Leader Rolls and Halted Attacks

[Wiedrock]

Since, some people seemingly are not affected by the random halting of attacks, I chose this as the subject of this Thread.
Some believe in the possibility to predict every combat outcome with accuracy to get close to 100% win rates, even with often relatively close final odds. All this without having accurate CV (and what they are made off), SU committments, RESERVES, Fort reductions, Leaders and Leader Rolls known before combat.
Therefore we will have a look at some "Halted at range 123" attacks and see what we can make off it to see if there is actual a way to make/predict 100%'ish winning attacks.

This post contains assumptions (maybe the Devs give more input) and simplifications. It focuses on single on single Divisional action.

---

Leaders in Combat:
As was established some time ago, Leaders are (likely) not affecting the Combat performance of Ground Elements directly, maybe there is still some caveat to Initiative.
They certainly do it indirectly through Supplies and such.
More important for results are their Combat Rating rolls for the overall final CV calculation.

For combat relevant ratings are:
Initiative Rating:

• SU committments
• RESERVE activations
• converts Hasty Attacks into SCOUTs
• involved in halting attacks in some unknown way (with higher Initiative)

The Initiative leader rating is used for determining the actual number of movement
points a unit will have during the turn, [...] the ability of support units and combat units in reserve status to commit
to a battle, and the ability to reduce casualties by turning a low odds hasty attack into
a reconnaissance in force [SCOUT].
Equally leaders with an initiative rating of more than 5 are more likely to halt a poor
odds attack at a greater range, thus reducing overall attacker losses.

Mech/Inf Ratings:

• Final CV
• (very likely) rolls "intermediate CV" in a similar way to decide whether an attack is halted at "Halt range 123"

Mechanized (Mech) and Infantry Ratings: These ratings are used to determine [...] Successful rating checks will increase combat value [...]

Mech ratings apply to motorized units and the infantry ratings to non-motorized.

The Leader shown in the Combat window is only rolling the final CV for his own units, units of other HQs roll their own units, I assume this works the same for intermediate CV.
...there is certainly more rolls related to things like retreat losses/retrat distance, but that is "after combat", so does not matter here.

TLDR for CV of Ground Elements, each Element has a CV, CV is impacted by many things, at start of combat the CV of an element are determined/fixed depending on those many factors, if an element is DEStroyed, DAMaged or DISrupted during combat it does no longer contribute to the final CV (or the intermediate CV - assumption).

start.
So, as the title states we want to figure out how/why attacks are halted. For that we first of all need to understand how the final CV are generated/impacted by leader rolls. Since - assumption - the intermediate CV are calculated in the very same way.

For that I will quote Joel (can't find it in the Living Manual somehow):

Each block of elements in a unit (i.e. 500 Rifle Squads, 25 37mm AT guns, 5 T-34 76s) rolls the leader rating and can be multiplied by 1.25 or divided by 1.25. In addition each distinct unit in the combat (i.e. 3rd Rifle Division, 7th Artillery battalion) does a leader roll and can have it's CV value halved.
[...]

So, this being established, we can look further into the halted attacks.
This, as mentioned, I expect to be the same calculation (to the largest degree at least), since it shows the very same changes to CV you can expect in a Final CV. Whether this is being checked after/before every range or Leaders decide whether they "want to compare" after each range....or whatever is hard to say.
It certainly is possible that the Initiative Ratings are additionally used to start/"accept" the comparison, but I expect the Mech/Inf Ratings to be used for the actual CV, since as said, it behaves the same way. Joel mentions it this way:

Also, initiative comes into play with who gets to join combat and I think it is initiative that determines the disengagement decision once the combat starts, but I didn't look into these today, [...]

So the Initiative rating may be involved, maybe as
- start a comparison at a range X or Y
- decide whether the comparison is "accepted"
- ...
Another assumption is that the Attack always gets halted once a intermediate CV is in favor of the Defender (...again here a potential Initiative roll could come in (if it exists) to prevent this "loss" being applied to the attacker).

Now..., accepting that the actual CV shown in Halted attacks behaves the same way as Final CV and using the Mech/Inf Ratings for that, and assuming that the actual "decision" is either started or approved (or whatever) by Initiative we can draw some preliminary conclusions.

