In Tsarst Russia, Survival Puzzles You! -- a Darkest Hour WW1 "AAR" AAR

[JasonPratt]

ARMY HQ: last, and kind-of least. I said divisions are the smallest unit on the map, but I'm pretty doubtful these are actually full division size, though they do have integral combat capabilities like a militia division (or somewhat better). In vanilla DH I have occasionally seen the AI using army HQs for recon, which is hilariously wrong! -- but to be fair, I only saw that in situations where it hadn't brought up many divisions yet.

HQs need a general for best effect, but I'm fuzzy about how the game treats relative rank in this case; in fact I haven't used HQs much yet, since I couldn't get around to building them as Ethiopia and Russia doesn't start with many.

Basically, HQs provide various organizational, supply, and combat bonuses to corps (and divisions) under their command. I'm not at all clear yet how the game parses this command chain; it doesn't seem possible to create "armies" per se, but in effect that's what's happening: an "army" is a "corps" with an Army HQ attached, but HQs give bonuses to other corps nearby. I think. I'll talk more about that as I get more experience with them I guess.

As I learned to my chagrin while playing Ethiopia, HQ tech is upgraded on various land doctrine technology paths.

Wow that was a short entry.  :D

[JasonPratt]

Ack, away from keyboard much longer than I was expecting today, so no work on Russia today.

It may be several days before my next update: if you've read this far, you know I just got finished talking about sorting divisions into corps and corps into armies (maybe, if I can recall how to do that), and shuffling brigades around and into divisions -- and that doesn't even begin dealing with what to produce first! (Or to stop producing as the case may very well be.)

That was no joke back in vanilla DH for Russia, and now we have significantly more than 150 divisions: I seriously don't know how many yet, because I only know we have more than 150 which need upgrading! -- presumably they aren't all behind our current technical curve, though most must be.

Anyway, next time I check in, I hope the western map will show much smaller "stacks" of divisions. In fact, I probably won't check in again until I can show a plot for moving (or starting to move) corps (and armies) into post-initial positions.

[MetalDog]

Is there still a check box for prioritizing which divisions get taken care of first?  I know in HOI2, you could go to each division and prioritize it for upgrades.  That would allow you to spend less IC on trying to get ALL of them up to speed at one time.  It may also be a good idea to look in to disbanding some units.  I am fairly certain that the manpower goes back in to the national pool.  At least a fraction of it does. 

[JasonPratt]

Yeah, all the manpower goes back in the pool (whatever was left in the division) once the division (or brigade) disbands.

Manpower isn't really a problem for Russia, though. I have two different ways to mobilize more already; with mobilization research on the way as well; and if we go to war we instantly not only get full manpower topped up everywhere, but we suddenly get no kidding around twenty thousand manpower. Where one, remember, equals about a thousand troops. In other words, we'll have 20 million troops to send wherever we want. The Germans won't be able to kill any divisions for a long, long time, unless they cut divisions off from supply -- thus the question of how best to abandon Poland so we don't leave any divisions trapped there unable to get out of a pocket!

And admittedly the "AAR" mod may change how much we get; but right now I'm not worried about it.


And yes, I can click on a toggle switch for reinforcements or not, in each division or per corps -- obviously it would be easier to do that per corps, once I get the divisions sorted around! But again I'm not worried about manpower.


Upgrades for divisions, though, that's a tougher problem. I'll be toggling a lot of divisions off (in fact I already did so, for reinforcements, too, for all naval and air units, except some reinforcements for the Baltic fleet), in order to concentrate on upgrading a few at a time as quickly as possible while doing other production things, especially making new arty brigades (and some armored car brigades).

[JasonPratt]

So now I've created a new spreadsheet, imported all the new figures into it for starting Russian infantry and arty and armored techs in the "AAR" mod, and run my comparisons.

Keep in mind that I'm necessarily looking at what's most effective right now, because I don't really have figures for what a tech would be later; and even if I did (presumably they're buried in one of the data files for the game, the main one being literally as long as ALL OF ROBERT JORDAN'S WHEEL OF TIME BOOKS PUT TOGETHER so yeah I'm probably not going to search through that thanks) by the time I researched up to that point some other factors might have also improved which could affect the relative efficiency outcomes.

If I was feeling particularly heinous, I would now inflict a screenshot of the spreadsheet at my reader, but I'm feeling merciful so I'll spare you.  O0


(But obviously I'm not going to spare you from my tedious number crunching FLEE YOU FOOLS FLEE NOW BEFORE YOUR REGRET DESPAIRS YOU!! )


In order to understand my report, it's important to get a handle on what's available for Russia to build; at what tech levels; what the unit stats mean; and how I'm gauging unit effectiveness by the stats.

Russia at the start of the game, a little more than a month before the historical kickoff of WW1, can spend production power (industrial capacity units, IC) and days and manpower building five different kinds of divisions, and four different kinds of combat brigades. (There are also non-combat brigades which help increase some non-combat abilities for divisions, so there's no point gauging their combat effectiveness.)

Those divisions are Infantry-3, Garrison-2, Militia-2, Light-3 (actually abbreviated Mtn-3 because later this becomes Mountain infantry), and Cavalry-3.

(Britain has access to Marines already, but normally nations can't get those until 1922 or thereabouts. Tank brigades haven't been invented yet, much less tank divisions. Motorized, Mechanized, and Airborne Divisions won't show up until much closer to WW2. Ditto Special Forces brigades, which are basically Mtn Infantry.)

Each of those is capable of adding at least one combat brigade (though not all divisions can add the same kind of brigades); and standard infantry divisions can add two brigades of pretty much anything (but not two of the same kind). This increases the possible options for production.

Brigades can also be produced independently, to be attached to existing compatible divisions.

Available combat brigades are: Artillery-1, SuperHeavyArty-1, Cavalry-3, and Armored Car-1.

