AGEOD Forums

A couple of new games with the gold version (1.08I official). My machines are all reasonably high end. I run Vista on a powerful desktop and Win 7 on a top line laptop. No crash bugs (at last!), and the game runs fast and smooth almost all the time. Indeed, it is not necessary to put mouse sensitivity all the way to the top, if anything it is too sensitive.

I ran both games as CP on WEGO. One was the 2-player GC, the other the 1916 scenario.

On the GC I went Moltke with Diplo poker and got Italy on my side. The AI fought back well initially on the diplo front, getting England in and trying to stop me from getting more Balkan powers on my side (e.g. by giving Bulgaria concessions). Unfortunately for the AI, the crucial for my plans Romanian ambassador duel ended up in my favour and the AI then stopped even trying. From 1915 onwards the AI didn't run any diplo missions at all -it just vegetated and watched me get Bulgaria, Turkey, Persia and China on my side. I repeat: the AI didn't run any missions at all after the first 3-4 turns. Obviously with so many allies I won easily. Russia capitulated (straight up, no revolution was needed) on July 1915. I stopped the game then as too easy.

The AI tends to keep a front line much better than before. E.g. France now holds the south tenaciously against the Italians, and my assault petered out after I got Savoy and the coast out to Marseille and faced an unbroken line of French defenders. Russia also tried hard to keep an unbroken front line but ran out of troops.

This was because on the attacking side the AI is too reckless. It went for WW2 style deep penetrations: the French 2nd Army brpke through and reached the outskirts of Mainz, while the Russians made two massive incursions in Galicia and East Prussia (where they got to Danzig). In all cases however, the penetrations were not properly flanked so all these forces were easily cut off by me and the troops isolated. I like the aggessiveness of the AI but it must try to attack on a broader front, or at least hold back the breakthroughs if the flanks are not following, if the WWI experience is to be simulated better.

The 1916 scenario is much more realistic by comparison, since it is much tougher for the attacker and such deep incursions are impossible. There are some strange occurences though:

-Belgium capitulated on the 2nd turn although I had not done anything in that sector. There must be a special rule for Belgium not capitulating at all (like Serbia) in my opinion.

-Similarly, Luxembourg (on the side of the CP!) also capitulated in the 2nd turn with no prompting by anyone. This seems to be like a bug. What is the point of Luxembourg anyway in this game (in 1916 no less)?

-Both my Grand Offensives (Verdun and Trentino, I didn't change anything) succeeded easily on their first turn when the defender retreated after the 4th or 5th round of battle. It's too easy. Historically these were slautherhouses. Also, although I repeatedly tried to stop the offensives afterwards so as to plan new ones, the game ignored me.

-The AI sent a load of penny-packet attacks against my well prepared defenses and of course it got slaughtered. Also it redeployed its troops too much and spread them out thinly, so that when its offensives came around (Somme and Brusilov) they were laughably weak (2-3 corps and no bomabrdment to speak off). Perhaps it should be made to focus on one massive grand offensive during the Firepower doctrine period, and hold back on the penny-packet stuff?

In general the AI, although improved, behaves rather oddly. In short it goes bonkers on the aggression on the strategic front but is rather spineless tactically. E.g. on the 1914 GC I sortied with the HS Fleet and met the Grand Fleet in a major action. I had surprise on my side, but the British shot much better than me and sunk 7 of my ships to 3 of theirs. But, although they had me, they retreated!

I'll try the 1918 scenario now on the 1.08K beta to check that out.

Thanks for the feedback!

Much has been improved on AI's handling of Major Offensive battles, its not always perfect of course, but on the defense in theory it should hold out until it has no other way than to retreat. Dont forget if you apply flash bombardment succesfully (also perhap inflitration, I am not sure), the defenders cant really get any reserves to the battle prior to start, so it IS possible to force breakthrough, if you can wear them down faster than they can bring new troops in.

As for letting its offensive be outflanked, yes, but also I have seen occassions when such advancing thrusts stopped and turned back to counter enemy armies. So there is hope this can be further improved.

HOWEVER, that behavior is often the result of the WEGO system, I believe.

Of course, the AI is clearly performing less optimally in army-by-army mode.
I THINK I finally managed to spot the reason for this, and already notified Calvinus about it. It appears (please deny this if isn't true), that if the AI fails to coordinate a given number of Armies with the active army, these armies it wanted to coordinate will not actually move during the turn. Yes, they will be activated separately, but they will be left in place.
This would explain a lot of things I have seen from the a-by-a AI behavior, and I sure hope that this is the case, because then the a-by-a AI could be raised to the level of the WEGO AI, and I would stop playing WEGO altogether.