• Each Roll can fail. A 90% roll chance means, that still 10/100 rolls will fail.
• The worst roll to fail is the "/2" for the whole unit.
• To have a 100% success rate and not getting a single Attack halted you need a way larger CV superiority than one might expect, this superiority can be to some degree replaced by Artillery and Air.
• the chance to halt early is directly related to initial CV (minus what has been hit until that range) which then gets "rolled" like final CV
• The more Units are involved in a battle, the less influental the "/2"-roll becomes - since not all Untis will fail the roll.

To have a 100%'ish success rate your initial CV would have to account for the worst case. Which means, you fail all /1.25 rolls and the /2 roll. Let's say this case is highly unlikely and assume Germans on average increase CV through the 1.25 rolls by (on average) about +20-25% (let's say 25%). Soviets stay "neutral" for sake of easing things.
So a German would need to have about Soviet CV/1.25*2=1.6→160% of Soviet initial CV. This ofc. ignores what gets hit until the "range roll" (we come to this now).
To calculate how many CVs actually have been hit/removed from calculations until a point is rather hard since targeting varies between combats/cases/tests and targets can get hit/counted multiple times. So all we can do for now is an simple "all elements hit"/ready elements=leftover percentage of CV and assume elements all have same CVs per element on average. So If out of 1000 elements 200 are DIS, 100 DAM and 50 DES, we would get (200+100+50)/1000=35% got hit, meaning 35% of initial CV removed. As said a massive simplification.

example1.
The following is the very same attack, here you can see the difference between failing a "intermediate CV roll" and not failing it (or in contrary - Soviets failing it and not failing it).
I could add way more runs, but this is the magnitude what to expect for this encounter.
As can be seen the combat CV predictor tooltip had a rather large difference (bigger than usual), being off by -~62%. But this is another subject, if you have info on that add it to the linked Thread.

The Soviets in the first halted attack barely hit any Germans, we can see that only ~6% of German ready elements have been hit, but 26.6% of Soviets ready elements. So whatever we talk about, the actual hitting does not play a large role - but as mentioned may tip the balance occasionally in other cases.
So the German CV of the halted attack clearly shows a sign of failing the "/2" roll. 99/2=49.5 6% were hit, so 49.5x0.94 and the average increase of German CV by (see above) +25% gives us 61.88 Intermediate CV. This is not science and assumes lots of things I have pointed out, but it shows that this leader in this case failed the "/2" roll with 100% certainty for the Division.

So what CV superiority would he have needed had he wanted to win with 100% certainty, even when failing the /2 roll (which happens 15% of the time with 85% roll chance) while Soviets aren't failing theirs?
He would have needed:
- 26.6% Soviets getting hit
- 6% Germans getting hit
- German average increase +25%

Soviets stay "neutral" and lose 26.6%, means the German would need to have more than 130CV*0,734=95.42CV after failing his German /2-roll.
Germans lose 6% and x1.25 the rest, so it should be around 163CV for the attacker to make it a 100% win.
[need]/2*0,94*1,25=95.43CV → [need]=95.43CV*2/0,94/1,25=162.43CV → rounding up ~163CV!

All this is excluding the (in this case very large) Combat Tooltip CV prediction. So even assuming a Tooltip variance of ~+/- 25% would give a need for ~204CV to make it a 100%'ish win.
Obv. if you bring more CV, you will hit harder (usually), so this would reduce the needed CV by a bit again, because you would hit more than 26.6% of the Soviet elements.
And yes, this example accounts the Soviets as "rolling neutral". It just gives an idea (an idea which I can tell you that this battle (and similarly set up battles) tend to have up to 50/50 win rates when leaders are somewhat equally skilled.

Halt2.jpg (1.18 MiB) Viewed 684 times

Halt_Leaders.png (129.33 KiB) Viewed 684 times

example2.
Another example where I "reproduced" (obv. not perfect/contains assumptions) an engagement in the Editor to see what chances this encounter actually had to be won/lost. Also here - somewhat equally skilled Leaders, single Divisions and we got a 50% success rate for the Germans in the shown 10 attacks (again).
Noteworthy is that the Germans even lost battles where they entered Close Combat.
Furthermore Germans hit ~35% before close combat with 3 Artillery SUs and ~65% after and still failed with this massive fire superiority.
Soviets this time hit at range ~8% and around 15% after close combat.
So German Artillery performed 4.3times better than Soviet ones, and overall performance was also about 4.3 times higher.

