Welcome to the American Civil War Gaming Club's RU-vid Channel. This channel has been designed to highlight our online Wargaming Club and to show everyone just how much more fun strategic gaming can be when played against a real live opponent. The channel will be updated frequently with more tutorials and information so be sure to check back often!
Have really loved this series as I have with all of your videos. I'd love to see you refight Antietam vs the AI but with you as the Yanks. I think it would give a great comparison of the abilities of the AI on the defensive versus offensive.
I would think it would be much harder to program the AI for the attacker than the defender. While I'm sure the AI would still be abysmal compared to a real opponent I'd like to see you try this battle with you taking the Union to see if the AI better as a defender.
Enjoying the series so far, but I have to admit that it's a very stark reminder of how bad this AI is. As far as I can tell, it has improved very little since the '96 Battleground Shiloh game (a game which I adored back in the day), and I suspect that it's basically the same AI. I can't imagine why this AI hasn't improved, as hundreds of other game companies have made AIs that are at least competent over the years. There is no reason why the AI can't move, attack and defend in formations, and no reason why the AI can't keep leader, artillery, and supply units protected. The rationalization that there are "millions of possible moves every turn" is bogus to begin with...absolutely no human player considers "a million possibilities" each and every turn, but instead has a relatively simple awareness of what they consider to be important steps to victory. Compared to other games, these games actually have very little to consider...there is no research, politics, manufacturing, etc. to consider. The AI here simply needs to know if it is attacking or defending, and then take the steps of doing so in a logical manner. It should be able to recognize the relative strength of the enemy, and decide if attack, reinforce, or retreat are the best options at any particular point. There are plenty of other game series that are able to do this. My conclusion is that WDS considers any real work on the AI to not be worthwhile...but they also realize that trying to sell these games without an AI player would be a big mistake. The exact ratio of solo vs. multiple players probably can't be exactly known, but it's well understood that the majority of overall gameplay is solo. So in a sense, we are being cheated by their refusal to dedicate minor resources to improve AI. They've had plenty of time to do it. I could list many current games that I play, some developed by small teams (sometimes just one or two people) that take AI improvement very seriously and continually work on it to the satisfaction of people who have purchased their games. There is simply no excuse for WDS AI to remain this poor.
Agreed. WDS should either make the A/I competent and an actual challenge, or concentrate on giving their customers a stellar PBEM experience by opening up map editors, new scenarios, and a better level of PBEM security. These games should either be solo games or MP games - trying to do both isn't ideal.
I like all of your ideas for improvements. I do have one addition to the leader bonus: Chuck Norris on foot: add 100% Chuck Norris on horseback: enemy surrenders
I think we all know how this will end. I think the scripts just provide a strategic framework for the tactics programmed into the AI. I created a scenario for Sicily '43 (also by WDS) some years ago. I had to add scripting to get the German reserve units to move in "reasonable" directions once released. This helped somewhat. I think the answer to achieving a good AI opponent is that already used in the chess community. An online program that is truly intelligent. i. e. capable of learning from its mistakes. The current AI packaged with each game cannot hope to do this. But I do like to use it to learn the rules and give me experience as to what works and what is really stupid. Some scripting will still be necessary. (a3 or h3 will never be good opening moves. EVER!) I REALLY hate it when the AI just moves units aimlessly burning supplies and causing fatigue. I mean, this is just not realis...
I was curious as to why Burnside only crossed with a couple of units which exposed them to being surrounded. That seemed like a really bad scripting. Taking a peek at the actual scenario answered the question. Burnside's Corps isn't released until 8 AM. What crossed was a couple of units that the Rebel units triggered the release of because they got to close to them. Script kicked in and they made a suicide charge. Scripts will force units to move but won't make them any smarter once they get in contact.
The big problem seems to be in most of these games that the AI has a hard time to keep track of the overall and overarching objectives on a larger scale aswell as what to do with individual units and how to handle things on a smaller scale and ultimately how to bring these things into Unison.
