Should we just ignore Front/Army Group HQs?

[fbs]

A Front HQ has exactly 3x the command points of an Army HQ, and an Army Group HQ has exactly 4x the command points of an Army HQ.

For each command point that exceeds the HQ's capacity, that's the same as decreasing 1 from the leader's rating on whatever check it is doing.

Therefore, if one has 2 or 3 combat units attached to Front HQ or Army Group HQ, then one can only fill it with 2 (soviet) or 3 (german) fully built armies, plus one partially built army.

Now, we have more armies than 2 or 3x the number of fronts or army groups. The Soviets get 60 armies, for example, for 11 fronts, ergo should have something like 5 armies per front. But the front will only support 2.5 full armies, so the options are:

(a) Put 5 weak armies per front, but leave them at 50% capacity
(b) Put 2.5 strong armies per front, and leave the rest with STAVKA (that travelling circus!)
(c) Put 5 strong armies per front and just ignore the front HQ penalties

I'm thinking that (c) will lead to stronger forces, given that 12-15 divisions will report to the same army and suffer no coordination penalties. The downside is that the Front HQ will be severely loaded, so one can just put a bad leader there, as any checks are expected to fail anyway.

That seems better than having 6 divisions per army and hoping that the front HQ will provide assistance if needed (at best the front provides assistance in 25% of the checks, even with a good leader at zero distance).

What do you guys think?

[AFV]

My opinion is that you go with (c). Try and get die rolls at the Front level, and at the STAVKA level, and if you get any at the Front level its a bonus.
Be interesting to hear what some of the better Soviet players do.

[mmarquo]

Another option is to fully load the army headquarters, and then directly assign any extra units directly to the front headquarters. The units directly assigned to the front headquarters have a better chance of being committed as reserves.

[Walloc]

ORIGINAL: fbs

A Front HQ has exactly 3x the command points of an Army HQ, and an Army Group HQ has exactly 4x the command points of an Army HQ.

For each command point that exceeds the HQ's capacity, that's the same as decreasing 1 from the leader's rating on whatever check it is doing.

Therefore, if one has 2 or 3 combat units attached to Front HQ or Army Group HQ, then one can only fill it with 2 (soviet) or 3 (german) fully built armies, plus one partially built army.

Now, we have more armies than 2 or 3x the number of fronts or army groups. The Soviets get 60 armies, for example, for 11 fronts, ergo should have something like 5 armies per front. But the front will only support 2.5 full armies, so the options are:

(a) Put 5 weak armies per front, but leave them at 50% capacity
(b) Put 2.5 strong armies per front, and leave the rest with STAVKA (that travelling circus!)
(c) Put 5 strong armies per front and just ignore the front HQ penalties

I'm thinking that (c) will lead to stronger forces, given that 12-15 divisions will report to the same army and suffer no coordination penalties. The downside is that the Front HQ will be severely loaded, so one can just put a bad leader there, as any checks are expected to fail anyway.

That seems better than having 6 divisions per army and hoping that the front HQ will provide assistance if needed (at best the front provides assistance in 25% of the checks, even with a good leader at zero distance).

What do you guys think?

Since u mention 3* army = Front or 24*3 = 72. This leads me to believe u talk early on. Since every thing is a mess early on i just try and get the best C&C possible up and going. Mostly worring about keeping the army within its limits. As soon as the initial unrush settles down i try and steamline my armies/fronts much more. I try and look ahead so i dont attach more then 4 armies per front. As changing HQs from front to to front is so expensive.

Since the army CP limit fair fast falls to 18 the math looks like 4*18 = 72. This means 4 armies per front, but overloading with airbases as they count for 1 each. Then from 3/42 the front capacity raise to 81 so u then can have the 4 armies of 18, plus airbases and a few units usually directly attached to the front as reserve units and keeping all limits.
I at this point usually have several armies attached directly to Stavka, but i put those in the areas of static front so rolls dont matter as much.

Later on in 4/43 the front capacity raises again and u can then start too put a 5rd army usually at bit less str or more attached directly to teh front for even more reserves and since u recieve another front about there, there really shouldnt be all that many armies attached to Stavka any more.