Tolbu test.jpg (2.83 MiB) Viewed 684 times

example3.
Just a short PvP example of what can happen (and will), even with good Leaders which are even having a CV advantage. Nothing else to say here but that it can happen and 90% chance means you fail 10% of the time.
Similarly Soviets having 65%'ish roll chances does similarly not mean they will always fail.
The funny bit here is, that the 88 defender CV may actually have been enuff had the German simply rolled neutral (like only losing 2-3CV), if my assumption is correct, that a 1:1.01 Odds will always cause a halt (depends on what exactly the initiative rating is being used for).

Manstein fail.png (843.41 KiB) Viewed 684 times

final remark
I do not think that you can predict a win with 100% certainty in this game as long as you are not either overstacking CV (as calculated) or you are bringing massive amounts of Artillery/Air which hits before the "halt at range X"-rolls take place. Otherwise there will always be a (rather large) portion of RNG messing with your attacks. Especially with the uncertainty of Predictor inaccuaracy (FoW), SUs, Air, RESERVES on top of the definitive RNG of the Leader rolls.
As mentioned, if you add more units to the attacks the RNG factor gets reduced - I could give examples of multi-Divisional cases but let's stick to single for now.

Now You!
Now, what do you think?
Are you affected by preemtively halted attacks - or are you playing lottery already?
What do you do to avoid early halts?
Do you like the RNG?
Do you have another unerstanding of the mechanics (pls share)?

[Wiedrock]

Stochastics1. - single on single Division

Some stochastics.

An attempt to figure where the tested 50% win rates may be coming from.

Initially let's ignore the hitting and the 1.25-rolls and only look at chances of success as a general figure depending on what "/2"-roll-events it is based on.
We assume the Germans having 90%, the Soviets 65% roll chances.
Furthermore we assume that we are talking about an attack that, if not halted, will always succeed (so all CV talks are about the intermediate rolls).

There are 4 combinations of events, based on your setup success and failure of an attack depends on one or multiple of these chances added up.
Event 1: Both do not fail, chance 0.9x0.65=58.5%
Event 2: Both do fail, chance 0.1x0.35=3.5%
Event 3: Germans succeed, Soviets fail, chance 0.9x0.35=31.5%
Event 4: Germans fail, Soviets succeed, chance 0.1x0.65=6.5%

So if we take this and go back to example1.
guess1: We can either say that the 50%'ish success rate I assumed/tested comes from the need for Soviets to fail and the Germans to not fail. Which is 31.5%. If we now add the (so far ignored) usual increase of German CV by some 25% and add the expected 25-30% Soviets getting hit at longer ranges, this may give us the 50%'ish from the tests I did.
guess2: The other way to look at it is to assume it always requires a Soviet to fail the roll, this would put the chance at 3.5%+31.5%=35%. Again, adding the hitting and the 1.25 rolls may get it close to the 50% I've tested.


If we now assume this ~50% chance being the case I have tested (in whatever way it is created), then we may have a look at how big of a chance it is to succeed in multiple attacks in a row using these chances.
The chance to succeed one time in a row is 50%.
The chance to succeed two times in a row is 25%.
The chance to succeed 5times in a row is 0.5^5=3.13%
The chance to succeed 10times in a row is 0.5^10=0.1%

If we'd put it at 75% (by changing leaders/units/SUs) win chance to succeed in multiple attacks in a row would be:
1: 75%
2: 56.25%
5: 23.73%
10: 5.63%

Some stochasics.

[Wiedrock]

Stochastics2. - multi on multi Divisions

Some stochastics.

As mentioned before in this Thread, the impact of failing a "/2" gets lessened with multiple Divisions taking part in a combat.
To give some examples on what to expect, here some chances when attacking with 3 Divisions.
Assuming a 90% roll chance again and ignoring the x1.25 rolls and combat.
Combination 1: 3 succeed (loses 0/6th CV), chance is 0.9^3=72.9%
Combination 2: 2 succeed (loses 1/6th CV), chance is 0.9x0.9x0.1x3=24.3%
Combination 3: 1 succeeds (loses 2/6th CV), chance is 0.9x0.1x0.1x3=2.7% loses
Combination 4: 0 succeed (loses 3/6th CV), chance is 0.1x0.1x0.1=0.1%
As can be seen, the chance to reduce the total final/intermediate CV by /2 is rather low with only 0.1% chance, quite a difference to the single-on-single battles where it would be 10%, which is a 100times larger chance.
If we want we could put this figures into an average loss of -5.05% amongst battles, the single variant loses also -5% on average but understandably there the average has less of a say and results switch between 2 "extremes".
So it is save to say to at least account for 1 possible fail, meaning 1/6th (16.67%) CV lost (ignorng SUs) when you plan your attack (again/still ignoring the other rolls and combat) with 3 Divisions (and without/with only small SUs).