The -# suffixes indicate the level of technology. Garrison and Militia are researched along with regular Infantry, and occasionally upgrade. The suffix provides a quick reference to relative tech, but in fact some lower suffixes are higher tech than others. Mil-2 is 1897 tech for example (not quite 20th century), but Gar-2 is 1907, which is actually better than Inf-3 at 1903 tech, but Gar-2 is on par with Mtn-3 and Cav-3 for 1907 technology. AC-1 and Arty-1, on the other hand, are 1911 tech. (In this game there are no arty brigades less than 1911 tech, although color descriptions indicate Russia has 1900 arty and 1910 super-arty.)


Each division is rated for attack against hard, soft, and air targets (not any very hard targets right now, but hard attack values kind of suck anyway); vulnerability against defenders, attackers, and air; organization; morale; softness (everything is a 100 percent soft target right now except an armored car brigade and its still 90% soft); how many days it takes to make at top speed (affected by industrial efficiency); how much IC it needs to be produced at top speed; and the manpower which will be deducted to make it. And the supplies (and oil) it will use per day under normal circumstances. (Each division has some other stats, too, but those don't affect combat efficiency in relation to production factors.)

Each brigade is rated somewhat similarly, but in terms of what it adds or subtracts to a division's figures. Sometimes those effects stack, sometimes not: if a brigade (or a set of two for an infantry div) is made along with a division, the brigade(s) won't add to the days made, but will add to the IC and manpower needed.


Attack (in defense against) air, and defense against air, won't really be a significant factor until late in the war if ever, so I'm not counting those in my calcs right now.

Defense values are better the lower they go, since what they rate is vulnerability.

Organization affects how effective a unit is at attack or defense, so higher values are better, and it is in fact possible to be better than 100 per cent organized. Up to that point, a unit is relatively disorganized and this affects its combat effectiveness negatively. In other words a unit at 75% efficiency can only apply 75% of its strength 5 attack against soft targets, whereas its vulnerability on defense or on offense goes up by the same proportion.

Morale may or may not affect attack and defense figures (I'm not sure), but it definitely affects whether the division (and any attached brigade(s)) will rout into another county (or surrender if it cannot retreat).

Morale and Org are somewhat affected by higher unit tech, but usually more affected by other factors (including other techs).

Softness affects how strongly soft attacks will hit against the unit.


Got that so far? Okay.


First I rate combat efficiency, which according to my current efficiency theory means starting by separately summing up hard att, soft att; and def vul, off vul values; and then dividing sum att by sum def (because the higher the def values the worse); then multiplying by the organization percentage; then multiplying by the morale percentage (either one of which being less than 1.00 makes the result worse); then dividing by the soft percentage (which any less than 1.00 is great). Finally I multiply by 100 to give a result somewhere in the range of 1 (to normalize the result so I'm not comparing tiny fractions with each other).

That's basic combat efficiency for a division, which I sometimes call raw power, which I can then adjust based on how adding particular combat brigades affects the raw power. A Inf-3 division for Russia (it might be different for other nations, especially insofar as org and morale are affected) has a rawpow of 1.4132, which naturally any brigade adds to, and any mix of two available combat brigades even moreso. (No brigade, so far, makes any combat stats worse.)

In that sense, Mtn-3 and Cav-3 have the best rawpow values; and they still do even with one brigade added compared to any two brigades added to regular infantry. This isn't so much due to attack values (which are inferior compared to a regular infantry division, and moreso with a brigade mix) as to superior defensive, organizational, and morale values.

One factor I haven't reckoned positively yet by the way is unit size, which can be considered (at least) the hit points to the various armor classes. Unit size is directly proportional to manpower requirements (sort of), so a larger unit has a quality based on its quantity. I haven't figured out how directly the manpower relates to final combat resistance, though; and the game doesn't bother to tell me ahead of time how many troops will be in a brigade or division, much less how many of those troops will have combat strength (horses may or may not reflect combat strength for example). Consequently for want of any clearer way to treat the issue, I regard manpower constraints and manpower contributions to neutralize each other pro and con. But I mention this to register some uncertainty as to whether I'm underestimating divisions (and brigades).

Therefore, the top five division and brigade mixes currently available to Russia so far are, in descending order:

Cav-3 + Cav-3 (4.1098)
Cav-3 + Art-1 (4.0031)
Cav-3 + AC-1 (3.9594)
Mtn-3 + Cav-3 (3.3872)
Mtn-3 + Art-1 (3.1909)

So that's what we ought to be building, and building toward, right?

Ohhh, friend... the darkness of the hour goes deeper than that...  >:D

[JasonPratt]

While those figures can give us an idea what to build and upgrade toward, they don't yet take into account limited industrial capacity and the fact that we need the most efficient troops we can get asap because war is on our doorstep.

Those eliteish divisions take more time to train than regular infantry, and much more than a garrison or militia. We're talking about the difference between 176 days for militia (who have a 10 thousand troop upper limit by the way) and 494 days for light infantry. That's 1.37 game years! (Darkest Hour years are 360 days by the way; all months have 30 days.) We wouldn't see a new light-inf division until Christmas NEXT YEAR! Crikey even a diddly little militia division takes almost half a year to produce.

Which is realistic of course, but gives you an idea how screwed Russia is in vanilla Darkest Hour despite having an approximation of infinite manpower: new divisions just can't be cranked out.

We're also limited in industrial capacity, despite having a very respectable starting IC of 172. Out of that budget, though, we have to keep supplies in production (or buy them) for troops already in the field and maybe soon in combat operations (movement also increases supply usage, and we're dang sure going to be moving lots of troops soon); and also train reinforcements for any short divisions; and also produce consumer supplies; and produce equipment (and training) for any divisional upgrades.

So we don't get all that 172, not even close to it. Maybe around a third or a half if we're lucky.

