Ultimate General gameplay footage..

[fabius]

Yep they added scrolling.

And they really seem to be listening, wanting feedback and acting on it.

[Queeg]

Yeah. Humans will do the same thing when nearing end game. So the AI is smart. It is a matter of game design.

Humans who play historically won't.  No need to program the AI to play like a gamey human.  Better to program the AI properly and leave it to the humans to control themselves. 

In any event, they've said they plan to fix this.   

[JudgeDredd]

Yeah. Humans will do the same thing when nearing end game. So the AI is smart. It is a matter of game design.

Humans who play historically won't.  No need to program the AI to play like a gamey human.  Better to program the AI properly and leave it to the humans to control themselves. 

In any event, they've said they plan to fix this.

I had a BIG reply to your concerns - and lost it like an idiot!

So here it is in a nutshell.

When I play a game, I want to play the game. I don't necessarily want to be confined to what happened. Otherwise, as the rebs, I'd never win - and I can do without the AI beating me (or me losing) just because "that's how it was".

I can understand why it irks people who have an in depth understanding of the war.

And I actually don't agree it's "gamey" either. I often break a unit away and check round a flank or try and come in from the rear and even try and get a victory location. If for no other reason than the enemy has to drag units away from the front to deal with the incursion.

Either way - I have no interest in playing a game that is coded in such a way that you cannot change the outcome. Why bother playing if I already know the result.

I don't have your knowledge of the Civil War - so I have to bow to that - but I don't agree a "game" should be exactly as it was. I have my own ideas on how a game (certain games, not even all) should stay within the realms of reality.

I probably didn't really put any of that coherently - but oh well.

[Jarhead0331]

Yeah. Humans will do the same thing when nearing end game. So the AI is smart. It is a matter of game design.

Humans who play historically won't.  No need to program the AI to play like a gamey human.  Better to program the AI properly and leave it to the humans to control themselves. 

In any event, they've said they plan to fix this.

I had a BIG reply to your concerns - and lost it like an idiot!

So here it is in a nutshell.

When I play a game, I want to play the game. I don't necessarily want to be confined to what happened. Otherwise, as the rebs, I'd never win - and I can do without the AI beating me (or me losing) just because "that's how it was".

I can understand why it irks people who have an in depth understanding of the war.

And I actually don't agree it's "gamey" either. I often break a unit away and check round a flank or try and come in from the rear and even try and get a victory location. If for no other reason than the enemy has to drag units away from the front to deal with the incursion.

Either way - I have no interest in playing a game that is coded in such a way that you cannot change the outcome. Why bother playing if I already know the result.

I don't have your knowledge of the Civil War - so I have to bow to that - but I don't agree a "game" should be exactly as it was. I have my own ideas on how a game (certain games, not even all) should stay within the realms of reality.

I probably didn't really put any of that coherently - but oh well.

Interesting comments...I think there is a difference between a game that permits different and possibly ahistorical outcomes and a game where the AI cheats, or pursues an impractical, or historically impossible strategy in order to win.

Its great, in my opinion, if say the game permits reinforcements to arrive dynamically, or AI leaders to attempt strategies that did not occur historically. However, where the game pulls a trick that would have historically, been impossible, be it due to logistics, morale, supply, timing, exhaustion, etc., I think that crosses the line and takes something away from my enjoyment of the game.   

[Queeg]

Interesting comments...I think there is a difference between a game that permits different and possibly ahistorical outcomes and a game where the AI cheats, or pursues an impractical, or historically impossible strategy in order to win.

Its great, in my opinion, if say the game permits reinforcements to arrive dynamically, or AI leaders to attempt strategies that did not occur historically. However, where the game pulls a trick that would have historically, been impossible, be it due to logistics, morale, supply, timing, exhaustion, etc., I think that crosses the line and takes something away from my enjoyment of the game.

My complaint really isn't just about gamey tactics (though I think the rush-the-VP-wait-for-the-bell tactic is exactly that).  The bigger problem - and one the folks over on the Steam forum seem incapable of understanding - is that programming the AI to focus on last-second VP-rushes results in the AI dissipating its forces at the point where the real battle is being fought.