Tamas wrote:
Of course, the AI is clearly performing less optimally in army-by-army mode.
I THINK I finally managed to spot the reason for this, and already notified Calvinus about it. It appears (please deny this if isn't true), that if the AI fails to coordinate a given number of Armies with the active army, these armies it wanted to coordinate will not actually move during the turn. Yes, they will be activated separately, but they will be left in place.
This would explain a lot of things I have seen from the a-by-a AI behavior, and I sure hope that this is the case, because then the a-by-a AI could be raised to the level of the WEGO AI, and I would stop playing WEGO altogether.

Fully spotted, 100% true.

Not only that, but several time I have seen the AI activate an Army via Reaction during the resolution of my Movement phase, then not do any movement with that Army.

AI problems in army-by-army mode for coordination and reaction are well known. I just have to find some time and mainly some force to work on these two points. Unfortunately the hot hot weather of these days does not help much!

Hi Calvinus,

First, I could have sworn I read a post that said that the AI was better in A-A mode?

Second, since we're talking about the AI, and recognizing your valuable time and the fact that it's the AI afterall, I still had two curious questions:

What is the British Grand Fleet doing in the South Pacific?

What are British heavy tanks doing in Africa? Seems to me they should be in the desert supporting attacks there against the Turks.

Tamas wrote:Of course, the AI is clearly performing less optimally in army-by-army mode.
I THINK I finally managed to spot the reason for this, and already notified Calvinus about it. It appears (please deny this if isn't true), that if the AI fails to coordinate a given number of Armies with the active army, these armies it wanted to coordinate will not actually move during the turn. Yes, they will be activated separately, but they will be left in place.
This would explain a lot of things I have seen from the a-by-a AI behavior, and I sure hope that this is the case, because then the a-by-a AI could be raised to the level of the WEGO AI, and I would stop playing WEGO altogether.

I'm investigating about this AI inactivity in Army-by-Army mode, though unsuccessfully for the time being. I hope I'll be able to find the cause for when the next patch is to be out.

bk6583 wrote:What is the British Grand Fleet doing in the South Pacific?

What are British heavy tanks doing in Africa? Seems to me they should be in the desert supporting attacks there against the Turks.

Surely two weird strategies! Sometimes it's really difficult to understand what AI thinks and does!
However, I'm pretty sure such things do not happen regularly. So I'm not worried about that.

One more CG game, this time with beta 1.08K. CP with WEGO again. Slight variation on my favourite "what-if" war plans: Moltke/diplo poker/munitions for Germany with Prinz/Ultimatum Accepted/Munitions for Austria. I "won" (stopped game for being unchallenging) in August 16.

I am much happier with the diplomatic AI this time. It got Britain in first, then Romania, while I was getting Turkey and Bulgaria (Italy I got on my side again through the diplo poker ploy). I took advantage of the ambiguous Greek diplomatic stance when I suddenly realised in early 15 that their diplo level was 7 (!) and pushed them over to my side with a couple of well-placed diplomatic missions. The AI fought me very well on the blockade neutrals front by targeting Sweden and Holland repeatedly, while getting in a couple of missions to Spain as well. My only complaint remanis that in all my Gold games (all patch versions) the AI makes no effort whatsoever towards Japan which stays neutral in 100% of cases. Compare that with the vanilla where Japan was always number 1 priority. I suggest a slight change in the algorithm to ensure the Entente AI to target Japan after Britain, Italy and the Balkan minors. Oh, and perhaps try to get a right of passage through Belgium (I'm not sure on the rules there) to extend the front there (since I always go East, Belgian neutrality is a very nice way to keep my few Western forces concentrated, and an elongated front would cause me many problems).

Strategically, it was a decent showing by the AI on the West Front, keeping a nice tight line in the South once the first Italian rush was spent. In fact, clearing the French out of the Savoy fortresses took me all the way to 1916. On the main Franco-German border, I resisted the first onslaught in 1914 by the skin of my teeth and then it settled down to a continuous front, more or less. Some problems remain:

-Occasionally it still leaves the edge hexes of the front (Briey, Belfort) unoccupied.

