Surprise, Activation, and the Solo Wargamer

What is "solo gaming"? Well, clearly it is "playing a game by oneself", but it has to be more than that. To some people, however, it is not. Anyone who says "I just play both sides to the best of my ability" sees the definition of solo gaming as being just that and no more. Others believe that there has to be some "surprise" element added to the game because the player has "perfect knowledge". Even Stuart Asquith agrees with that idea, to a great extent.

A frequent comment made regarding solo war gaming is "What's the point? The player controls both sides and thus always knows what the opposition is going to do and thus there is no element of surprise." A not unreasonable statement, so irrespective of the period in which the solo player intends to war game, or the manner in which the games are to be conducted, various solo play mechanisms and techniques will be needed so that the soloist can introduce a random element into the unfolding situation.

- Stuart Asquith, "The Partizan Press Guide to Solo Wargaming"

I am on a forum for solo wargaming and there are a stream of questions like "what rules should I to wargame solo?" or "how can I play rule set X solo?" The responses generally fall along these lines, like:

• "Rule set X uses a card activation method, so you are halfway there."
• "You should choose rule set X because it uses cards to determine which side acts next. Solo rules should do that."
• "You should first determine how the player side is going to deploy, then randomly determine how the opposing side should deploy its units."

Actually, those comments also come from authors of books and articles on solo wargaming so they are in good company. All that said, I think the comments miss the mark. Let me explain why.

Surprise in Wargaming

Let me ask you a question, and be honest now: how many times were you truly "surprised" in a wargame by your opponent's move and because it was a really good move? The last one I can think of - and it did not happen to me - was when AlphaGo made move 37 in game 2 against Lee Sedol. That move so astounded Go players around the world that the Go community was abuzz about how it came up with the move. Lee Sedol took a full 15 minutes to recover from the shock before he could make his next move. Have you ever had such a game? I have not; not even close.

I can remember all of the times I was "surprised" (and the move was good) and they were always when I was playing a new game and did not know all of the rules. It is a game that has a fairly straightforward core set of rules, but with too many "special abilities" that were constantly being added as new supplements came out for you to buy. Essentially, a unit made a move that was illegal, according to the core rules, but perfectly legal according to the special rule in the supplement that I had not yet purchased. You know that kind of game. There are plenty of them out there.

Now, there have been some moves where I was surprised that the player made them, but only because it was either a bad move or a risky one. On the latter, some worked and some didn't. I made a few of those "you're a hero if it works, zero if it doesn't" type moves myself. But at the end of the day, playing the game is about exploiting the missteps your opponent makes while minimizing your own. The last thing you want to do is introduce a random element to force you to misstep.

Does that mean I don't agree with adding in random chance elements into the game? Of course not. The thing is, the games already have them. Is adding another element really to facilitate solo play?

Card Activation

Let's take card activation as an example. Is adding card activation for units conducive to solo play?

First, there are three types of card activation mechanisms. The first use cards to determine which side activates next. When a red card is drawn, the red side acts with all of its units. A variation of that is that the card indicates that one unit on that side activates next. The last type of card activation - and one used infrequently - is assigning a card to a specific unit. When the card is drawn, that specific unit then acts.

Let's think about this in terms of decisions. Unless you are in the "play both sides" camp, what you are looking for in a solo system is a means for answering the questions every time a decision point comes up. So, what are some of the basic questions?

• Which side gets to act next?
• Which unit gets to act next?
• Which action(s) will the acting unit take?
• How will the unit execute that action(s)?

If your system is not answering these questions then it is not helping you as much as it could. You, the player, are making decisions for the non-player that allows you to let your bias creep in.

Which Side Gets to Act Next?

As you can see in the table below, all activation methods, including the traditional IGO-UGO method, takes away the decision of "which side acts next" from the player.

Activation Type	Result
Card to determine side	Yes
Card to determine side that chooses a unit	Yes
Card to determine a unit	Yes
IGO-UGO	Yes

Which Unit Gets to Act Next?