Since Germany isn't playing alone, we should add the Soviet perspective here, since the Leader Rating gap and the missing of Corps HQs (mostly) indicates a rather bleak situation for the Soviets, but with multiple Divisions it doesn't even look that bad.
Assuming a 65% roll chance and applying it to the combinations before, we get:
C1: 27.46%
C2: 44.36%
C3: 23.89%
C4: 4.29%
On average the CV will be reduced by the /2 roll by -17.5% (82.5% left), while in single Division the average would be -27.5% (72.5% left).

Some stochastics.

[Wiedrock]

Alright, before we come to the Multi-Divisional attacks we may have a look at some other types of attacks. The ones where Germans while advancing make a Hasty to save MP and for that are using multiple Divisions against a single Soviet Division which is somewhat buffed by terrain/fort but should be an easy thing to deal with, heh?

I won't get into depth, especially not into SCOUTing attacks (I am not sure about them) but for Hasty Attacks (that are not converted into a SCOUT) in general we assume here that:

• ...the attacker CV always starts at 50% (fact) and can not by any mechanic/roll be reversed back to normal
• ...the intermediate CV calculations and final CV calculations are working exactly the same way as with Deliberate attacks
• ...in about 50% of Hasty attacks, the committed Artillery SUs do not fire at all and in the other 50% only in limited amounts.
• ...the combat intensity is rather low (depending on the Artillery RNG, but way lower than regular attacks), so initial CV need to be larger and everything is more dependent on the Leader Rolls than in deliberate attacks

So, since combat is less intense, we (as Germans) need actually a better CV ratio than we'd need in Deliberate attacks, since we will not be able to hit such high percentages of Soviet elements. The RNG rolls stay presumably the same and their impact is greater due to less intense combat.

example4.
At first I want you to have a look at this pre-combat situation. It is a situation in a Player vs AI in Road to Leningrad Scenario on normal difficulty (so no morale modfiers or so). One of the Divisions was in contact with the enemy during logistics, the other one got moved ~4 Hexes to join the attack. Leaders and the SU situation is shown.
Just make a guess if this is a win/loss or if you think both events may be possible, make a guess to what percentage of attacks may succeed (e.g. 60%, 70%, 80%, 90%). You can also make a guess towards Soviet Air and SU committments and think about if it would make a difference.
MAKE YOUR GUESS NOW (after watching the image, but not any further).

Combatpredictor_1.jpg (833.64 KiB) Viewed 481 times

So the combat predictor shows a 2.14:1 Odds.
I can tell you now that there will not be any Soviet SUs or Air bombing involved. We (as attackers) will have about 20 planes doing bombing each attempt.
ANY CHANGE IN YOUR GUESS?

So here the results come.
The forum is annoying in uploading procedure since you can't copy/paste and only 4 images per post, I will give 10 results in 2 Pictures which were the first 10, but are reflective of overall 50++ tests. (except for one SCOUT I got after 40+ tests, which I reported about in this Thread - but as said we will not worry about SCOUTs for now).

results1-5.jpg (648.65 KiB) Viewed 481 times

results6-10.jpg (644.55 KiB) Viewed 481 times

This first 10 tests contained 3 HELD attacks, this is a bit more than the average, on average amongst the 50 tests I would put the win chance to about 80% for this very example (to keep simple numbers with generous rounding) - but as we know randomness is random at times.
So who put the win chance at 80% has won. Congratulations.

This example does not contain a single "Halted at range XXX" combat and I needed 40+ runs to get the first SCOUT. So why use it in a Thread in which I mostly look at those (missing) "Halted at range XXX Attacks" you may wonder. I chose it because it shows/supports my assumption that as an attacker you need to have continuously at least 1.01:1 Odds when those range check(s) happen (whenever/how ever often they may be checked and in whatever way Initiative plays a role in them).