-It never launches a concentrated GO during the trench period, with lots of guns and troops, but goes for penny-packet efforts all over the line and obviously gets slaughtered. It would be much more proper from a WW1 simulation perspective if it remains on the defensive everywhere apart from the main offensive sector which it should pack with troops

-All my Grand Offensives on the other hand always succeed. Perhaps I am good in concentrating strength and picking the right spot, but I wouldn't mind getting slaughtered and frustrated just once. It's WW1 after all, slaughter and frustration is supposed to be the whole point! As it was I broke through Belfort and then steadily through Burgundy and got all the way to Lons while I rolled up the French line in Nancy and isolated the troops at Epinal and the Vosges hex. Nice, but a bit WW2-ish. I want more hard slogs, not sweeping encirclements in 1916!

On the main East front the AI went for the main East Prussia effort (Czar plan?) and again got badly isolated in deep penetrations. It fought well, but too many troops were lost to isolations and eventually Russia revolted and capitulated in two successive rounds in summer 1916 with my troops outside Vitebsk and inside Dvinsk.

I was really happy with the AI on the Serbian front. I got my 3 Austrian army noses repeatedly bloodied (the Serb troops are really tough) and it was only through the Bulgarian and (eventually) Greek backstabs that I killed them off. Perfect WW1 simulation here.

Similarly, the Romanians were also accurately useless. With a bit of help from the Turkish 1st Army (I risked leaving the Straits uncovered as Britain seemed to be massively commited in Mesopotamia and Armenia (more on this below), plus I have NEVER, EVER, EVER seen the AI go for a Gallipoli gambit) I killed them off by 1916. Accurate again.

Britain really made my life tough against Turkey. My forces in the Sinai were even nicely encircled by a push up from the Red Sea (I got careless) and most of my IV Army got isolated and destroyed . In Mesopotamia I thought I had bottled them up North of Basra, using those lakes or whatever they are to keep the line short, but the marvellous AI went through Persia (which was on my side along with Afganintan) and over in Armenia! Oops. It was even advancing there when I finally beat Russia and rushed over the Turkish 2nd and 3rd armies along with German and Austrian expeditionary forces to stem the flow when Britain capitulated! (see below)

General occurences and apparent bugs:

-Technology goes a bit fast. I ran out of advances in mid-1916. Granted, I had three majors on my side researching 5 advances at a time.

-On the other hand, doctrine advances are infuriatingly slow. I only got gas, and that was in 1916. Since there is a limit of 2 per year anyway, perhaps increase the probability of success? The "Country X failed in adopting the following doctrine:" message gets infuriating, particularly as ALL my GOs (even the Turkish ones) succeeded.

-Whenever my fleets venture out to sea, even seas controlled by me (e.g. the Italians in W. Med, or the Turks just outside Bosporus in the Black seas) they immediately get wounded (usually the best ships too). I may be missing something in the rules about enemy subs/mines (although there were no enemy minefields in view), but it is annoying even in that case since there is no message of enemy activity causing the damage, it's as if my fleets magically fall sick when they get out of port

-The AI makes the silly mistake during the redeployment phases of sending troops to isolated but controlled hexes. With all the deep penetrations the Russians did, and the subsequent getting cut off by me, there were always lots of enemy hexes at the back of my front that my cavalry did not have the time to convert back. Well, every redeployment phase these isolated hexes got filled with troops again, which were promptly isolated and died. Perhaps put a rule, that a stack can't move to a hex which will be out of supply/isolated immediately.

-Isolations and poor supply are a problem even for prudent advances sometimes. In all Turkish fronts supply is awful. Fair enough, but the British AI in particular loses so many stacks through isolation when it is doing even historical things like trying to get up to Bagdad methodically, one hex at a time.

-National Morale and Parliament movements are completely bizarre. I would call them a random walk: there seems to be some causal drift due to the fortunes of the war but the variance is tremendous. France, with the South, Savoy and Burgundy lost, with Russia out of the war and with oodles upon oodles of losses was still Sacred Union/20-odd morale in 1916. Ok, let's put that down to tough Frenchmen wanting to get Alsace-Lorraine back so much they absorb punishment without demurr. OK, so the Austrians and Italians get disillusioned amazingly quickly despite costly victories in the front - that's also fine. Britain on the other hand was manhandling the Turks as above, plus kicking my butt every time the HS Fleet ventured out (I was exchanging capital ships 1-1 but lost so many cruisers and destroyers it wasn't even funny). OK, so my subs got 90 and 110 points out of their economy on the 2 phases I ran them (restricted), but I would say Britain was doing rather well overall. Well, Parliament went defeatist around the end of 1915 and morale around -7 by summer 16. It didn't ever try to form a new govt (only France did, once, it seems the AI doesn't like the option in gold as much as it did in vanilla) and promptly capitulated!