Unless you are using a mechanism that assigns a unit (or sub-command) to a card, card activation does not answer the question of which unit you should act with next. If you provide no mechanism for making this decision, you leave it up to the player to make the choice.

Activation Type	Result
Card to determine side	No
Card to determine side that chooses a unit	No
Card to determine a unit	Yes
IGO-UGO	No

Which Action Will the Unit Take?

Activation Type	Result
Card to determine side	No
Card to determine side that chooses a unit	No
Card to determine a unit	No
IGO-UGO	No

How Will the Unit Execute Its Action?

Activation Type	Result
Card to determine side	No
Card to determine side that chooses a unit	No
Card to determine a unit	No
IGO-UGO	No

As you can see above, none of the activation methods answer the remaining two questions that pertain to the unit acting. That is because activation methods stop providing use once you get past the decision on which side or unit activates next. Put another way, card activation is not a solo gaming mechanic unless you use it for unit assignment. It answers no more questions than IGO-UGO does (again, with unit assignment by card as being the sole exception).

What card activation does is create "surprise" to the player as to which side will get to do something. It is a command-and-control mechanic, not a solo mechanic. However, it has side effects that you need to consider.

Variable Activation

Whether variable activation - activation where the order of the side acting varies from turn to turn - comes into play the most is where each activation only allows a very granular action. For example, if a unit can only move or fire in a single activation then moving into the weapon range of the enemy becomes a significant decision. In IGO-UGO games when a unit moves into range the enemy is guaranteed the chance to fire first. Changing this to variable activation does not provide that guarantee any more. A unit could activate at the end of one turn, move into range, and then in the next turn draw first, allowing it to fire before the enemy can respond. It is actually these sort of "double activations" that players attempt to set up and exploit as they are so effective.

Even if you are using rules that allows units to move and fire, variable activation can have an impact. You could move in and fire on one turn and then fire again on the next turn with a lucky draw.

So, you might be thinking, what is wrong with that? Some games are rather delicately balanced regarding the volume of fire in relation to distance moved and the passage of time. Changing that affects the core balance that the rules author put into their rules.

The main point is that these sort of mechanics are attributed as being "solo friendly" and my point is that they have nothing to do with solo gaming. These are ways to add additional random chance elements into the game, often in the name of modeling "the fog of war", "chaos", or "friction". A player will either like such a mechanic or not, but do not attribute it to facilitating solo game play. If a mechanic does not take away a decision from the player when acting with a unit for the non-player side then it is not a "solo gaming" mechanic.

Injecting Chance into Non-Player Decision-Making

Another response I see in forums regarding "how to game solo" is to use random chance elements to answer the type of questions above. Donald Featherstone calls that 'instant' solo wargaming.

But even the most volatile spontaneity can be dampened in the case of the solo-wargamer if he has to waste a large part of his precious few hours in so organizing the battle as to give him the enjoyable interplay of tactics that comes from dividing himself down the middle and being two generals at once. Briefly, this means that he requires some ready-made method of 'instant' solo-wargaming that enables him to set up armies and get on with the battle in a manner that allows for a realistic demonstration of both tactics and the fluctuations of the fortunes of war. He need not despair because, if his inventive mind has not already though up a system of his own, then there are a number of other methods that he can utilize or adapt to suit his own requirements.

- Donald Featherstone, "Donald Featherstone's Solo Wargaming"

I admit to using some of these mechanisms myself, on occasion, but for me the goal is that it is a fallback for when you want to game, but don't care as much about the quality of the result. That typically occurs when I am testing out a new set of rules. I tend to push the boundaries in such games to see if the rules allow "crazy" results and whether or not it punishes "bad" tactics. More often than not though I simply play both sides without regard to "being surprised". I tend to have less bias (or at least I like to think so) when I am simply testing than when I am gaming for fun.

The simple fact is, a random die roll that includes a chance to select a sub-optimal option will always be, at best, on par with the decision that a thinking human would make, given all of the factors present. A far greater portion of the time it will produce an inferior decision. Multiply that over all the decisions that will have to be made and it is not hard to see why the more you inject random chance to make decisions for your non-player forces, the dumber and more erratic the opponent will be.