Fortification as a subject
Important when looking at these results is, that the 10-15 Engineering value is not sufficient to reduce the Fort Level to 0 before final CV are calculated, since the initial Fort Level is 1(+62%), meaning 62/2=31"HP"+1=32 Engineering value would be needed to reduce to to below lvl 1 and we at max got 15.
This is important when it comes to the chances. Since a reduction of a Fort level is quite an reduction in final CV, I personally often use/used a simple multiplicative formula for Forts but was corrected on that by ppetar recently - since when Fort and terrain bonuses interact it's additive, so we can not make a general statement on how CVs are affected by a fort level reduction.
If we look at results of a combat we will not be able to see from the combat screens if the fort was reduced by the engineering value or Artillery, since a won battle always shows the Fort level set to 0 (for whatever reason this was decided to do) and we are lacking the info on the starting X(+XX%) Fort Level. So the results we see here may only/rather be applied to Fort 0 and only carefully to anything containing a Fort (with any Engineering/Artillery reduction shown).

EDIT: I forgot to look at the final CVs and rolls.
Over these 10 Battles I would put the German "/2" failure rate at 2/10 (altough with 2 Divisions it's a bit complicated as explained in the Stochasitcs Posts) and the Soviets failed around 4/10 battles.
This (once again) reflects to some degree the ~85-90% German Roll chances and the Soviet's ~60-65% chances (there are many caveats to Soviet rolls in 1941 but nothing that would change a 60% into a 30 or 40% roll chance, so it's all smaller changes). Some results can obv. be interpreted in different ways if one adds the x1.25 and /1.25 rolls to the conclusion.

Upcomming:
In a following example we will again try to reproduce a close combat result. Germans will have less initial CV and we may see in how far those "halted at range XXX"-attacks appear or not and we may see if the win chance will be lower than 80% and how many HELD attacks we get.

[Sammy5IsAlive]

I guessed 70% - close but no cigar!

I'd normally predict that CV ratio as being a 100% Axis win but I think that low CPP units perform worse in combat which is why in this situation I was predicting that the combat would be less one-sided and the leader rolls would become more significant.

[Wiedrock]

I guessed 70% - close but no cigar!

I'd normally predict that CV ratio as being a 100% Axis win but I think that low CPP units perform worse in combat which is why in this situation I was predicting that the combat would be less one-sided and the leader rolls would become more significant.

Actually a good question you bring up is on how far CPP works same/different between deliberates/hasties (combat performance/SU committments). I leave this to you figuring out.
The additional CV from already 20 CPP each (...or so, maybe even less?) would likely make it a rarity to fail.


...okay I can't go without numbers....
Highest Soviet was 108, with better CPP shooting it may be 100 (veery loosely eyeballing - the testing is up to you!).
Lowest German was 150.

Without CPP they would have had:
96x(1-0.17)=~80CV
72x0,99=~71CV

We need 200CV at least.
Since we will hit better we will need less, but we ignore this to make it more save.

200=(80+71)*X ......../(80+71) and <->
X=200/(80+71)
X=1.32 → So 32CV each to have 200 CV min, while Soviets have 100CV max. (MAKING LOTS OF ASSUMPTIONS - and hoping the math is correct and makes any sense).

[Sammy5IsAlive]

In terms of my understanding of halt attacks. We know that there is the fact d initial CV - this is not affected by leader rolls (if you don't count the commitment of reserve units). We know that the final CV is affected by leader rolls as you describe. My understanding is that these rolls take place at the end of the battle and serve to increase/decrease the CV values of of the elements that remain after battle and then to increase/decrease that total value to reach a final value that is used to determine the battle result.

So what happens in between with halt attacks? I think that a lot of this is in the 'black box' that is undocumented. The manual says that a 'special' CV calculation is performed that is used to decide whether a leader check for a scouting battle is needed. I assume that either the same calculation or something similar is used to decide whether a 'halt attack' check is needed. One important thing to note (I don't know if this fits or not with your understanding) is that there is no evidence or documentation to suggest that these 'intermediate' CV calculations are moderated by leader rolls in the way that the final CVs are. The halt attacks CV calculation is pretty much completely undocumented. All we can really assume is that it calculates a CV that is more accurate (but not equivalent) in terms of representing the likely final CV. So you'd imagine that the calculation goes some way to correcting some of the ways that starting CV might be over or underestimated in comparison to final CVs.

So assuming your goal is to avoid unwanted halted attacks, I would suggest that the skill is twofold. Most importantly to be able to predict final combat outcomes and the likely corresponding final CV values. The second skill is to accurately predict how the 'intermediate'' CV check is likely to go to get the best likelihood of that CV ratio being sufficient to avoid the 'lottery' of a leadership check.