So my approche is yeah early on u just hafta do but i do try and think ahead to when i start to streamline the system keep within the limits from around 3/42. It becomes a pain with the new tank army with 15 CP "arrives", but well they made that for a reason [;)]
If u look at the modified vs initíal CV in attacks it seems to me that leader ratings matters a whole lot, up to 2-3 times as high mod as initial CVs . So i see no need to lower leader ratings by overloading armies/front when unavoidble. Even incuring a 10-20% coordination penalty is much less than the possible modifier in modified CV comming from leader rolls.


Kind regards,

Rasmus

[fbs]

Hmmm, that's a great point. It remains constant as Army Group = 4x Army for the Germans, but for Soviets after 1.05.59 it is now:

6/41 to 8/41: Army = 24, Front = 72; Front = 3x full Armies
9/41 to 3/42: Army = 21, Front = 72; Front = 3.4x full Armies
4/42 to 3/43: Army = 18, Front = 81; Front = 4.5x full Armies
4/43 to 3/44: Army = 18, Front = 90; Front = 5x full Armies
4/44 to 3/45: Army = 18, Front = 99; Front = 5.5x full Armies

So with other attachments to Front it should be like:

6/41 to 8/41: Army = 24, Front = 72; Front = 2x full Armies + 1x weak army + 9cp
9/41 to 3/42: Army = 21, Front = 72; Front = 3x full Armies + 9cp
4/42 to 3/43: Army = 18, Front = 81; Front = 4x full Armies + 9cp
4/43 to 3/44: Army = 18, Front = 90; Front = 4x full Armies + 1x weak army + 9cp
4/44 to 3/45: Army = 18, Front = 99; Front = 5x full Armies + 9cp

Thank you for pointing that out.

[CheerfullyInsane]

ORIGINAL: fbs

A Front HQ has exactly 3x the command points of an Army HQ, and an Army Group HQ has exactly 4x the command points of an Army HQ.

For each command point that exceeds the HQ's capacity, that's the same as decreasing 1 from the leader's rating on whatever check it is doing.

Er, no.
Something wrong with your math there. Check the example in the manual (11.3.2).
For units directly attached, you lose one leader-rating per 2 CPs overloaded.
Once we get to the second and third level, I'm too mathematically challenged to crunch the numbers exactly, but since range and base-chance both increase, the overloading becomes less important.
Then again, as the levels increase the odds of activating in the first place become pretty low in the first place.

(a) Put 5 weak armies per front, but leave them at 50% capacity
(b) Put 2.5 strong armies per front, and leave the rest with STAVKA (that travelling circus!)
(c) Put 5 strong armies per front and just ignore the front HQ penalties

I'm thinking that (c) will lead to stronger forces, given that 12-15 divisions will report to the same army and suffer no coordination penalties. The downside is that the Front HQ will be severely loaded, so one can just put a bad leader there, as any checks are expected to fail anyway.

Or d) Use the upcoming 18 CP limit for Armies, attach four of them to Fronts (until their CP increase) and only overload the Front with the few airbases necessary. Meaning you'll have 4 full armies, each covering a 2-3 hex front.
The coordination-penalties can be annoying, but they're usually avoidable with a little fancy foot-work.
Have a handful of reserve-armies (complete with SUs) attached to STAVKA to use as fire-brigades, and anything left over gets attached directly to STAVKA while training/refitting.
This way you can cycle individual units in and out of your Armies to recover fatigue (for 1 AP a pop), instead of having to constantly (re-)attach complete Armies to Front HQs.

That seems better than having 6 divisions per army and hoping that the front HQ will provide assistance if needed (at best the front provides assistance in 25% of the checks, even with a good leader at zero distance).

Yep, leave the frontline decisions to the Army HQs. Having a Front (and thus another leader-check) is nice, but I wouldn't count on it. Using units directly attached to a Front HQ would work, but that'll mean fewer troops for the armies under that HQ, plus they'd also use the Inf/mech rating of the Front commander, meaning you'd have to find more good combat-leaders that are already scarce.

[pompack]

CI, I read 11.3.2 carefully then I went back and re-read 11.3

Everything including the example shows that you lose one rating point for every cp it's overloaded. What am I missing? [&:]

[fbs]

ORIGINAL: CheerfullyInsane
Er, no.
Something wrong with your math there. Check the example in the manual (11.3.2).
For units directly attached, you lose one leader-rating per 2 CPs overloaded.
Once we get to the second and third level, I'm too mathematically challenged to crunch the numbers exactly, but since range and base-chance both increase, the overloading becomes less important.
Then again, as the levels increase the odds of activating in the first place become pretty low in the first place.

