entries as mentioned above in the "AACW_DB_Models.xls" file for specific models cohesion differences.

Lastly.

Thanks to Pocus for agreeing to implement these variable exports and let's have some patience in waiting for him to work out the complications.

Also, though we have had some disagreements, Thanks to Jagger for bringing this problem to the attention of everyone here and trying so hard to find a solution within the current game engine constraints.

Excellent! These parameters will help immensely in creating accurate march rates and attrition rates. Although I do not envy whoever is going to create the updated weather data. That is going to be a very challenging project producing realistic results for heavy mud conditions considering the many differing weather regions within the game.

The CohMove model parameter is useful for modeling the higher attrition rates of new troops while marching. Militia/volunteers/conscripts all have a higher CohMove rate which results in a higher marching, attrition rate for new troops. In addition, the CohMove parameter could be useful for modeling a very fast but low endurance unit or vice versa when combined with family type.

Unfortunately, the CohMove factor continues to tie attrition to movement. I agree with Runyon that marching was not a major cause for permanent loss of men. Stragglers returned to their regiment often on the very same day or by the start of the next march.

It does seem realistic that high cohesion loss due to marching temporarily reduces the combat power of a regiment. However the loss of combat power is only temporary as troops return to their regiments. Even low cohesion due to battle should not represent a permanent loss of men as again, the men eventually return to their units. Unless, of course, the army is completely destroyed.

Historically attrition rates were closely tied to the time in service and quality of the unit. Many civil war regiments lost hundreds of men within the first three to 6 months of entering service. Large numbers of recruits were lost to illness as they were exposed for the first time to various easily transmittable diseases. Many recruits were physically not suited to campaigning or the army life and were discharged. Those that enlisted for a bounty would often desert shortly after muster. For many reasons not tied to movement, a regiment lost a high percentage of their troops within months of the regiments creation. However as time passed, the survivors that remained were reasonably motivated, acclimated and suited for the army. At some point, the attrition rate drops to a very small percentage requiring severe events to produce high attrition.

Ideally, it would be very useful to have a separate attrition factor directly linked to models, but unrelated to movement, which would reflect factors such as illness, desertion, unfit men, bounty jumpers, etc.

With a separate model attrition parameter, we could adjust attrition rates to the units time in service. Volunteers and conscripts could have high rates of attrition which are reduced to very low rates once they upgrade to early war or late war status.

So a 1000 man volunteer or conscript regiment may lose 200 to 300 men over the 2 or 3 months required for upgrade to early war troops. Once the regiment upgrades to early war status, the attrition rate of the early war model would be significantly lower than the volunteer/conscript model. Thus a seasoned early/late war regiment might only lose 50-60 men a year to illness, desertion or various incidents vs the 200-300 over the first few months after creation.

The attrition parameter would simply deduct a percentage of men per turn. If you wanted, various factors could slightly adjust the base attrition percentage such as the season of the year, cohesion and whatever other defined factors might lower or increase attrition. The parameter would probably need to work with fractions of a percent such as 0.25% or 1.5%-very small percentages.

This sort of attrition factor, tied directly to models and separate from movement, could model the initial high troop losses when a unit is first created and then acceptable small attrition per turn once the regiment is seasoned.

Anyway, just a thought.

As an aside, it would also be nice to have the ability to model the lack of winter campaigning by cavalry. Lack of ready fodder kept most cavalry dependent on a steady and reliable source of supply which often was not available in the field. Cavalry could operate in the field during winter only if they brought along huge amounts of supply to feed their horses. So during winter, cavalry usually stayed near a steady supply of fodder. Although I don't have any quick and easy ideas on how to model this characteristic of cavalry.

[berto]

This also might be somewhat easier to replicate, once movement rates are adjusted, and "attrition" due to movement reduced somewhat. Though it should not be an everyday possibility. Fair weather would have to be the major cooperating/determining factor for the entire move.

I totally agree. Anything like the 1862 Kentucky Campaign would have to depend on 1862-like weather conditions--i.e., fair weather over the entire campaign area for an extended period.

To some extent, too, anything 1862 Kentucky Campaign-like would also be dependent on "lucky" activations: all three commands--Bragg, Smith, Buell--avoiding inactivation, else at least having a reasonable forced march alternative. (Meaning to say, in the game: If the CSA side has the good fortune and opportunity to launch an 1862 Kentucky Campaign-like northward thrust, even if Buell is inactive, the USA side should be able to respond to that thrust in "reasonable" fashion--i.e., use forced marches to chase after Bragg & Smith as necessary, but without facing force decimation, much less destruction due to extreme cohesion and attrition loss.)

[berto]

Thanks to Pocus for agreeing to implement these variable exports and let's have some patience in waiting for him to work out the complications.

Also, though we have had some disagreements, Thanks to Jagger for bringing this problem to the attention of everyone here and trying so hard to find a solution within the current game engine constraints.

Yes, indeed, thanks to all concerned, but especially the "two Phils" for creating such a multi-layered, adaptable game engine that leaves nothing essential out (and that scales across their entire game product lineup); also their commitment to overall excellence and "getting it right." It's why this game holds out so much promise and why I have such faith in it.

Exciting game, exciting company, exciting player/modder community, exciting times...

[Gray_Lensman]

deleted