But a Cav-3 division with a Cav-3 brigade costs 6.3 IC to produce at max speed. Any less assigned, and the production speed drops in proportion -- and it already takes 388 days! (That's like early August NEXT YEAR starting now.)


Taking rawpow (whether for a naked division or with attached brigades), dividing by the fastest days to make, and dividing by IC required to make at top speed (and then multiplying by 1000 to normalize the fraction back to something usefully around 1) gives what I call punchspeed: how fast can rawpower be put into service.


Looked at from that perspective, the top five punchspeediest divs or div and brigade combos are:

Cav-3 + AC-1 (1.7594)
Gar-2             (1.7338) (which has 0 movement)
Cav-3 + Cav-3 (1.6813)
Cav-3 + Art-1 (1.6121)
Cav-3 by itself (1.5441)

Mtn-3 + Cav-3 is still better than any division or div/brig which can actually move (Garrisons can't except by strategic movement, maybe), but its Art-1 brigade though still also a good punchspeed combo can be beaten by infantry + AC, or by inf + cav + AC.


Moreover, before war kicks off Russia is somewhat limited in manpower, and since that's a constraint I could (and on the sheet I do) take the punchspeed and divide by manpower requirements (times a normalizing 10); or divide the rawpower by manpower requirements if I had time to fiddle around but not much manpower to do it with.

(And I could take the punchspreadspeed figure and divide by supplies to get an efficiency at doing all that for upkeep spent per day.)


As it happens, Russia may very well be drawn into the war whether we want it or not -- and if so we may get a ridiculous jump in manpower, so that effectively this constraint can be ignored and all we need to consider is punchspeed.

BUT: as I talked about upthread, in time eonian past , since DH gives Russia an opportunity to avoid supporting Serbia, I plan to take that option and delay entry into the war as long as feasibly possible. If so, even if we spend money (and some tech research) on upping our manpower, we could still have a fairly limited manpower pool to work with; and yet with a world war firing off in Europe (and a Chinese civil war constantly spooling up on our Eastern border, not even counting what Japan might do), we could still be under some time constraints to get effective power on the map.


Looked at like that, in an ideal world war one scenario (where Russia stays out for a significant and substantial time, if not altogether), the top five punchspreadspeed divisions or div/brig combos are:

Cav, without any brigades at all (2.5735)
Cav with AC (1.9549)
Cav with Arty (1.7912)
Cav with Cav (1.6813)
Gar-2 without an arty brigade (they can only get arty as a combat brigade) (1.7338)

Or Militia-2 without arty or cav (for a movable division) (1.7338)
Or Light Infantry without arty or cav (for something not embarrassingly weak and liable to run away ;) ) (0.8937)

After that the best is a plain infantry div with no brigades; and then back to the light infants with cav or arty. Then regular infantry plus armored cars.


So, thinking along that line, we ought to be building Cav divisions and not worrying about brigades, if we're at peace and likely to stay that way for a while yet still want to get our manpower spread out as far as possible as soon as possible in the most effective combat strength possible keeping those factors in account.

Or rather, we ought to be converting what we have into Cav divisions, since that takes less IC and time (if anything, though really about the same if we're cross-converting from other division types) but we'll get some manpower back.


Or,  if we think we're going to be dragged into war sooner rather than later (or at all), we should be upgrading existing divisions into Cav-3 and producing armored car brigades.

And since our IC, though large, is limited, we should start with upgrading any low-level Cav first (since that's quicker than converting a division to Cav), and then once that's done start upgrading any low-level light infantry to Mtn-3; then any militia (lowest levels first if any) to Cav-3; then any regular infantry (lowest levels first) to Cav-3.

Of course any of this might change significantly with tech researches. But the comparisons give us a good idea what to focus our research in, even though we can't be sure infantry (or militia for that matter!) won't catch up being as cost-effective as cav or mtn-inf divisions.

We should research Cavalry (which also gets us improved cav brigades); Armored Cars; Artillery; and Light (soon Mountain) Infantry. Then we can look into tanks and catch up on basic infantry research; and look into researching marines perhaps if we survive into 1922 (or get the prereqs to try a little earlier, though DH punishes early research with tough time penalties.)


Oh, wait. First we should look into how comparatively cost-effective it is to just build brigades at first; because they definitely take less time and IC than full divisions! -- we'd still want to upconvert the huge swarm of divisions we already have (no need to build any new ones for a long time), but adding Cav, Arty, or AC might get more rawpower out on the map more quickly than upgrading divs.

I'll get back with those numbers soon.  :)

[JasonPratt]

Back now.

It's probably impossible to get a real apples to apples comparison between brigades and divisions in how effectively combat power is produced. Brigades don't operate (in this game) apart from divisions, so some of their stats are the same (like fastest days to complete, and IC for full speed production, and manpower, supply/oil, and attack values) while other stats are quite different (defense values, organization, morale, and hard-target effects). The latter group lists as improvements (along with attack values, but they can be treated as a raw stat, too); the former group as raw stats (mostly to do with costs).


This is no small problem, because brigades add fairly small improvements anyway, at rather significant costs; but the improvements are provided in a fashion different than the comparable division stats.

The best I could do in the end was create a sum of the improvements to attack and defense values; multiply that by a sum of the improvements (if any, usually not) in organization, morale, or hardening, expressed as a percentage more than 100 (so for a 2% hardening improvement that contributes 1.02 to the sum for multiplying the sum of the attack and defense improvements); and then divide as usual across the costs (with a little normalization along the way for ease of reference).

Looked at like that, the best brigade production for punchspeed is currently armored cars and then arty (by far, but AC significantly more than arty); then SHArty; then cav at rather lower than sharty.

And those ratios remain constant for punchspeedspread, although if speed isn't a factor only manpower for effectively spreading the raw power increase, SHArty comes first, followed closely by regular Arty, then AC at a distant third, and Cav at a very distant fourth.