Our wargames are a lot tougher for real AI to handle than the games AI works in. For example, it took almost 20 years for Big Blue to reach the level where it could beat a master on a game that has only 32 units of 6 types on a map (board) of 64 squares (hexes). Antietam's map has 28,000 hexes. It has five general counter types with subtypes of each. And it has over 500 of these units to track in a system that allows many to be completely hidden. Add in each unit has multiple ways of moving and you have an AI nightmare. Most games that have a completive AI achieve it by limiting what the player can do and/or by having the AI play part of the game for both sides.
Your comment at about 5 minutes in "... which might actually be an improvement on McClellan himself." alone was worth watching the whole video :). And great job presenting the matter btw.
Anyone who has played these games knows how poorly the AI performs. Hopefully, this series is to remind the developers that this issue is a real drawback to their products. I have to play with house rules, such as not capturing enemy cannons or officers, but that doesn't help the AI's lack of any kind of sound tactics. They claim to have two kinds of AI, one being scripted. It's obvious that the scripting is just as bad as the general AI and it's almost pointless to even start a battle against the AI. There is no sound logic to the way the AI performs. I know it can be challenging to create a competent AI, but many other systems have done it well. Even old games like TW Rome II and TW Napoleon have an AI that can keep formations intact, attack in lines, and protect ranged units. I would probably buy more WDS games if the AI was better. It's well past time for some of these games to be rebuilt from the ground up, because as they stand now, I find them hard to recommend to new players.
Preaching to the Choir. I get into much more detail about the A/I and how it works (and the why it works like it does) in the upcoming episodes. But I think it is safe to say the A/I does not offer a competitive game to even "average" human opponents. Thanks for watching!
@@acwgc873 I'm really curious as to why Antietam Union side was chosen to test the AI. That is a tough nut to crack even for a human player. I'm 0-2 at it. The superb PBEM system for WDS has allowed me to play 4-6 hours a day for the past 4 years against a very fine player in Scotland! I have been wargaming since 1958 with the first G-Berg game by AH. WDS is by far the best wargaming experience I have had along with miniatures which I am now just getting back into. Sorry, I love WDS games and hate to see them get picked on regarding their AI. Others have told me it has improved the AI on some of the games, so I am testing Vicksburg as the Union. Each series in the vast WDS collection has some excellent 'Getting Started" scenarios to learn the games. You learn the basic WDS system and you can play any of their games. Again, these games are best enjoyed with a PBEM opponent.
The WDS AI has always been abysmal. It's fine when you're learning the games, and it does okay-ish in shorter scenarios. But once you've learned the mechanics of the games, anyone with a basic grasp of military tactics will run rings around the AI. It doesn't consider fatigue, it doesn't consider command and control, it has no idea of flank protection, and its idea of an attack is to stuff as many men into a hex as possible and move forward.......sigh. Even worse, since the AI is entirely scripted, there is zero replay value. It'll always do the same thing over and over again. Don't get me wrong, I love the WDS games (though I'm more into WWII than black powder), but since I don't have the patience for PBEM, and the AI is laughably bad, my only choice is hotseat games against myself.
1 -4 are excellent ideas, and #5 would be nice (not that important for me, but as you stated, should be the easiest to implement into the game). Being mostly a mult-player game person, I have been a proponent of your #1 multiplayer suggestions for a long time. In conjunction with your thoughts, I would like to be able to click on a leader and have the "next unit" button (whether for movement or firing) highlite only those units under that leaders control. In battles with large armies, it is sometimes difficult to make sure you have moved/fired all the men under your command. Let's hope WDS is taking notes!
Regarding Leaders I would add that the amount also includes the leadership quality of the leader. Quite a few leaders would not inspire them men to fight harder. Might inspire them men to shoot the leader by accident instead. Some games use to add die rolls for bad leaders that included being drunk resulting in a negative effect on the men. Bragg would be Army level but probably no one would follow him if he tried to lead a charge. Lee on the other hand even attempting to join the fight would be enough to make them whip anything in front of them.
