Grognard Wargamer Thread!

Here’s another update on my ongoing solitaire play of “World in Flames.” It’s November/December 1942. The Germans have taken Stalingrad and one of the oilfields in the Caucasus Mountains, but they just lost a big attack against a Russian HQ near Tbilisi. If the Germans can get four corps to the Turkish border, the Turks can join the Axis. But the snow is slowing them down.

Japan has taken most of the objectives it took historically; it is now focused on taking Free-French Indochina, Burma, and perhaps invading the USSR. The Americans, however, are about to invade the Marshall Islands and perhaps Guadalcanal. The Americans have also invaded and conquered Morocco and are now completing their liberation of Algeria. The British are pushing the Italians out of Egypt.

Overall, I think the Axis is doing better than historically, and the Allies are behind schedule. It’s been an exciting game!

This is fantastic thank you for the update!!! Exciting the way your game playthrough is developing!!!

p.s. Also fwiw I agree with the above. AI is not a particularly hard development problem to solve. UI/UX is.

I feel like almost every wargame/4x disagrees. It’s the thing most often complained about: braindead AI.

For sure. My point was its a relative thing. Programming a strategy AI for a specific game really is not very hard. Usually you have to tone it down because most wargames (but not actual wars obvs) are solvable logic puzzles which machines are great at. Hell just throw in some random moves and most players will think there is a lot of deep intelligence behind it.

Often when wargamers are complaining about AI they are actually complaining about tuning. “Dumb AI” - tuned too easy vs “cheating AI” - tuned too hard. (obs you can have broken AI that cant actually play the game but thats rare imho and usually only happens with first time AI programmers.)

Some wargamers are like golfers and a golf course. They want it to LOOK hard but actually be EASY.

They want to win and tell a story of how they engineered victory within the games mythos/historical context but also have the fantasy they ‘out thought’ the computer. All of which is not so hard an entertainment experience to make.

But UI/UX, now that is HARD. You have to cut your way into a very complex system (to human minds) and show the appropriate decision nodes at the right time.

all imho obvs but that’s my experience making strategy games.

Yeah, I know your background which is why it surprised me to hear you say that after years and years of hearing about AI being hard to program and terrible much of the time. Interesting perspective, thank you.

I could be wrong of course! I dont want to be rodsplainin here. I just find the subject very interesting! As a gamer I want deeper feeling wargames. I just think defining that problem as AI players is not going to get us the results we want. Now AI as a game director or generating dynamic content that’s exciting stuff and hard no doubt! Anyways great discussion!

I think if it was that easy, there’d be a wargame with it- but while some wargames probably could be solved with some AI techniques, it would involve tacking on an enormous amount of development time while the game is pretty much complete, and that’s a non-starter. I also think people underestimate the sheer depth of the game state of even somewhat simple wargames. I think, for example, if I made a computer version of No Retreat!, it would actually take a long, long time to develop an AI for it that was competitive with a top player.

For example, most of the really impressive AI work is done on games like Starcraft 2 which are pretty much known quantities- it wasn’t blizzard making the AIs that the articles suggest, and given the constraints of development, it’s probably not realistic to expect that.

True, though I’d add for historical wargames what breaks the spell for me is the computer doing things that are completely at variance with the historical situation. I would be happy with an AI that plays more or less following the same sort of principles taught to and practiced historically by the generals on the side it is playing. Then again, I am less concerned with whether the AI is good at victory conditions or not, though I know most players are.

Well put! I feel similar. Bad AI = implausible AI for me. Not whether its utility value algorithm is good per move but whether its moves feel plausible.

In my experience, in games AI can mean two things.

First is AI-as-heuristics. Having a game being by able to play a game and have a good performance. Bots moving around a complex 3D environment and stuff like that also falls in this bucket. For strategy games this is a solvable problem given a fixed game design and enough development time with said design fixed (note that these two factors are not necessarily there in most development schedules. It’s easy for the design to refuse to solidify until very late and then there’s no time for this AI).

Second is AI-as-game-design. That is, a system or set of systems that create an experience for the player that is satisfactory. In a very simplistic way many solo board games have this kind of AI. It can be clever or quite basic and driven by randomness. The question is whether the experience/challenge it provides “works”. This is a much harder problem and part of the game design per se. It’s not necessarily “solvable” in the sense that there’s no right answer. It’s not a technical problem but a set of creative/design choices that will work for some players and likely not for others.

Then you have UI/UX which is beyond hard to crack…

Indeed. Though with computer games, it does intersect the tech, but in complex and often obscure ways.

Quick update…

My new tweezers arrived from Amazon. They are way too narrow. Tweezer fail. Back to the drawing board.

Saipan, End of Turn 4 Update

Things look grim for the US forces. The Japanese have destroyed 4 of the 6 beach landing zones, and have the US forces split and isolated.

The die roll gods have not favored the US, and a ground support rule misunderstanding further stacked the odds against the invaders.

As things head to the ninth day of the battle, the US would need some significant help from the die roll gods to establish any sort of perimeter.


It’s been interesting to revisit a board game like this, after probably 20 years away from playing.

I confess I’m not a fan of how the combat rounds play out. There is the adjacent forces’ strengths, then ranged barrage strength, then off-board ground support strength, then final ranged protective fire from the defense, all added in to each battle. This means that you’re fiddling with writing numbers down on paper to track what’s been used or not. More than half the combat strength in a battle often comes from elements not directly involved in the combat. It often feels like you’re just shifting numbers around that aren’t connected to the units present. I’m not sure if I’m articulating this well, but the combat feels a bit too abstract for what I was expecting from a board game.

I don’t know if I was just more tolerant of these sorts of things back then, or have gotten spoiled by computer games that track all these things, or this is just … weak design. Perhaps it’s a little bit of all these things.

Having said that, I’m enjoying getting away from screens, and the tactile nature of the game has a charm to it that is relaxing and fun. I’m looking forward to exploring more games going forward.

I suspect nowadays with the costs of production for quality components low enough, those sorts of additional design elements for combat would have been placed somewhere other than the players brain, either on cards if they are fixed, or a track to move a specifically marked counter along with or a calendar with the values provided on the map or some such. I think gamers nowadays expect their games to do much more of that heavy lifting in terms of remembering factors or rules or such…It might be a design element that couldn’t be mitigated due to the cost of production rather than bad design. Mind you “back in those days” wargamers would also put up with an awful lot more overhead in terms of managing the game simply because most games made the player do it!

Great points, and after watching some Let’s Plays of people playing newer games, I can definitely see what you mean. Boards now seem much more surrounded with aids to help track these sorts of things, just as you mention.

Not only that, but there was a segment of grogdom that reveled in the complexity and mind-boggling minutia that wargaming often required.

@ Godzilla_Blitz

Look for a vacuum pick up tool on amazon. They use these in the electronics industry to pick up ICs, it will likely work great for counters.

It is basically a little suction cup on the end of a pen with a hole in the center. There is normally a hole in the pen on the other end. Cover the pen hole with your finger and press down on the part to pick up. Release your finger to drop.

Whoa. Mind blown. I didn’t even know this was a thing. Not to be critical, but won’t this only get the top unit in a stack? Would tweezers be better for picking up stacks?

No the suction is so powerful it can grab through chits.

Yeah, it will only get the top chit. Tweezers are generally too narrow to go around the short side of the chit, but you might find some that have longer arms. I have some like this in my lab, maybe you could find similar and bend them more permanently open (as long as they are sufficiently long enough).

A chip extractor tool might work for a small stack.

I used to know a girl… Oh, never mind.