But then once continuing supply costs are factored in, the spread goes back to Arty by a more than twice the value for AC, which in turn is almost twice the value of SHArty, which in turn is almost three times the value for Cav. (And supply usage isn't something to sneeze at, as it affects how much IC we can put toward doing other things than keeping our troops in supply!)


What does that mean? Dang if I know.  :))

It probably means that in the very short term, as in WE MAY NOT SURVIVE LONG ENOUGH TO GET DIVS UPSCALED, it's better to create armored car divisions, since they put more competency on the map more quickly even than arty.

But that's only compared with other brigades or making divisions from scratch (sort of, or cross-upgrading). I just don't have proper figures for whether upgrading lower-tech cav (and light infantry) up to date (maybe also some out-of-date brigades for that matter) is a more or less effective use of our time and resources than building armored car brigades.

And even making divisions (with or without brigades already attached) from scratch might well be more time-and-cost-effective at putting military competency on the map than brigades, even though four(ish) more armored car brigades can be produced than cavalry divisions (with brigades) in the same time.


That means I don't really know whether it's better to kit up everyone with an armored car brigade pronto before upgrading their divisions, or upgrading their divisions before producing new brigades for them (which might or might not be AC, could be Cav or Arty depending on our priorities).



....annnnnd now you are insane.

Sorry. When you look too long into the Darkest Hour, etc.

[JasonPratt]

So, I spent today doing the following things in preparation for actually starting Russia's campaign.

0.) Copy-pasted all recent entries into my main doc archive, so I wouldn't lose them accidentally.

1.) Before I forgot again to do so, I split the Baltic fleet into two roughly equal groups (because one group was too much for any one admiral to handle), assigned tactical experts to each of them as admirals, named the new one the Empress Baltic Fleet, and rebased both fleets (plus the sub group) near the northern border with Germany. (They haven't started to sail yet, but the courses are plotted.) Saved the game.

2.) Using the army list I went through every corps, turning off resupply and upgrading for each corps, and sending any brigades attached to divisions back to the brigade pool for potential reassignment. Then, with my spirit crushed by the gears of administration, I went off to punch robots for a while. And saved the game.

3.) Seriously considered getting Shandalar to run in DosBox (or the D-Fend shell more specifically); couldn't recall where I had put Shandalar; went back to Darkest Hour.

4.) Searched all corps for any level 3 divisions of any kind. Found exactly none. Searched for any light infantry divisions of any level. Found exactly none. Wept. On the other hand, unlike vanilla DHFull, Russia in the "AAR" mod starts with all cav and infantry divisions at level 2 instead of some still at level 1! And no militia have been added or substituted. So it could be worse. A lot better but also worse.

This simplifies and simultaneously makes more difficult any upgrade strategy. Everything is almost entirely upgraded; so we don't have to waste time cross-upgrading militia or level 1 infantry to light infantry. And everything is almost entirely upgraded, so everything still needs final upgrades equally and we have no up-to-date troops at all.

5.) My most eastern group of corps, the 4th army, will take until early Sept to get anywhere near where the fighting may have already started by then, and will be so pooped they'll have to rest for weeks, and are at only ten percent strength anyway (except for the HQ which is at full strength). I'm somewhat unsure whether my initial strategic decisions (discussed at length some time ago upthread) will eliminate unrest in Finland, where a significant number of our factories are located. So at the moment I'm assigning this corps (without its HQ, which I'm aiming toward the front instead) to march to Teriyaki, just over the border from Petrograd, from which if necessary I can either assign them individually to anti-partisan duty near factory areas in Finland; or after resting and reinforcing them I can march them south to battle if Finland looks unrestless already. Also, this way I can joke about the territory being quite literally called Terijoki for some reason.

6.) Several "Turkistani" divisions are already staging near the probable war zone, but a couple of corps remain behind at the Turkish border (partly shared with Persia, with whom we also share more of a larger border eastward past the Caspian sea). I don't dare move them in case Turkey gets frisky, but on the other hand most of the divisions down there are pretty well full already. I take the two least full infantries and set them on anti-partisan duties at the center of clusters of factory territories (they'll suppress any revolts in their own and adjacent territories); and divide the cavs and remaining infantries into their own homogenous corps for redeploying a little closer to the Turkish border across a four territory front. This will cede a few counties to Turkey if they go to war, including some factories, but makes for a tighter defensive bulwark.




Oh, and in a burst of inspired experimentation, I learn that I can rebase my air groups beyond their actual flight range somehow; so I send all four squads down to Kutaisi on the Turkish border, in case I need them (since I have no airbases on our western border, and couldn't possibly have any for a year under the best of circumstances.)

Save the game.

7.)  Ditto for a number of small corps in our next most-eastern line, not counting Kiev.




8.) Ditto again all remain corps except the superstacks.

Note that I'm not reassigning brigades yet, mainly because I'll have to run the clock a few days before they become available to teleport around again. ;) But also partly because I need to have a clear idea what my first defensive strategies and groups will be. With so many divisions kind-of ready to go in "AAR", I'm thinking of trying to keep a Polish corridor open across several rivers. But I'll get to that later.

Also note that I'm not assigning reinforcement or upgrade orders yet either. I'll get to that once I have all the corps resorted around homogenously; meanwhile I'm making sure new corps groups turn off supply/upgrades.




Next up, the superstacks.

[JasonPratt]

Catastrophe.

When I loaded up my initial save this afternoon to work on sorting and assigning initial moves to the two remaining superstacks, I happened by some faint premonition to click on another much smaller stack first...

...to make a horrifying tale shorter, none of my move assignments, corps sortings, or even turning off the upgrade/reinforcements... none of that was saved.

And not because I only imagined doing so. To test I quickly rebased my four tac bomber squadrons again; saved the game; reloaded... and their orders didn't save.