-Bizarre messages when it comes to capitulations remain. Romania capitulated first (got the message), declared war to me again (!) in the same sequence and ended up with Right of Passage to Entente diplomatic level. The same happened with Afganistan (I said the Brits were doing well): surrender, war, surrender, war in one sequence. Russia capitulated properly, Britain however went through the same bizarre cycle.

-After Russia surrendered, in the next interphase I got a series of messages (the yellowy-paper type, not the white-newspapary sort) for each of my countries (Germany included) that "Due to the capitulation of major power GERMANY (or Austria, or Italy etc) the morale of Germany (Austria, Italy etc) has been affected by +x points." Despite my horror, I had obviously NOT surrendered and all the morale bumps were positive so I suspect this had something to do with the capitulation of Russia.

-Peace treaty messages still have this ugly "none was ceded to the winning powers, none was ceded to the winning powers, none was ceded to the winning powers, none was ceded to the winning powers, none was ceded to the winning powers" message.

OK, I know this was a long post but this game for me is so fascinating in its potential, that I really want it to be as perfect as possible. I can provide save games for every month of this game if it is required to help locate the bugs.

Agree with most everything written by PK205 and echo the observations regarding sometimes bizarre AI behavior.

I have yet to see an American army actually fight anybody anywhere, one game they sat in French Africa and another they made it to France where they sat in Sedan, far from the front lines while three corps partied down in Britain for the duration.

In my last game the Japanese army landed in Suez and sat there inert while the German's overran French Indo-China and Singapore!

In my current game the German AI has sent an army of African Askaris across the South Atlantic to the Falkland Islands, an action even more more rediculous than it is implausable. (see attached .jpg)

German Naval units cannot be intercepted and can exist at sea forever as raiders without the benefit of friendly ports. I would argue though that the naval attrition is not too much out of line but the agent of that attrition should be more clearly presented to the owning player rather than being lost in any of the 50+ pop up messages that appear in a typical turn.

I am less concerned with the tech tree as is and submit that in the actual event, most of the techs represented in the game were known by the end of 1916 and the last two years was spent refining them and finding effective doctrines to use them, something the game reflects pretty well given average die rolls.

For all that, La Grande Guerre 14-18 Gold is a superb game and the support has been outstanding. Note that recent patches have cured the conquest of colonial Africa by Germany and noticable improvements in the enemy AI in the 4-Person campaign have been implemented even if one still sees the Kaiser's African Askaris in Poland fighting Russians on occasion. Perhaps they tired of the Falkland Islands.

Overall however, am very happy with WW1 in its current state even if sometimes frustrated by some of the niggling details and the occasional inability to execute a historically reasonable operation. Have learned to hate the red text in the message box when it says what I am trying to do is impossible for reasons that appear to have no bearing on the actual situation. Still, so much of the game has been done right that the minor details stand out in sharp relief, perhaps more so than is reasonable to expect.

Cheers

Normally, CP are not allowed any transport mission outside European waters

PhilThib wrote:Normally, CP are not allowed any transport mission outside European waters

The solution might be just to lock the German colonial garrisons in their starting locations, except maybe those in East Africa after Portugal enters the war or by event. Germany never had any realistic prospects of increasing her colonial holdings during the War and even with von Lettow-Vorbeck's successful offensives into British and Portugese territory, he could not hold German East Africa in game terms.

The AI makes the silly mistake during the redeployment phases of sending troops to isolated but controlled hexes. With all the deep penetrations the Russians did, and the subsequent getting cut off by me, there were always lots of enemy hexes at the back of my front that my cavalry did not have the time to convert back. Well, every redeployment phase these isolated hexes got filled with troops again, which were promptly isolated and died. Perhaps put a rule, that a stack can't move to a hex which will be out of supply/isolated immediately.