Playing Against Mr Babbage

Whenever I read someone's post about Dan Mersey's The Men Who Would Be Kings (TMWWBK) inevitably someone will mention the solo rules Playing Against Mr Babbage (PAMB) that are included in the rules. To be honest, that is the only reason I purchased those rules. So, what are they?

PAMB is essentially a programmed opponent, mixed with some rule changes to TMWWBK to compensate for the fact that the programmed opponent will never be as smart as the player. In general I do not like modifying rules, even for the programmed side. As a solo gamer I acknowledge that gameplay will never be as challenging as a live opponent, so I solo game to try out tactics or strategies, refine lists or decks, or to create an interesting or amusing narrative.

So, how is PAMB programmed? Basically it is a simple rule-based system with priorities identified for each unit type and general actions to take if the unit-based ones do not apply. The first system is for programming the natives, but includes a second system for programming the Imperial forces. Here is a small sampling, to give you an idea. (But not all. You should buy the book.)

1. I like my shooting units to find cover and stay there. Failing that, I like them to advance to short range and fire, avoiding melee.
2. I like my melee units to advance quickly into contact, towards the closest enemy they can see, except as noted in rule 3.
3. ...

Pretty simple, but descriptive enough for the player to reasonably determine what should be done when it comes time to act with a unit. Note, however, that the rules are not written in order of precedence (as indicated by rule 2 stating there is an exception in rule 3). So you have to read through all of the rules to see which one makes the most sense.

As stated previously, PAMB also has rules that modify the TMWWBK rules, such as randomly appearing native units (similar to the Two Hour Wargames' PEF concept), native action modifiers (which include the unit leaving the table), and recycling native units.

My dream has always been for solo gamers to be able to 'trade' programmed opponents. My schemes have always been too complex - essentially requiring other gamers to have computer programming skills to define the rules - and I can see that PAMB's method is a good start towards codifying such rules for programmed opponents. The one change I would make is that rules have to be in order of precedence and that when you hit a rule that applies, you stop evaluating the rules.

I still think that generic opponents would be harder to create than ones for a specific scenario for a specific time period. Maybe that will be my next project.

New Solo Game Design Book

Upcoming Book Purchase

I have purchased the first two volumes on game design by Wally Simon (and edited by Russ Lockwood) and really liked them. Now On Military Matters has published a third volume called SOLO SECRETS OF WARGAME DESIGN: Volume 3: A Tabletop Toolkit of Ideas, Analysis, and Rule Mechanics. The book's description is listed as:

The third volume of the SECRETS OF WARGAME DESIGN series concentrates on solo wargaming ideas. Peel back the layers of game theory to generate great solo games for a variety of periods. Delve into the real nuts and bolts of game mechanics to create a more compelling – and often more entertaining – tabletop battle.

• Simple Solo Guidelines: Morale, Data Sheets, and a Napoleonic River Line Attack
• Good Morning Vietnam: Cards, Squares, and Sequences
• WWII Efficiency: Too Much Damage, Loss Points, Reaction Points, and Combat Points
• Napoleonic Equality of Inequality: Simon's Theory of Fire Power
• Scenario Generation: A Line in the Terrain
• Speedy Solo Siege: Zones, Building Points, and the Passage of Time
• Third Time's Charm: Modern Solo Squad Skirmish
• Renaissance Squared: Grids, Groups, and Grinds
• Bad Blood in the Balkans: WWII Command, Control, Activation, and Quality
• Rules are Greener: American Revolution Response Chart and Efficiency Levels
• Solo Hastings: Flights of Fancy?
• NATO Rescue in Serbia: Pop-Up Squads, Prisoners, and Running Battles
• BONUS: The Wally Quadfecta: Universal Guidelines for Good Game Design

The book is $19 and is 50 pages.