[Wiedrock]

My understanding is that these rolls take place at the end of the battle and serve to increase/decrease the CV values of of the elements that remain after battle and then to increase/decrease that total value to reach a final value that is used to determine the battle result.

So what happens in between with halt attacks? I think that a lot of this is in the 'black box' that is undocumented. The manual says that a 'special' CV calculation is performed that is used to decide whether a leader check for a scouting battle is needed. I assume that either the same calculation or something similar is used to decide whether a 'halt attack' check is needed. One important thing to note (I don't know if this fits or not with your understanding) is that there is no evidence or documentation to suggest that these 'intermediate' CV calculations are moderated by leader rolls in the way that the final CVs are. The halt attacks CV calculation is pretty much completely undocumented. All we can really assume is that it calculates a CV that is more accurate (but not equivalent) in terms of representing the likely final CV. So you'd imagine that the calculation goes some way to correcting some of the ways that starting CV might be over or underestimated in comparison to final CVs.

What I could imagine is that the "final Leader Rolls" we expect at the end, may be actually already be performed at the start of a combat (so in the same moment these SCOUT rolls/decisions/calculations take place). Like, this way the code would be more effiecient I would imagine (like not rolling several times for SCOUT, different ranges and then at the end), but who knows...
And then hit elements are just substracted from that amount and that way "intermediate CVs" ....and later on final CVs are created, this way the final CV result would be the same as if you would roll at the end of combat and you could use the CV already for the intermediate CV-check(s) (in whatever further way/how often/at what ranges the leaders check with their initiative).

But as you say, it's a complete blackbox.

And yes you are right, I pointed out earlier that the use of the Leader Rolls in the ranged halts is an assumption I make, to have "something". This I do because the results look like when you do mod Units to not shoot at all and then attack. Then the final CV (so after actual close combat - without shooting) looks similar to intermediate CV after reduced shootig before those ranged halts (in battles where you do not have many Artillery SUs involved.

So assuming your goal is to avoid unwanted halted attacks, I would suggest that the skill is twofold. Most importantly to be able to predict final combat outcomes and the likely corresponding final CV values. The second skill is to accurately predict how the 'intermediate'' CV check is likely to go to get the best likelihood of that CV ratio being sufficient to avoid the 'lottery' of a leadership check.

1. Not sure why the expected final CV would matter for the ranged halts (if I understood you), because they are affected by all the close combat. So if usually Germans hit harder in close combat and Germans will have 10:1+ final Odds, with your suggestion this would mean that there can not be a 'ranged halt' if a German high Odds battle is to be expected (again,if that's what you suggest). I can say that this is not the case (just working on an example about this odditie).
2. As the Germans to avoid them a higher initial CV is suggested I'd say, but unfortunately also does not offer a guaranteed avoidance (see example3.). This was probably just a single digit % chance to happen I'd guess.
Also in this example I'd expect a high final Odds if it enters close combat, but still it failed, it has +5k Germans, low MOREX Soviets (below 40) and Soviets are massively buffed by the Heavy Woods (so they just have CV, nothing else to offer).

[Wiedrock]

Upcomming:
In a following example we will again try to reproduce a close combat result. Germans will have less initial CV and we may see in how far those "halted at range XXX"-attacks appear or not and we may see if the win chance will be lower than 80% and how many HELD attacks we get.

Continuing.

example5.
Another PvP example we want to investigate if it is one of a long list of basically 100% won battles (not containing a single "Halt at Range") to be evaluated on it's general win chances.
Additional observation, as mentioned, here will be in how far the halted attacks appear (as a comparison to example4.).
Setup:
To make it more clear, you will see in this example that:

• Germans have more CV (CV varies a bit because MOT units have some kind of check before/on way to combat that damages some elements)
• Germans have more SUs (2 additional Werfer SUs being able to be rolled for the second HQ (you see it in the number of Guns))
• Germans have Better Combat Performance (compare the Soviet hit Elements to the Original in the German wins)
• Soviets have worse Combat Performance (compare the German hit Elements to the Original in the German wins)
• Soviets have less SUs being able to be committed (and likely of way worse quality)

As always, no exact information is known in the definitive make up of individual units.
The unknown Corps leader is assumed to be Hube, as he was in charge 3 turns before.
The SUs of the 7th Panzer are also being asusmed to still be the same 3 Pioneers (since the first one is still the same), but in case of Doubt I added another 2 Werfer SUs (Turn 1 Units without CPP).

