OOB Scenario Design Woes

[MikeAP]

I design scenarios for real world military training exercises and I've done some professional design work for other games on the market. I really dislike what I see from almost every vanilla campaign and some of the trends I see with user-posted work.

I would offer- this is really not how scenarios should be designed....

Turn 1 - spawn enemy units / add income
Turn 2 - spawn enemy units / add income
Turn 3 - repeat
Turn 4 - repeat

This is simply called spamming units and not really a realistic way of designing a scenario. It puts the player at a significant disadvantage because its just wave after wave of enemy spawns and it's impossible to try and determine enemy force composition/disposition with relation to their ORBAT. My friend and I are playing through a terrible cooperative scenario right now where we're just fighting brand new waves of enemies - we've taken to calling it "Nazi Zombie's" because every turn seems like we're fighting fresh German units.

Sure enough, I opened the scenario in the editor....and the Germans are getting 5 new units EVERY turn plus more income. That's not really what I would call optimal design. You're just smashing the player and not even offering a fair chance.

What I would offer from my experiences: These are just 3 steps within a whole sub-set of the process of building a tactical scenario

1. Determine the OPFOR Task Organization/Order of Battle - this also includes determining their income and supply. Do not give the OPFOR any advantage that it would not normally have (ie - the technique above - "Nazi Zombie Waves")

2. Determine what the OPFOR objectives will be: This doesn't mean re-define the scenario objectives, but rather how does the OPFOR interact with the scenario objectives and set sub-objectives to the main objectives

3. Determine HOW the OPFOR will employ their organization to achieve those objectives. How will the OPFOR fight to achieve their sub-objectives and ultimately achieve the scenario objectives? The designer must determine if they will lightly defend one area in support of a more important area. Will they launch counter-attacks or will they fight a delay? This is where the designer needs to really be creative, but also work within realistic constraints of the pre-defined OPFOR task organization (#1), so it doesn't become a gamey mess of Nazi Zombie Waves. In a professional military setting this step is done by someone outside of the scenario process. Someone who is totally unbiased with the exercise.

I know this post comes off very aggressive. It's attempted to be more informative, if anything. OOB is a great game with a unique set of features a really well considered game rules. It is also very accessible to create new content and offers a fun cooperative platform. The frustration is tied to the CONTENT - both vanilla and user created. It's just very gamey and unfun.

[stevefprice]

Out of curiosity which scenario specifically is your example from?

[StuccoFresco]

Spawning overwhelming amounts of units is a (poor) way to overcome the inability of the AI to match a decent player's tactical ability. I agree with you it's frustrating. I also don't like when the supposedly flanking move seems to meet much more enemy troops than what was supposed to be the main enemy force.

I always prefer to give the player space and time constraints, as well as starting them already at a numerical disadvantage but with the possibility of taking the enemy apart piecemal with good maneuvering. Enemy reinforcements are used, but they happen a few times, and are usually announced via event and/or during mission briefing.

[MikeAP]

Out of curiosity which scenario specifically is your example from?

It doesn't matter. I'm not trying to focus my complaints towards anyone specific. I will say though, I did download it from the community box in game.

Spawning overwhelming amounts of units is a (poor) way to overcome the inability of the AI to match a decent player's tactical ability. I agree with you it's frustrating. I also don't like when the supposedly flanking move seems to meet much more enemy troops than what was supposed to be the main enemy force.

I always prefer to give the player space and time constraints, as well as starting them already at a numerical disadvantage but with the possibility of taking the enemy apart piecemal with good maneuvering. Enemy reinforcements are used, but they happen a few times, and are usually announced via event and/or during mission briefing.

Yes. Exactly.

Put the player in a position where they must make decisions based on mission variables and the analysis of the terrain, enemy, time, and friendly forces available.

[conboy]

MikeAP,
The problem with the AI is that it will not mount a sustained, coordinated, determined attack on a dug-in defender or go deep into the enemy's rear. Erik2 and I were working together on the Battle of the Bulge 1 v AI (German AI attackers) and I had the responsibility of getting Peiper's battalions to attack West through weakened allied units. After a while, 5 turns or so, the AI always stalled, no matter how I set the attacking forces' characteristics.
Units wander off from the objective, don't concentrate for an attack, disperse artillery, the AI doesn't rest and reinforce units, etc, so a determined, focused attack on a critical point is difficult to achieve without adding units after the initial attack. Actual orders of battle for the AIs attacking units are not enough to conduct an effective attack, they have to be supplemented by follow-on units that aren't in the original set-up. Even then, the follow-on units start doing the inane things the initial echelon did.
OoB AI is very good on defense and even good on small counterattacks but for the main, deep-penetrating attacks on the enemy rear or strong points, it's not good.
BTW, I was never able to get the Peiper attack to work.