My last game is a perfect example:  I'm the Confederates, in the last few minutes of the first battle.  I have only one capable brigade left, but it's a big one and is pressing toward the large VP on the SE edge of the map (below Seminary Ridge).  The Union has one smaller brigade, plus an artillery battery, on the VP.  My guys are getting blasted with canister but are slowly pressing forward on weight of numbers alone - it's an awesome, very CW-like fight, and it's going to be very close. 

Suddenly, the Iron Brigade marches up from the Union rear (where they had previously retreated after being mauled by another of my brigades).  I think, "Oh no!  Game over."  I figured the Iron Brigade would extend the Union left flank, overlapping and enfilading my lone brigade.  No way my guys could have withstood both canister and flanking fire. 

Instead, the Iron Brigade marches in column, past the fight, and on a bee-line toward Herr's Ridge on the NW edge of the map.  There's no doubt in my mind they were targeting the VP.  They marched past my artillery (who raked them with canister from the left) and past my skirmishers (who pinged them from the right) and marched like a Napoleonic column toward Herr's ridge.

They never made it.  My artillery finally drove them back.  In the meantime, however, I captured the SE VP by the skin of my teeth.

So programming the AI to be VP-happy isn't just gamey.  It actually makes the AI weaker where it matters most.   

[magnus]

 The thing that I see in your post is that the AI is working the exact way that human players do.

They exploit the " gameyness " cracks that almost all games have.

Don't get me wrong, I don't like it either.

Which is why I only play single player games now.

[Queeg]

Which is why I only play single player games now.

Same here.  Which is why I'd hate to see the AI programmed to mimic a gamey MP human. 

[MikeGER]

Its great, in my opinion, if say the game permits reinforcements to arrive dynamically, or AI leaders to attempt strategies that did not occur historically. However, where the game pulls a trick that would have historically, been impossible, be it due to logistics, morale, supply, timing, exhaustion, etc., I think that crosses the line and takes something away from my enjoyment of the game.

+1

i think to do the trick the AI programmer of a wargame must balance the love the AI has for taking VPs against the losses it will cost to reach the VP with penalties for destroyed own units. also all units need 'historical' (sensors and message travel times) situational awareness and selfpreservation instincts in their 'tactical AI' 
that will help to collapse a 'gamy' attack or march order, even so the 'strategic AI' of the game had calculated to wage it. (cause it cant be allways perfect)

FP:RS  is a good example which gives the historical timeperiod correct feeling, but allows different approaches and even delivers different approaches by the AI.  its a walk a fine line to achieve that and need a lot of beta testing and fine tuning to reach the crest (and staying there when features get added/tweaked and bugs fixed from patch to patch)       

       

[JasonPratt]

For the record (seeing as how I'm "one of the folks over at the Steam forum"), I'm 100% against the kind of behavior Queeg was talking about, and I said so there, too.

I'm not necessarily against opportunistic rear-area raids on lightly protected VPs, or even ones planned and executed by drawing out protection forces. But I realize without supply and C&C issues those are going to highlight some of the unrealistic portions of the game system (which doesn't feature supply and most C&C issues); and I think there are ways of doing it which would make better sense within the game system while still nodding in the direction of historical tactical necessities. For example, requiring a threshold of force, proportionate to the importance of the VP, to really take and secure it for ostensible usage (which translates in game terms back into "getting victory points for having it".)

So for example a 5000 VP ridgeline ought to need 5000 men to initially secure it, and perhaps should only be scored in proportion to the relative level of force parked there at the end of the match. (I say "perhaps" because once an area is initially secured then non-combatants would hold it independently of combat support on the defense -- they would be proportionately overrun or hassled by undefended spoiler attacks but two brigades wouldn't have to camp there permanently to keep the VP once initially secured.)

There must be some kind of code like this in the game already, for arbitrating conflicts over VP areas during combat. I'm thinking the principles of that code could be extended a little to achieve this result.

[Queeg]

For the record (seeing as how I'm "one of the folks over at the Steam forum"), I'm 100% against the kind of behavior Queeg was talking about, and I said so there, too.

Yep, you did. 

Having played a bit more, I see now that the problem probably is most acute in the opening battle, where there aren't many units on the field and the Union has a bunch of small, disposable units. 

Having also paid more attention to the briefing, I see that the goal of the Union AI is to slow the Reb advance, which is exactly right.  The Confederate goal is described as not losing Herr Ridge.  I think perhaps the Union AI is splitting its attention between achieving its goal (slow the Reb advance) and denying the Reb goal (hold HR). 