Hasty_2v1_original reconstructed.jpg (806.37 KiB) Viewed 258 times

In the attachment I did 20 attacks this time to get some better view on the win chance of this individual battle.
What stands out is the sudden (re-)apprearance of "Halts at range", which was the intend with using a "lower initial Odds as German attacker".
I have attached one of the "high Odds" wins at the end to show it's combat performance separate. As one may see a battle that ends with Odds as high as 12.8:1 still is not not a 100% won battle (speaking off Hasty attacks).
The overall win ratio in these 20 tests is 14/20=70%. Unfortunately, with the Attacks in the second halve we had Bad Luck Yet Again several times in a row.
So in summary, this is not a 100% Winning Battle and there is about a 30% chance to lose this attack (with how buffed the Germans are, in the original you probably can add another 5% losses, let's say we ignore this and correct with it the bad luck of 4 fails in a row).
So this is a 70% winning battle which causes Halts at range.
The sole cause are the Leader Rolls (in whatever way they are performed - I pointed out how it could be/is assumed before). Even adding another 5 Artillery SUs would not change that to levels even close to 90%, especially not 100%.
And again, see the Setup's conditions and think for yourself how the win chance changes if we look at the original battle.

Conclusion concerning the "Halts".
To avoid them, I'd say this attack could have ignored all Artillery and better added "CV-giving" Divisions/SUs to maybe reach 160+ initial CV. Since the HELD at close ranges for this encounter seems unlikely (at least with the massive combat performance difference in the test setup - the original likely alsoi could get a HELD at close range (further worsening the win chance on total)).

It's all Editor tests, but I also did 5 Attacks in a "live game", there I got 2 "Halts" in 5 attacks. There I also have de-activated the FoW, so you see average Soviet MOR/EXP now.

example5_live game.jpg (378.89 KiB) Viewed 258 times

In the comming example we may leave the sphere of the Hasty Attacks and take a look at an interesting 2v1, in which we will again buff the Germans, just to be sure.

[Sammy5IsAlive]

What do the artillery combat results look like in those examples? Are there noticeable differences in artillery performance between the halt attacks and the successful ones?

Three potential 'models' of how these halt decisions are made.

1) The decision is made using the 'raw' starting CV (perhaps reduced as elements are disrupted/damaged/destroyed) and a leader roll. This seems unlikely to me - if a halted attack is seen as a 'good decision' based on the raw CVs then the pattern would be the other way around (i.e. roughly 70% attacks halted and 30% continued when the leader fails his roll)

2) The decision is made using a 'special' CV calculation that better reflects the likely eventual combat results, even if the starting CVs odds are poor. So in these examples the 'special calculation' would show that if combat proceeded to close range then a win was likely and so the "positive outcome" of the leader roll becomes the attack proceeding and the "negative" outcome would be the attack being halted.

3) The leader rolls that affect CVs are made at the outset of the battle and these are what are used when checking if a 'halt attack' leader roll is required. So in these examples the halted attacks are a result of the axis leader/soviet leader failing/winning their 'CV rolls' and then the Axis leader wins their 'halt attack' roll to prevent the battle closing to close range.

[Wiedrock]

What do the artillery combat results look like in those examples? Are there noticeable differences in artillery performance between the halt attacks and the successful ones?

The usual no Artillery fire/hitting at all or only very little.

1) The decision is made using the 'raw' starting CV (perhaps reduced as elements are disrupted/damaged/destroyed) and a leader roll. This seems unlikely to me - if a halted attack is seen as a 'good decision' based on the raw CVs then the pattern would be the other way around (i.e. roughly 70% attacks halted and 30% continued when the leader fails his roll)

I'd agree the raw initial CV can not be used from the results we get.
On what's "good"...
For Soviets it makes sense.
For Germans...I imagine it being good in the past because in previous games CV may have been differently created and combat being affected by Leaders (directly affecting CV/Combat(?)), so the CVs more reflected the actual shooting performance.
Or the forces were more equal (WitW) so you could not expect to completely run wild in close quarters, so a "behind" in CV at a certain "range" was more likely/accurate a predictor of a failing attack.