I stand corrected; the relationship is not linear. This is from my other post:

From that, I think that the chances of success for those leader checks are:

1st chance: leader_rating / ( 10 + (CP-CC))
2nd chance: leader_rating / ( 20 + (CP-CC) + distance/divider )
3rd chance: leader_rating / ( 40 + (CP-CC) + distance/divider )

Therefore, exceeding CP by 10 at any level will halve the chance of getting a first-chance check at that level. That is, if an unit is attached to an Army Group and the leader has a rating of 8, then 10 excess CP makes the check to be the same as if the leader had a rating of 4.

That is, exceeding CP by 10 is terrible for all units attached to that HQ, whatever its level. It is the same thing as decreasing the leader stats by 50%.

Therefore, the decrease in the effective leader rating is as follows:

Code: Select all

 Excess CP  Decrease leader rating
 =========  ======================
 1          9.1%
 2          16.7%
 3          23.1%
 4          28.6%
 5          33.3%
 10         50%
 15         60%
 20         66.7%

Therefore the effect is more pronounced leaders with higher ratings that are very slightly overloaded; a leader with rating = 10 will go down to 8.3 on exceess of 2 cp (that is, lost 1.7)

But, a leader with rating = 3 will go down to 2.5 on excess of 2 cp (that is, lost only 0.5).

So leaders with high ratings will lose some 0.8 point for 1 cp (initially, and less as they get more overloaded), while leaders with low ratings will lose about 0.25 for 1 cp (and then lose less as they get more overloaded).

Of course, all that is for first-chance checks. For second and third-chance checks the effect is much smaller, as these checks are already have a small chance of success.

I think the bottom-line is to not overload armies, never.

[marty_01]

ORIGINAL: fbs

ORIGINAL: CheerfullyInsane
Er, no.
Something wrong with your math there. Check the example in the manual (11.3.2).
For units directly attached, you lose one leader-rating per 2 CPs overloaded.
Once we get to the second and third level, I'm too mathematically challenged to crunch the numbers exactly, but since range and base-chance both increase, the overloading becomes less important.
Then again, as the levels increase the odds of activating in the first place become pretty low in the first place.

I stand corrected; the relationship is not linear. This is from my other post:

From that, I think that the chances of success for those leader checks are:

1st chance: leader_rating / ( 10 + (CP-CC))
2nd chance: leader_rating / ( 20 + (CP-CC) + distance/divider )
3rd chance: leader_rating / ( 40 + (CP-CC) + distance/divider )

Therefore, exceeding CP by 10 at any level will halve the chance of getting a first-chance check at that level. That is, if an unit is attached to an Army Group and the leader has a rating of 8, then 10 excess CP makes the check to be the same as if the leader had a rating of 4.

That is, exceeding CP by 10 is terrible for all units attached to that HQ, whatever its level. It is the same thing as decreasing the leader stats by 50%.

Therefore, the decrease in the effective leader rating is as follows:

Code: Select all

 Excess CP  Decrease leader rating
 =========  ======================
 1          9.1%
 2          16.7%
 3          23.1%
 4          28.6%
 5          33.3%
 10         50%
 15         60%
 20         66.7%

Therefore the effect is more pronounced leaders with higher ratings that are very slightly overloaded; a leader with rating = 10 will go down to 8.3 on exceess of 2 cp (that is, lost 1.7)

But, a leader with rating = 3 will go down to 2.5 on excess of 2 cp (that is, lost only 0.5).

So leaders with high ratings will lose some 0.8 point for 1 cp (initially, and less as they get more overloaded), while leaders with low ratings will lose about 0.25 for 1 cp (and then lose less as they get more overloaded).

Of course, all that is for first-chance checks. For second and third-chance checks the effect is much smaller, as these checks are already have a small chance of success.

I think the bottom-line is to not overload armies, never.

Looks roughly logarithmic -- at least from the data points you provided above. Quick regression yields:

%Decrease in Leader Rating = 20.169 x LN (CP) + 3.7194
Correlation: R^2 = 0.98

But the affect is pretty brutal for even moderate overages in CP for a leader. A lot of Front Commanders, as well as STAVKA; and OKH and Army Groups get schwaked pretty hard. Seems like the CP overage should be on some sort of sliding scale based upon HQ level. Moreover, perhaps the leadership decrease breaks for say an overall Command HQ should be double or triple the CP overage for a Korps or Army HQ.