Good luck, please let us know what you come up with if you find a solution!

conboy

[MikeAP]

Another example of awful scenario design-

Playing through Allies Victorious DLC- M03 - Falaise Gap.

I lost my first attempt by 1 turn because I had supply issues for several turns.

Not just once - but several times, the Germans ran suicide counter-attacks deep into my rear areas which caused problems with supply. When the hell did the Germans mount suicide rushes into the rear areas of the Allied lines in '44? Just curious where that came from...

What's even more irritating - nothing is mentioned in the briefing about these suicide attacks into the rear areas, or any threat on supply lines. The entire focus for the player (mission/objectives) are on the complete opposite side of the map, so ...it just felt like a trick-f*ck of a scenario. A real 'GOTCHA!' moment - not because the AI is brilliant or the scenario is good - but again, because its just a total surprise and not even something that could've been prevented.

This is the kind of 'GOTCHA!' stuff that I find with most of these terribly designed vanilla scenarios.

The screenshots from the scenario that I posted in my original post: my buddy and I are still playing this scenario, for kicks. We're coming down to the last few turns and still nowhere near our objectives - I literally saw a German unit spawn 3 hexes in front of me - so even at turn 24 of 30, the Germans are getting fresh reinforcements and income per turn.

[MikeAP]

MikeAP,
The problem with the AI is that it will not mount a sustained, coordinated, determined attack on a dug-in defender or go deep into the enemy's rear. Erik2 and I were working together on the Battle of the Bulge 1 v AI (German AI attackers) and I had the responsibility of getting Peiper's battalions to attack West through weakened allied units. After a while, 5 turns or so, the AI always stalled, no matter how I set the attacking forces' characteristics.
Units wander off from the objective, don't concentrate for an attack, disperse artillery, the AI doesn't rest and reinforce units, etc, so a determined, focused attack on a critical point is difficult to achieve without adding units after the initial attack. Actual orders of battle for the AIs attacking units are not enough to conduct an effective attack, they have to be supplemented by follow-on units that aren't in the original set-up. Even then, the follow-on units start doing the inane things the initial echelon did.
OoB AI is very good on defense and even good on small counterattacks but for the main, deep-penetrating attacks on the enemy rear or strong points, it's not good.
BTW, I was never able to get the Peiper attack to work.

Good luck, please let us know what you come up with if you find a solution!

conboy

I would suggest to then focus on things the game does well instead of trying to jam a "square peg into a round hole". I've learned this lesson with other games that I've designed for. You have to work within the capabilities of the medium that you're using.

My design philosophy has always been:
Focus on the things that work, then focus on the things that are fun, interesting, and challenging

[terminator]

The AI is difficult to program for the attack, the solution is to play on the number of attackers (too many attackers).
It is also interesting that the attacker can see the layout of the defenders so that the attacker controlled by the AI does not attack blindly, using ground or aerial reconnaissance units for example.

[Erik2]

Rather than simply adding more AI units I like to use unit experience, AI commanders and AI specialisations to even the odds. I would also use AI faction minefields if these are part of the scenario.
You can help the AI by assigning the units to several smaller AI groups and give them separate orders suited to the situation. Scripting trip wire tasks/unit spawns may work fine.
The editor is quite flexible, but it may be a lot of work getting it right. Especially to test the beast.

[StuccoFresco]

Rather than simply adding more AI units I like to use unit experience, AI commanders and AI specialisations to even the odds. I would also use AI faction minefields if these are part of the scenario.
You can help the AI by assigning the units to several smaller AI groups and give them separate orders suited to the situation. Scripting trip wire tasks/unit spawns may work fine.
The editor is quite flexible, but it may be a lot of work getting it right. Especially to test the beast.

Yeah, more AI teams are always the way to go. It's a lot of work, but it's still the best way. It wll still not match the way a player can concentrate and move around forces, but it's the best chance we modders have. Encirclements are especially painful to code, unless you give the AI a wide corridor to exploit so it doesn't wander around.