Since the Union is unlikely to ever take HR, I think the AI would be better advised in this instance to focus all its effort on stopping the Confederate forces.  I've now re-fought the first battle on a variety of difficulty settings,and the Union AI always depletes its forces by sending token forces (and sometimes even a full brigade) on a fruitless charge for the Confederate rear.  Far better, in this instance, to focus on the very realistic goal of mauling the Confederate units (the Rebs only have four brigades) rather than the highly unlikely one of taking Herr Ridge.

I think the AI actually is quite good overall.  Sometimes, though, it can best accomplish its goals by destroying the enemy units without focusing so much on specific VPs.

I'm not sure exactly how the game calculates the "winner" at the end of a scenario, but perhaps more weight should be given to casualties than VPs - to better distinguish real from Pyrrhic victories. 

[fabius]

Yes, I've only found it a problem when force to space ration is low or I've really left a flank open. When it happens it's annoying as hell. I'd be ok with it if it was a rare occurrence, but seems too often.

That said it's not that often and reserve deals with it. And in no way a game breaker.

[mikeck]

Well Grant`s Vicksburg campaign was deep inside enemy territory.... How about Sherman`s march to the sea? There were many cavalry raids deep inside enemy territory.. Grierson`s raid comes to mind pretty quickly.......Are we thinking about the same Civil War?

Well, in fairness, there is a difference between sending a fully self-sufficient army behind the lines and a single brigade. The latter lacks the ability to do much of anything behind the lines.

Also, Vicksburg was not behind enemy lines. The first time, Grant came from the north extending his supply line from Grand Junction. This was cut by a confederate cavalry raid (VanDorn) and compelled Grant to retreat. On the second attempt, Grant landed south of Vicksburg and Marched north to pace himself between Jackson and Vicksburg. In this case, he could still receive supply from the river and the Union largely controlled the area north of Chickasaw bluffs.

So Vicksburg was an amphibious operation into enemy territory but not behind the lines. The Mississippi River from Vicksburg north was a "line".

I'm not aware of any instant where a regiment or brigade sized unit accomplished anything of consequence behind enemy lines. If you march your brigade behind my lines, you are out of communication and very soon...ammo. Then you die

[JasonPratt]

Then you die get creative!

:D

Okay back to serious game discussion. I'm sure I've played the opening Confederate mission twice (on Dynamic, the most advanced balanced level) and while I saw the AI making spoiler runs at my cannons on the ridge I never saw them trying to flank it from behind which I was really expecting since cav charges from behind on arty would surely work better (and the AI definitely had room to maneuver had it wanted to). Then again I was mainly holding the ridge since those were my orders and until late I didn't have reinforcements to try much more, only pushing forward leftward a little for an opportunistic strike on that 500 point hill to the north. But I was defending in good cover far enough ahead of the ridgeline that the AI could have flanked me hard with a little effort and ruined my rear lines (or whatever taking that ridge would represent). Sure, I would have been able to turn around and crush them (maybe) but doing so would have exposed me to horrible rushes from the main Union line.

Of course, as Queeg says, if the AI is simply taking an opportunistic straight-line shot for a last minute VP win when conditions present themselves, that wouldn't be at all the same as rear-flanking my VPs or even going after my anchoring arty from behind with cav raiders.

I'm not sure exactly how the game calculates the "winner" at the end of a scenario, but perhaps more weight should be given to casualties than VPs - to better distinguish real from Pyrrhic victories. 

And then again, my impression (admittedly recorded a night's sleep later in each case) was that losses do count more than VPs, since I was solidly ahead on VPs on both missions so far but the game scored it as a draw with about equivalent casualties. I shall try to pay more attention next time... {pondering}

[Grim.Reaper]

New patch...interesting....


Patch 0.81 with elevation map and realistic battle speed

June 27 - DARTIS   


We thank you very much for the warm reception of our game and your continuous feedback and suggestions. Many things you have asked to improve or fix are now available in our newest patch. Important new features are the elevation map layer and the revamped, more realistic combat pace accompanied by an enhanced AI. Have a look at the patch notes in our site's blog (includes some informative images):
http://www.ultimategeneral.com/blog/ultimate-general-released-2ndpatch

or read below:


Ultimate General: Gettysburg Patch 0.81 rev4590
---------------------

NEW SPECIAL FEATURES
• Added Elevation Map Layer as many players requested.
- You toggle on/off with the M key (you can re-assign button in settings).
- You can use this additional map layer to understand which hills are good for placing your artillery. Height indications are measured in feet above sea level.
• You can now restart a battle. This is a temporary solution until we finalize the multiple saves system.