Or alternatively and perhaps a more elegent function would be – same percentile decrease for all leaders for CP overages, but based upon PERCENT of CP over maximum CP for a given percentile decrease in leadership rating.

[CheerfullyInsane]

ORIGINAL: fbs

Therefore, the decrease in the effective leader rating is as follows:

Code: Select all

 Excess CP  Decrease leader rating
 =========  ======================
 1          9.1%
 2          16.7%
 3          23.1%
 4          28.6%
 5          33.3%
 10         50%
 15         60%
 20         66.7%

..(snip)..

I think the bottom-line is to not overload armies, never.

It's not that linear. Increasing the overload only increases the divider, so the net effect will be reliant on the original Leader Rating.
Like this:


So a high-rated leader will be hit harder, but he'll also be able to manage more troops with some sort of effect.
But you're right in the bottom line. Don't overload Armies. [:)]

ORIGINAL: marty_01

But the affect is pretty brutal for even moderate overages in CP for a leader. A lot of Front Commanders, as well as STAVKA; and OKH and Army Groups get schwaked pretty hard. Seems like the CP overage should be on some sort of sliding scale based upon HQ level. Moreover, perhaps the leadership decrease breaks for say an overall Command HQ should be double or triple the CP overage for a Korps or Army HQ.

Perhaps but then again, it's not the job of a Front HQ to allocate reserves on the front-line.
Most (myself included) have leaders with high Admin in the higher HQs to make attachments cheaper. Their combat-ratings are of secondary importance. Granted, it would be nice to have high Initiative leaders everywhere, but there are only so many quality guys around.

[Walloc]

Just wanting to add one thing. Considering how rounding off works in other places in the game, and through years of playing GG games.
I have a gut feeling based in experience, underlining its nothing more than that. That any fractional numbers are rounded either off, but most likely by my gut are rounded down. Its for example how it works with the modifer for factories. We know that. U cant roll 3.8 with a 10d die. So the check would be against 3.

If that is the case for example a rating of 4 minus 5% might be 3.8 but might in case im right its actually a lot higher percentage loss.
So the usual thing will be that the actual loss is higher than the ones presented in the tables in particular for lower rated ratings some what cancelling out the note from others that higher rated ratings are hit harder.

Underlining again its a feeling,

Rasmus

[fbs]

ORIGINAL: Walloc
I have a gut feeling based in experience, underlining its nothing more than that. That any fractional numbers are rounded either off, but most likely by my gut are rounded down.

That's a good point.

Notice, though, that the "equivalent rating" is my way of looking at it, to make it easier to understand what is the final effect of the procedure. The actual procedure is this:

if random( 10 + CP - CC ) < leader_rating then success

therefore the chance of the check passing is

leader_rating / ( 10 + CP - CC )

which is the same thing as

( leader_rating * 10 / (10+CP-CC) ) / 10
= leader_rating' / 10

where leader_rating' = leader_rating * 10 / ( 10 + CP - CC ) is what I've been calling "equivalent rating" thing, because it has the same chance of passing the check (without cp penalties) as the original leader (with penalties).

I have no idea how the actual code works, but chance are it doesn't calculate that fractional rating equivalent, but just uses the procedure above.


ps: by the way, I find it highly unrealistic that the very first command point in excess has a great effect in the leader. I means something like "Oh, I'm the big super-dude general and I'm super-happy, but if I get one more company to lead then I'm distressed".

I would rather see that effect kicking in more progressively, rather than having a big impact on first extra cp and then decreasing the effect with additional ones. I mean, if the guy has 100% efficiency with 90,000 men, how come he has 90% efficiency with 91,000 men, and 80% efficiency with 150,000 men?

[Walloc]

ORIGINAL: fbs

I have no idea how the actual code works, but chance are it doesn't calculate that fractional rating equivalent, but just uses the procedure above.

Hi fbs.

Neither do i, so u could very well be correct. Any how over all one way or the other it doesnt matter much, this is just minor details. [:)]

Kind regards, suffering from to much time due to no turn in mail box syndrom

Rasmus

[fbs]

ORIGINAL: Walloc

Hi fbs.

Neither do i, so u could very well be correct. Any how over all one way or the other it doesnt matter much, this is just minor details. [:)]

Kind regards, suffering from to much time due to no turn in mail box syndrom

Kind regards too, my friend