And yet, when I turned off their upgrades/reinforcements last week, and saved that -- that saved. That change is still there.


I don't even know. Experimentally, there's no point in re-administrating several dozen divisions again, not even counting the other half of my divisions, if the game has broken somehow and won't save my orders. Maybe this is an extension of the code-bug that cancels any ongoing fights during a save/reload. I don't know.

I know I'm awfully nauseated at the thought of having wasted all this time and effort for nothing, and at the prospect of trying to get it to working again. sigh.

Man. I don't even...

OH CRAP I FIGURED OUT WHAT HAPPENED!!

I was expecting the game to default to the "AAR_Russia_date" format I had first saved the game in, from which I loaded it; but it doesn't, the game always creates a save title based on the nation and date which the player has to add to in order to distinguish it from other saves of that nation and date.

So all that time I was in fact saving my assignments, but to a save name I didn't load from.

And then in my 'test' I inadvertently overwrote my save game, which until then had in fact been perfectly preserved.

So all those organizational tweaks really are gone now, thanks to my inattention and/or stupidity.



....well, the good news is that I can recreate all that, close enough to never mind, and be assured it really will save. The bad news is that I'll have to spend another several hours doing so, just to catch up again to where I had gotten.

I can't handle that tonight, though. yeesh. 

[JasonPratt]

Well, it took me a week to find the time and energy to get back to where I was, but I have now reconstructed my previous assignments, with some minor differences in where various things shall be placed.

Now for the two superstacks.

Essentially each superstack has one of two options: I send them marching out to help line defense, or I create assault groups for trying to take the fight to the Germs or the Huns. Each has benefits; so I decide to send out the larger superstack (in Kiev) to bolster line defense, and create assault corps with the smaller superstack in Sia.

Up in Siauliai, I divvy up the 28 divisions between three infantry corps (one with the Army HQ, though I have at least one more HQ on the way in from Petrograd which should be there by early August), and one cav corps. Everyone gets a 3 skill general (we only have one 4 in the pool, though there may be another few elsewhere, it's hard to tell yet, but he's only a Lt.Gen so I promote him up to full General reducing his skill to 3), except for the most understrength infantry corps whom I give a skill 2 Trickster general. The next most understrength InfCorps gets an Offensive Doctrine general; the strongest gets the Army HQ and a Fortress Buster general; and the cav corp gets an engineering general who will act like he has an engineer brigade attached and help his corps get across rivers faster. I'll also assign an engineering brigade (we only have a few) to the Army HQ a few days from now when my brigades arrive in the pool, which should help everyone get across rivers faster (but the cav won't be dependent on staying nearby for that).

Hopefully we won't need this stack of corps for a while, but when-if-ever we do they'll be ready.




I'm very unsure whether it's better to salt them with our few arty brigades, compared to letting those help with defense elsewhere. But once our few engineering brigs are done helping march far corps in more quickly, I'll send them over, plus cav brigades for the cav corps (since Cav + Cav gives the best raw power and this isn't a production question).

The Kiev superstack I consolidated a little into infantry corps 6 and 7 strong (where the sevens have an HQ, three of which started there), plus a 4 and 3 div cav corps which I ended up sending to the same place so I combined them anyway into a 7 div corps. Once everything kind of arrives in place, or gets close to it, I'll be anchoring the line with infantry corps and keeping my cavs back to help move around for defense. I did send a 7 div infantry corps to Warsaw, as that's the tip of the Polish salient/corridor which I'm willing to try to defend for a little while, so I figured they should be better oomphed, in case I see an opportunity to launch a small superstack counterattack -- though by the same token I've got to be more careful to get them out through the Polish corridor if the CPs seriously hammer Warsaw or rather the counties behind it, as I don't want them trapped without supply and reinforcements.




And that gets me a final initial deployment scheme that looks, overall, like this:




Next up, I think I should try to shuffle my generals around more effectively. Thoughts on that next time.

[JasonPratt]

And then, realizing I hadn't saved the game since adjusting my Sia superstack, I went back, thought I clicked the menu button for bringing up the save, accidentally clicked the unpause button immediately next to it, didn't realize for a minute what the hell was going on, lost a couple of days on the clock, had to quit and reload and do Kiev all over again.

Jesus wept. This game, you guys... 

[JasonPratt]

Okay, fortunately the Trickster and the Fortress Buster generals at the Sia superstack also both already have Offensive Doctrine boosters. All other generals, including at Sia, I have temporarily disassigned from every corps (and saved the game, properly, so I won't have to do that again.)

This will give me a clear full pool of generals to scan through and assign. First, do we have any (more) 4 skill generals?

Yes, one Lt. General who has both Offensive Doctrine and Logistic Wizard skilz. As a LtG however he can only effectively command six divs in a corps; more than that and they suffer various penalties including him not being able to buff them with his skills!

I could simply promote him, but then he'd go down to a skill 3. Since the Sia HQ Army has a general who can command 9 divisions, I add the level 4 LtG to the infantry corps there who needs a general (the Trickster being still in command of the 3rd infantry corps); and I move one of the 7 divs of that corp over to the 8 divs of the Fortress Busting Army HQ, leaving 6 and 9. Nice!

Next, since this guy is a logistics wiz, it makes sense to give him the six most understrength infantries nearby so he can kit them out faster, and shuffle the other divs around accordingly. Done, save.

Next, finding all the level 1 engineer generals and assigning them to the largest and/or farthest out infantry corps. One guy has to be promoted, reducing him to level 0! But they aren't meant to fight, they're just helping infantry get to the dance faster. While I'm at it, though, I find some level 3 engineer generals and assign them to the large infantry stacks heading from the Kiev superstack all the way to Warsaw or elsewhere in the Polish salient; that way if they arrive at a fight I won't be stuck with incomptents and unable to switch them out. Once a few days have clocked off and I can assign my few engineers out of the pool, I'll look into putting level 1 logistic wizards on incoming corps with lots of reinforcement needs.