GAMEPLAY IMPROVEMENTS
• Various and major AI improvements for better challenge. Union AI especially should be much more effective in holding their ground.
• Melee balance improvement. Fixed cases that could result in too low damage. Cavalry should be more effective but still preserve its vulnerability to ranged fire.
• Artillery more vulnerable to projectile damage so that it cannot be used as "tank" as it could be in some cases.
• Terrain importance enhanced. Cover, Altitude bonus to range, Movement penalties increased for making the good defensive positions even tougher to capture.
• Various improvements to morale, damage rate, combat mechanics, including the above mentioned result in a more satisfying combat pace. Not too slow, not too fast. We need your feedback if this is noticeable and works as you like.
• Flanking/Rear damage increased for more rewarding tactical maneuvers for you and the AI.
• Several improvements in Scenes concerning time of reinforcements and random events. For Union especially this will help them to have better chances in 1st day as the reinforcement routs and random delays are not as strict as before, so the flow of battle is more intense. Slocum reinforcements sometimes can delay too much and so Union AI has difficulties withstanding a possible evening attack. We will fully address this problem in next patch.


PERFORMANCE/GAME STABILITY
• Performance and FPS improvement. Lag issues should now be much less.
• Possible 3-4 seconds freezes should now not happen as frequently as before. Please let us know if you still get this problem.
• Fixed additional bug issues that caused game crashes in moment of any battle start. Next week we will sort any other pending issue related to supported resolutions.
• A new troubleshooting button is added temporarily to menu to lead you to solutions to common problems (needs internet connection).


CONTROLS
• The LoS of your selected unit (the area it can see) now no longer creates a so dark area around the unit.
• [NEW] You can now tune the LoS intensity and duration. Go to Options=>Gameplay and set it to your liking.
• [NEW] You can now use alternatively a key to rotate your units instead of middle mouse button. Default button is TAB and you can re-assign it in settings. To use, select a unit, and while pressing TAB use your mouse to rotate.
• Improved Formations pathfinding.
• Now you move only the visible units belonging to a group. This helps to control your saved formation groups better.


TEXT
• Various Corrections/Additions in main menu and battle description texts.
• All battle description messages available for Days 1,2,3. Decision room text descriptions still missing from 2nd day battles and onwards. They will be available in the next patch.
• All battle action messages available for Days 1,2,3.


BUG FIXES
• Possible issue of reinforcements not appearing at all after you load a campaign is fixed. This was a particularly annoying bug that damaged the gameplay and caused difficulties to AI especially in scenarios that reinforcements are crucial for its defense.
• Fixed additional problem that made some units invisible after loading a campaign.
• We addressed many campaign scenario issues that caused the battle to continue wrongly or to end prematurely with wrong win/loss messages to appear (more often there was a wrong "Dead lock" message). We remind that we need to finalise the battle to expand to the 4th day to completely eliminate this problem. Please continue to report such problems if you find them.
• Fixed condition that could allow artillery reinforcements to fire way out of map borders.
• Fixed cases which made units to evade or retreat in wrong directions.
• Fixed additional bug that could result in units not to shutter out of the map, despite having enormous losses.
• Fixed targeting bug that made AI and your units to prefer artillery and skirmishers too much and to ignore more serious threats near them. This bug fix makes your units to think much better with their own initiative without any micromanagement.
• Fixed major melee bug that resulted sometimes to see units having no melee collision damage at all and AI to become idle because of this.


***WARNING***
Because we had to cure some very serious bugs in the saving process, the saving system was refactored. As a result, this patch will make your old saves incompatible and not loadable. You will need to restart your battle campaign. We are sorry for the inconvenience.

[TacticalWargames]

Well with the new patch his is shaping up to be on the road to classic status. Not only is the gameplay great but the attention to detail is excellent. Naming all the outlying farms after their owners.

I see a great future for this engine. Napoleonic era comes to mind and really any pre WW1 era war from the start of gunpowder.