Brother let me second, third, and fourth this! It's ruining my current game as the German AI in France just redeployed its GHQ with over 6 corps, air, and arty to an isolated area - all doomed to die an isolated death. I've never expected greatness from AIs but I also certainly expect that stupidity has at least been weeded out. This is really stupid. Really excellent summaries above though that will give anyone pondering buying this game food for thought one way or the other.

bk6583 wrote:Brother let me second, third, and fourth this! It's ruining my current game as the German AI in France just redeployed its GHQ with over 6 corps, air, and arty to an isolated area - all doomed to die an isolated death. I've never expected greatness from AIs but I also certainly expect that stupidity has at least been weeded out. This is really stupid. Really excellent summaries above though that will give anyone pondering buying this game food for thought one way or the other.

Indeed this is a point to work on!

OK, I ran 2 more games on WEGO, CP, 1.08K beta. Difficulty and AI aggresiveness are both set one notch below maximum. The two games were the 1916 and 1918 scenarios respectively. Generally, the same impressions as above. The game is great fun and very stable, and for me the biggest problems which remain are:

-Very quick capitulations of countries. Romania in 1916 capitulated sort of normally after I battered it from 3 sides, as in Real Life, although historically it took up to 1918 for the Romanians to do so although they had lost 90% of their country in 1916. The Russians also capitulated in late 1916 without real change in their circumstances from the beginning of the 1916 scenario- historically I was concentrating west. The AI needs to be made more tenacious. WW1 should be tough to win.

-Buggy capitulations. One thing is the weird message sequence "surrender-declare war-surrender-declare war" when it happens. But that's a bit cosmetic in nature. The real bug is that in both 1916 and 1918 games Luxembourg (what's the point of it, I repeat?) and Belgium capitulate on turn 1. Belgium should be like Serbia IMHO (never surrender) and Lux should not exist.

-The AI attacks with tiny stacks all over the front and its Grand Offensives are puny. It should IMHO stay defensive apart from massive GOs.

-Similarly, without changing anything in the Redeployment phase, in both 1916 and 1918 I plowed through the defenses with my Verdun and Kaiserslacht offensives respectively. I cut through them like a hot knife in butter. Even the Austrians blasted through the Italian front in the Trentino and Piave offensives respectively. To make sure, I had not reinforced these stacks and simply used the historical set up. I should have been slaughtered as in real life, but in the game I completely ran the Italians over. I strongly believe that it's partly due to the Redeployment phase, in which the AI moves too many troops around and becomes super weak all over. But that's only a theory.

-The British AI is probably the best, making my life as the Ottomans very tough. It even tried a Dardanelles effort in 1916, when I had moved I Army to the Romanian front! The problem for it is supply, it gets isolated on its own too easily.

This game is very close to being perfect. Once you can smooth out little niggles like those above, it will be there.

bk6583 wrote:Brother let me second, third, and fourth this! It's ruining my current game as the German AI in France just redeployed its GHQ with over 6 corps, air, and arty to an isolated area - all doomed to die an isolated death. I've never expected greatness from AIs but I also certainly expect that stupidity has at least been weeded out. This is really stupid. Really excellent summaries above though that will give anyone pondering buying this game food for thought one way or the other.

I'm addressing this point, not only for redeployments, but also for placement of reinforcements more generally.

pk205 wrote:-Buggy capitulations. One thing is the weird message sequence "surrender-declare war-surrender-declare war" when it happens.

Strange, I fixed this "cosmetic" bug already with version 1.08J!

Well, improvement worked out in 1.08L:

- Military AI now carefully checks the supply status of areas when considering the redeployment or reinforcements opportunities if the difficulty level is set 4+

Calivinus,

The great effort and hard work on your part is very much appreciated let me tell you! I don't know how big your to do list is but I did want to add just a few things based on the above posts.

The naval game could still use tweaking. It's very perplexing to launch a raid into the Dogger Bank or the German Bight or the Black Sea and to have your ships damaged from mines! I completely turned off all fog of war and there are no mines in any of those areas. Also, I believe you said you were going to work the sea movement paths. Try to move the Austrian navy into the Eastern Med - it hangs up around Athens and sits there. I have to painstakingly move from zone to to zone clicking enact movement each time to get there. When I did get there with Austrian and Italian (was playing Central Powers and Italy joined my side) fleets set on control, the next turn the control flag was still a British one. Tried this for several turns and the control never changed.

Lastly, would still love at least a mediocre AI but as it is it almost seems programmed for ahistorical early defeats. The penny packet attacks referenced above occur constantly and practically open the front door to easily won offensives when I'm the active player. One other observation: The AI GHQs pack themeslves with tons of corps and arty that in my opinion could and should be up front with the regular armies. Well, enough commentary from me. Keep up the excellent work!