Move fatigue would have a problem as fatigue is currently implemented. Fatigue in the game is Combat Fatigue not how tired the men are. To reflect this the recovery rate from Fatigue is extremely slow except at night where it is just very slow unless you use the optional rules. For movement fatigue you would need higher day time recovery rates to reflect the ability of men to recover from being tired by just taking a one hour break. A typical Civil War march using 50 minutes marching with 10 minute rest could go all day without fatiguing the men. However, if they changed to quick time or faster then fatigue set in very quickly.
Regarding LOS, I would like to see a more complex handling of it. Right now if you can find a hill and the game says you can see 40 hexes you can see a unit at that distance with the same resolution of two hexes. The information displayed on the enemy unit should degrade with distance. Probably as a percentage of max range since the range depends on weather conditions. For example, units at 36-40 hexes for a 40 hex LOS would display only as "?". Units at 31-35 might only show type. And, so on.
I agree on movement fatigue. Thought on it for years while playing or testing. Something that equals rush marches and quality forces should be a thought and maybe not to difficult to add for the WDS boys. Fatigue as it is today is in general a bit weird and how troops get back. It's not only to rest a unit with no movement and you get the fatigue lowered. There is a "diceroll" to it so some might be dead tired for a very long time. Today there is now difference if you go all out on movement points on a turn or just move the unit 1 tick. So fatigue could be in for an overhaul .
You missed one... there is no way that 1000 men could shoot from a line that is 125 yards wide (even the long diagonal is only 145 yards). Not going to do the math here, but even 500 men, firing from 2 ranks, is too much for the hex scale. The actual maximum should be 350 - 400. Yes, a regiment of 1000 men could fit into the area included in one hex (2.8 acres), but they'd be shooting their brethren in the back trying to fire. While 500 is still too much, it makes far more sense to break large regiments into "battalions" of equal size. And being of one regiment, these "battalions" should be required to maneuver adjacent to one another, or face a morale penalty. 100 men could still stack in a hex, and melee together, but firepower from a single hex needs to be restricted... jmnsho
I've watched all of these series, and the armchair general in me is consistently frustrated by the conservatism on both sides, but especially the Union side. No risks taken, and worse, rarely are advantages pressed. I don't think you ever mentioned it, but I thought the Union lost this game very early on when they conceded Midway for no apparent reason.
Very good detailed info on this subject. You migh need to touch the two other rules that is close related to this, Flank Morale and Rout Limiting. They play an important role when it comes to what happens thanks to using this Density modifier. Played, and tested, for Tiller games for many many years but I don't think I have seen something explained this good. Known about it, but that is what you need to read in manuals and talk about in chats. You should try and have a go at many more rules. I see you done some more but they all need to be explained in depth. Great work.
Just discovered this and am loving it! I'd love to get involved in one, especially a Napoleonic version. Have done a few 1v1 Operational PBEM's using the WDS/Tiller games before but this is just fantastic.
That was the most amazing battle, truly nail biting to the last turn, and one of the best of the AAR's Blake has yet produced. There can be, at this stage, for my money, few, if any other MP's to equal them. Congratulations to the CSA, commiserations to the Union and a huge thanks to Blake for creating, running and providing these for our pleasure. <Salute> Sir.
I think the Confederates are in a bad spot. I was surprised the Rebs didn't send a few units up the far west road from the start to prevent being flanked on the left so early.
I prefer having this turned on, to avoid gamey-ness and because I think it's cool to try and re-crew and haul the guns off the field to keep the points (or to prevent the other side from getting them). Think the rule would be a lot better without allowing the 'other side' to shoot at you with your captured guns, which seems goofy.
Other enhancement to non-optional rule would be instead of having the battery vanish and end up on the opponents VP list, maybe a battery defeated in melee should rout and drop quality (to F or whatever), so it's not a VP haul but still reflects that the unit should be relatively ineffective for the rest of the battle.