Save before the next step; and now I'm going to put level 3 logistic wizards on corps with lots of manpower needs near the front -- this may be a moot point if we go to war and Russia gets the instant reinforcement event, but until when/if that happens, they should help. Ideally I'll replace them off with other generals later (probably defensive specialists, maybe some old guard skills whatever that means -- I'm assuming something to do with morale boosts or experience gain boosts.) I don't think we have any cav specialists yet, so I can't shuffle them into place, although we do have at least two panzer specialists!

I still have some corps which need generals when I'm done, but I expect most of them will fold in with other corps as they arrive in place, so at this point I'm done assigning generals for a few game days until I can assign out my engineer brigades and then a few more engineer generals.

Consequently, my initial troop assignments are done for a couple of game days (until the brigades free up).

Next then: assigning tech research.

[MetalDog]

If the definitions for this game are the same as Hearts of Iron 2 (the one I am most familiar with), then the Old Guard trait is actually a detriment and there are no Generals with the Cav trait.

[JasonPratt]

^^ Thanks, MD! To be safe I'll avoid those generals if I can.

I... feel pretty sure I've seen generals pick up a Cav booster trait, though (in DHFull vanilla), from working with cavalry. Maybe that was new to Darkest Hour, or maybe I'm just misremembering. Cavalry weren't a big deal in HOI2, naturally, due to its WW2 setting. In DHFull vanilla, I remember Cav being researchable up to the early 40s (which I thought was amusing). In the "AAR"mod... up until 1982!?!?

:o Logistics and repair don't get researched that long. (Both to 1965.)

Oh well. That brings us to research, which I'm writing up and will post on soon.  8)

[JasonPratt]

Since we're giant-huge Russia, we get seven tech slots, which may be the maximum available. Tech slots are a factor of Industrial Capacity, and if we don't have more than 7 already with 173 (nominal) IC I doubt we're going to get any more.

Those seven slots will fill up fast, though. In theory we want to research (in no particular order of importance yet):

COMPUTERS -- We're way behind the date on this, and can research at least three levels (maybe 4 depending on how long it takes) before running into the future-tech wall.

Wait, have I talked about the future-tech wall? Each tech has a trigger year (starting Jan 1 of that year) which represents all the other related behind-the-scenes tech developments necessary to even begin researching this tech. Which doesn't count the tech requirements necessary which we ourselves can research. Once those are met, we can start researching tech which is ahead of the curve, but the game will sock us with terrible speed penalties, to represent us having to research all-the-other-behind-the-scene-tech first which society would be normally doing in the fullness of time.

So the future-tech wall isn't impenetrable, but until that actual year rolls around we might be better off assigning one of our research slots working on something else not ahead of the curve.

Computers (starting with the census tabulation machine) help boost our research speed, though -- they may be the only tech that solely helps with that, and certainly they're the most helpful. So computers might be one of the slots worth keeping active all the time to work on future tech.

MANUFACTURING AND/OR AGRICULTURE -- mainly these are helpful for Industrial Capacity in various ways, though the Production track helps IC most. Fortunately we're up to date on that already (at 1910 tech). Unfortunately, the next Production tech is 1930, so any IC boost will come from these two; though sometimes they only improve our effective use of industrial materials (so IC doesn't use so much metal, energy, or rares. Or we make supplies more efficiently, freeing up a little more IC to put to use elsewhere.)

1914 CAVALRY and 1914 MOUNTAIN INFANTRY -- probably in that order, since it takes us less time to upgrade our cav-to-cav than to upgrade our infantry to elite.

MOBILIZATION and 1914 ARMY LOGISTICS -- Moby is more important because it allows a division or corps to strategically redeploy. They can be researched together but need separate research slots; on the other hand, once mobilization is researched there are no further developments per se for it (any improvements being subsumed under logistics).

1914 ARTILLERY (field) -- this includes super-heavy arty, but at the moment we're after the normal kind. The next level, which nominally opens in 1916, helps activate tank research.

1916 ARMORED CARS -- these, light arty, and cav brigades, are our best brigades (so far) for adding punch to divisions; and 1916 opens up tank research, too. Unfortunately, this is still 1914 so any research along this line will be greatly retarded for about a year and a half. Better than nothing, but not necessarily better than other things we could be researching instead at a normal speed.

LAND DOCTRINES -- this is a family of unit buffs which we'll be able to research continuously all through the war since we start out quite behind in them: we're only up to turn of the century warfare! This research is utterly crucial to being able to choose a family of interwar, WW2, and post-WW2 doctrinal buffs. It also provides upgrades to headquarters.

On the other hand, we're about caught up on Naval Doctrines even though there's a few things we could still work on until others unlock. But our navy itself sucks right now: we could easily devote every slot to trying to catch up in naval tech, and until we start building new ships that tech wouldn't mean a hill of beans. Ditto for naval doctrines. In short, it's too late to try to do much here until after the war, or unless we have nothing else to work on for a while -- and even then we'd want to be careful we didn't handicap ourselves by strapping a slot in for a couple of years of research on something, since the game pretends all accrued research on a tech instantly evaporates if you de-assign a slot on it for one moment on one day. :P

That means, not incidentally, if we commit to any seven pieces of research we either have to see it through at the expense of any other opportunities opening up meanwhile, or we have to sacrifice a current research utterly in order to start on a new, more important opportunity. There is no setting something aside and picking it back up later to finish out.

Except for interceptors (1914) and seaplanes (1911 and 1914), our aircraft are up to par, we just don't have many; and our Air Doctrines are even more up to par, missing only Naval Support for now (which adds a few buffs and opens up an option to attack ports.) Since I don't foresee building even many airfields, much less aircraft, for a few years, this is a low priority.