2) The decision is made using a 'special' CV calculation that better reflects the likely eventual combat results, even if the starting CVs odds are poor. So in these examples the 'special calculation' would show that if combat proceeded to close range then a win was likely and so the "positive outcome" of the leader roll becomes the attack proceeding and the "negative" outcome would be the attack being halted.

Ideally, one would guess some type of "elaborate calculation" would be made to better predict the eventual outcome/know what's "good" (so only halting potentially lost attacks). It's actually weird how it's set up (what it produces), either way you look at what's god/bad. But nice mind game switching the good and bad around.
Our leaders prove they are confused as well on what's good and bad, by halting attacks which otherwise always succeed in close combat (e.g.example5.) and keep attacking when there is a chance to lose in close combat (e.g. example4.).
...or it's all just about confusing the enemy.

3) The leader rolls that affect CVs are made at the outset of the battle and these are what are used when checking if a 'halt attack' leader roll is required. So in these examples the halted attacks are a result of the axis leader/soviet leader failing/winning their 'CV rolls' and then the Axis leader wins their 'halt attack' roll to prevent the battle closing to close range.

That's about my current understanding/imagining.
Maybe it's even intended this way to cause some "tactical blunders" with small losses occasionally if Germans attack too weak. Otherwise if German Leaders would "succeed" at not halting attacks and get to close combat the result is more clear and favors Germans massivley.

[Wiedrock]

In the comming example we may leave the sphere of the Hasty Attacks and take a look at an interesting 2v1, in which we will again buff the Germans, just to be sure.

Continuing.

example6.
Again a PvP example to investigate if attacks like this enable us to avoid all "halts at range" or are refelective of a 100% win rate in deliberate attacks over multiple turns or even a whole match.
The Germans will be buffed again, the Soviets weaker than in the original.
We again look at the types of losses we may get and estimate a general win ratio for this buffed attackers battle.
One important thing in this setup to look at is the variance in final Odds.

Scenario:

example6.rar

(1.91 MiB) Downloaded 2 times

Here the scenario I've set up/used. So you can test yourself in the Editor or a live game if you want to. You could test how adding more SUs (on either side) may affect the win ratio. Also worth testing is some RESERVE activation on Soviet side.

Setup:

• Germans CV varies because of the MOT unit RNG. So Soviets have some CV less
• Germans have Better Combat Performance (compare the Soviet hit Elements to the Original in the German wins)
• German hexes are "Light Mud", which leads to more combat performance added due to reduced CV to achieve the same initial CV. The original's German hex ground weather is unclear (attacker's Hex's Ground Weather impacts his attacking CV).
• Soviets have worse Combat Performance (compare the German hit Elements to the Original in the German wins)
• Soviets have 0 SUs being able to be committed
• Soviets have no Air

As always, no exact information is known in the definitive make up of individual units.
The unknown Corps leader is assumed to be Materna, since he starts commanding this Corps.
Since the original was a deliberate attack with 2 HQs involved it is likely for the HQs being further away than 5 Hexes. In the testing setup all the HQs are in range, including all the higher HQs of the chain (further improving German rolls).

example6_org vs results.jpg (1.44 MiB) Viewed 58 times

Above I put 3 different results/Odds of 30 attacks together to have an comparison to the original.

Final Odds variance:
The main thing to me to point out here is the large variance in final Odds. A takeaway could be that generally a 30+ Odd victory is not automatically reflective of a 100% won battle and can still fail at times by the random halts at longer ranges if you attack with too weak initial CV and have no Artillery SUs supporting you.

Win Ratio:
This one, as can be seen is again not a 100% winning battle. With this setup in which Germans are better, Soviets worse off than in the original, the winratio over 30 attacks is 1-4/30=86.67%. Assuming some bad luck in the first 10 attacks we may call it a 90% win. So there is again no indication about an applied idea/trick/knowledge to achieve a 100% win rate for deliberate attacks over many of such attacks.
If you perform such attacks you will fail, at times, unavoidably.

Conclusion concerning the "Halts":
Adding Air and SUs to the Soviets would likely put it more towards a ~85% rate, tipping some close "intermdiate CVs" (...if that's how it works).
For the Germans to avoid these results I'd suggest to always have Artillery SUs in deliberate attacks to overcome such a potential "initial CV minority" (you never know the exact enemy CV before).
...a guess:
With Artillery Fort reduction now working it may also help with the interemdiate CV rolls if you bring some Heavy Artillery (just a guess - needs testing if this reduction is actually applied to those "intermediate rolls" (if they are rolls at all)).