Last but not least are Secret Weapons, but the earliest such project is experimental tanks in 1916, which won't open up until 1916 armored cars and artillery are researched. Amusingly, the "AAR" mod adds (among many other things) Computer Wargames in 2010, called "Hoiception", heh.

There are a few other things we can currently research like hospitals, oil, regular infantry, and anti-aircraft guns, but I've already covered the most important ones for our purposes. Again in no particular order

LAND DOCTRINES
1914 ARTILLERY
1916 ARMORED CARS (but can't do that yet)
MOBILIZATION
1914 LOGISTICS
1914 CAVALRY
1914 MOUNTAIN INFANTRY
MANUFACTURING
1915 AGRICULTURE
COMPUTERS

Not counting ACs (since we can't do those again until 1916 anyway, or not feasibly), or Agri (can't feasibly start for another six months) that leaves eight potential topics for seven slots.

And if we treat mobilization as more important than logistics per se (so we can get strategic redeployment), that gets us our seven slots.

Not every researcher or institute (the game doesn't distinguish between them except for color sake) is specially good at everything, though, so even in that list we may have to make choices about whether to go faster with one research rather than another. That's certainly true for Russia in vanilla DHFull; possible the "AAR" mod adds more researchers, I don't know yet.


But that means it's time to talk about how research is done instead of what is available and ought to be done first.

Let's take 1914 Arty as our example. Any research requires five stages in a particular order; with general research types (often though not always given particular flavor texts) assigned by the game (and/or mod) developers to each stage. Those types might or might not be repeated from stage to stage; a tech could even have five stages of the same type, or four and then a different type for the final stage. 1914 Arty, like all early artillery techs, requires research in artillery (duh), mechanics, munitions, electronics, and training. And if I assigned Igor Sikorsky to it (since he won't be working on aircraft for a while anyway), the page would fill in some color details like this:



But that's just flavor text, the general categories are what count. And any researcher (or institution) can work on any research; but if they have specializations in a category type then they'll get some speed bonuses when working on that type. So Igor will work faster than he normally would on the arty, mechanics, and electronics stages, but he'll work at normal speed during the munitions and training stages. Then again, he has skill level 7, which is freaking genius (you may have heard of Sikorsky ;) ). I don't think that means he works seven times as fast as a level 1 researcher, but he's certainly faster by some proportion.

On the other hand, you can see little numbers next to the research types in each stage, 2 2 2 2 and 4 in this case. Artillery 2 is harder to work on that artillery 1, but not nearly so much as artillery 12, or training 4 for that matter (other things being equal). As tech projects get more modern, the level of research types required goes up. Unfortunately, unless the "AAR" massively revises this (which would be great though I haven't heard so), the game doesn't track completed research in a category, so if some other tech also requires Arty 2, that's too bad it has to be done all over again. Nor does having completed Arty 2 in any way help complete Arty 3 or Arty 12 or even Arty 1.

Another highly annoying factor which the game keeps completely hidden, is that sometimes (not all the time but sometimes, and increasingly often as techs get more modern) the final research stage will need a completion of 200% not 100% in order to finish. In other words the final stage will go twice as slow as it otherwise would have after accounting for all other factors. This is totally aside from whether, as in this case, the final stage happens to have a category twice as difficult as the previous ones.

On yet ANOTHER hand, we can buy or be given Blueprints for techs (and any tech we complete we get blueprints for which we can sell or trade). Blueprints tend to speed up research twice as quickly as the basic rate, plus or minus other factors. We can also get blueprints semi-randomly by investing national money in research projects (which also helps reduce dissent a bit, by the way). Research projects also give a small permanent speed percentage increase to all research henceforth, even if no blueprint is generated. Research projects are great strategic decisions, and you can dang well expect to see me spending money on them when feasible! :D

Finally, computer research (and maybe a few others in very limited ways) adds a permanent speed bonus to all research whenever completed. It isn't much, but it adds up like compounding interest in a stock account.

Obviously then, assigning someone awesome to COMPUTER research, now and probably forever (despite the tech wall), is the way to go. And we like Arty brigades, so having Sikorsky working on 1914 ARTILLERY right now is a great idea (maybe permanently despite the tech wall).

Sikorsky would be a plausible choice for this, too, but he has a bunch of other specialties which will be useful elsewhere. Who we really want is Havannes Adamain who is a skill level 10 researcher and only specializes in electronics and mechanics.




As you can see, early computers (we don't even have late 19th century ones yet) are ALL ABOUT MECHANICS (and later electronics and math), though different levels of mechanics at different stages (5 4 4 2 10, so not in increasing difficulty either). It's too bad that, unless the "AAR" mod makes major changes to how research works, no institution or researcher can gain new specializations by working on projects with categories outside their field; nor will they ever level up in any other way. Sik will always be skill 7, never skill 10; and no matter if we have Hav researching computers for the next ninety years, he'll never learn to specialize in math. Similarly, we're assigning the Obukhov Plant to work on our cavalry, because they're a skill 7 institution who specialize in two of the categories (munitions and artillery) instead of the Russian cavalry genius Brusilov (also skill 7 with tons of specialties but only one, cavalry tactics, which applies to 1914 Cav tech) or the Tula Arsenal (which specializes in three of the five categories needed for 1914 Cav but which is only skill 3).

Meanwhile, we have exactly no one who specializes in mountain warfare (and may never get anyone -- sometimes researchers activate over time, though, or deactivate as they die off or the institutions change, and we can pick up new ones if we conquer a nation). And mountain infantry tech research (unlike cav research!) mostly requires mountain infantry tech. Out of the other two techs, general equipment and small unit tactics, we have no one better than Cavalry genius Brusilov.

So to recap so far: our best aircraft guy is working on artillery; our best artillery designer is working on cavalry; and our best cav guy is working on mountain/light infantry.

The Darkest Hour research protocols are weird as hell.

1910 PRODUCTION requires some less straightforward weirdness, meanwhile. The tech stages are:
5 management
5 management
5 manufacturing
6 manufacturing
10 management

Our two best options are Kalinin Polytechnical Institute, which has a basic skill of 9 and specializes in manufacturing (among many other disparate things -- they'll be great later for other purposes), and the Vyatka Factory Design Bureau which has a basic skill of 6 and (among other things) specializes in management.

If we go with Vyatka, their basic speed will be less, but they'll have speed bonuses on three out of five stages. If we go with Kalinin, their basic speed will be more, but they'll have speed bonuses on only two out of the five stages.

Oy.

I'm not sure if this is exactly the right representative relationship of research skill and specialization to the tech difficulties used in stages, but even if I'm wrong I think I'm pretty close: a researcher's speed on any stage (not counting the hidden possibility of a 200% requirement on the final stage) can be expressed by dividing their skill by the tech type difficulty; multiplied by 2 for specializations. Then adding up the results. The higher number will be speedier. (This is modified by blueprints, the future tech handicaps, and basic national research speed, but those factors apply in any case regardless.)

So Kalinin, skill 9, special manufacturing =
9/5 management +
9/5 management +
9/5*2 manufacturing +
9/6*2 manufacturing +
9/10 management = 11.1 or 10.65 if the final difficulty triggers.

Vyatka, skill 6, special management =
6/5*2 management
6/5*2 management
6/5 manufacturing
6/6 manufacturing
6/10*2 management = 8.2 or 7.6 final difficulty.

Kalinin it is.

(I did the same calcs for a couple of the previous picks, but just didn't talk about them because the answer was obvious "by eye". By which I mean I didn't do the same calcs for any of them because the answer was obvious "by eye". By which I mean I may have made a mistake on the previous ones, though I doubt it, but if so it can't be very significant. ;) )

There's a similar issue for Mobilization:
Vyatka Factory Design 6.833 -- they only know one relevant skill but it's important.
Nikolai Yudanich 5.97222 -- they know two specials but are a little less skilled.
Zhukovsky Air Force Acadamy 7.667 -- also two specials and as skilled as Vyatka.
Any level 8 researcher 8.9333 -- even with no specials! Might as well put the ace pilot Petr Nesterov on that. ;)

That leaves LAND DOCTRINES for last. Here we can go with either Leading By Orders, or Leading By Task. They both lead to the same chain of doctrinal techs afterward; one gives a morale buff and one gives an organizational buff, each of those important for how long a division (or air or sea group) can stay in a fight before routing, and for how effective they'll fight while in it. Organization, once researched, will affect our combat calcs for the most efficient divisions/brigade combos, but not by a lot. The Organization buff gives more effect than the Morale buff; but the general techs are somewhat different and so one might be researched faster than the other allowing us to get on with catching up on Land Doctrines more quickly and that should probably be our guide since the difference in Org and Morale increase isn't much (6 and 4 percent).

For Leading By Orders, which gives the fastest (probable) speed? A skill 8 math whiz who doesn't have any specials here; a skill 6 academy with two specials; or a skill 5 soldier with three specials (and two repeats for the general or the academy)?

Alexander Mathwizovich skill 8 no specials
8/3 infantry focus
8/3 small unit tactics
8/3 centralized execution (of plans not prisoners ;) )
8/3 centralized execution
8/4 training = 12.667 (or 11.667 if final difficulty)

Zukovsky Air Force Academy, skill 6, two specials (training and centralized execution):
6/3 +
6/3 +
6/3*2 +
6/3*2 +
6/4*2 = 15 (or 13.5 if final difficulty)

Nicky Yudolich, skill 5, three specials (infantry focus, centralized execution, training):
5/3*2 +
5/3 +
5/3*2 +
5/3*2 + (these two were both centralized execution)
5/4*2 = 14.1667 (or 12.916667 if final difficulty)

Zukovsky Air Force it is. If we go with this doctrine.

What about Leading By Task?

3 Infantry Focus
3 Training (which is weird, I would have thought the logic went the other way around: more training for teams to work independently)
3 Decentralized Execution
3 Decentralized Execution
4 Large Unit Tactics (again, weird, I would have expected large and small units switched around, but whatever)

Alexander Mathwizovich, skill 8, no specials:
totals will be the same as the relative tech categories haven't changed difficulties (3 3 3 3 4) and he has no specials to affect the speed, so 12.667 (or 11.667 if final difficulty).

Nikolay Yudenich again, still skill 5 but only 2 specials this time (as he knows centralized not decentralized plan executions):
5/3*2 + (infantry focus)
5/3*2 + (this is the stage for training now)
5/3 +
5/3 +
5/4 (he doesn't know large unit tactics) = 10.13889 (or 9.5139 if final difficulty)

Moscow Military Academy, skill 4 with 3 specials (training, decentralized execution, large unit tactics):
4/3 +
4/3*2 (training) +
4/3*2 (decent ex)+
4/3*2 (decent ex)+
4/4*2 (large unit tacs) = 11.333 (or 10.333 if final difficulty)

So:
Alexander Freidman (hey that's German isn't it!?!) the math whiz, doing Leading By Task at speed 11.667 (at worst)

vs

Zukovsky Air Force Academy, doing Leading By Orders at speed 13.5 (at worst).

Air Force wins. :)

And now my brain hurts. (save game)

But here's our final initial research plots:




Whew.


ARE WE FINALLY READY TO WAGE WAR???!

No.

And we're trying to avoid war anyway. But we still aren't ready to start the gameclock.

Still on the to-do list:

assign initial production (maybe with some tweaking after the clock starts);
make initial strategic policy decisions (previously somewhat discussed);
do the initial social slider move if any (that'll be quick though).

I can probably do that in one more post. And then we'll be set to really start the game clock.