Order Writing

The Solo Wargame forum on Yahoo has a thread about writing orders which triggered a thought or two in me. I was thinking about my younger days when I used to play the Colonial Skirmish Rules and how signals were used to change orders. As my first test of the Finite State Machine was to be using my Meso-americans and the Saga rules, I thought I would incorporate orders into my game. As signaling with banners is very much a part of that style of warfare, I thought I might come up with some rules for it, despite the fact that command and control doesn't really exist in Saga (or at least is largely abstracted away). I hope to have a battle report on that next, using a hex grid to regulate measurements. If you read my Dale's Wargames blog, you know that I am caught up in a BattleLore tournament over Vassal, so it is taking up a fair portion of my time.

Finite State Machines and ... Miniatures Gaming?

A word of warning: this blog is basically my way of thinking through ideas, even unfinished ones. I find that I do better writing things down rather than just sitting there pondering, and hoping I remember the good bits later. I hope you will bear with me as I sort through my thoughts and hopefully you will find something of interest and use.

The first AI technique discussed in Programming Game AI By Example (see last blog entry for the reference to this book) is a classic: finite state machines (FSM). In the words of the author, the definition of an FSM is:

A finite state machine is a device, or a model of a device, which has a finite number of states it can be in at any given time and can operate on input to either make transitions from one state to another or to cause an output or action to take place. A finite state machine can only be in one state at any moment in time.

The idea being using an FSM is to decompose an object's behavior (in our case, a unit's behavior) into manageable  states. A simple example provided in the book is probably one most of us can understand, that of the "ghost" in the video game Pac Man. The normal state of the ghost is to "chase Pac Man". However, when Pac Man eats the power pill, the ghost switches from "chase" to "evade". The ghost reverts back to chase mode once the timer runs down and the power pill wears off. The rules for the ghost might look something like this:

rule: if in 'Chase Mode' and 'Power Pill Active'
    then switch to 'Evade Mode'
    else continue 'Chase Mode'

rule: if in 'Evade Mode' and not 'Power Pill Active'
    then switch to 'Chase Mode'
    else continue 'Evade Mode'

The same sort of behavior can be encapsulated for units in a game. This is essentially what I was referring to when I discussed creating and using programmed opponents on this blog. What makes a FSM "better" than the rule-based approach I discussed previously is a matter of organization. The rule base mechanic simply lists a set of rules to follow, in a strict order, and to stop evaluating the rules once you find a rule that applies. What makes developing this rule base complex is setting the evaluation order of the rules. It is not usually impossible to determine the correct order, but sometimes it gets so complex that the gamer decides to forego the whole process and just "play each side to the best of my ability", which after awhile can get pretty stale and predictable.

Put another way, a single rule base evaluates a set of conditions in a strict order. An FSM allows you to evaluate conditions in order based upon the state you are currently in. You can essentially have a set of rules associated with every state.

Sample Game

Okay, enough of the theoretical; you can get that from any book. Let's make this concrete. The image to the right shows the start of a skirmish game. The red player (you) have four units: two Warriors (the groups of eight red circles, on the flanks), one Bodyguard unit (the four red circles in the center), and the Warlord (the large red circle in the center). The blue player (represented by a non-player general) also has four units, but they are Warriors (group of eight blue circles on the left), Levy (group of 12 light blue circles in the center, on the hill), the Bodyguard (four dark blue circles on the right), and the Warlord (large blue-in-white circle on the right).

The player's side is oriented towards melee. In fact, all four units have no missile weapons. The non-player side is oriented towards shooting; all four units have missile and melee weapons, but are weaker in melee than their red counter-parts. The basic blue battle plan is to:

• Use the Levy and Bodyguard to fire upon and weaken the opposite red unit.
• When the red unit has been weakened sufficiently, the Bodyguard will charge in with the Warlord and finish it off.
• The Warrior unit will skirmish with the red units, attempting to engage them sufficiently so they do not go support their unit being attacked by the Levy and Bodyguard, but not so heavily engaged that they become overwhelmed.

For now, I am going to ignore the orders of the Levy and Bodyguard and simply focus on the Warriors. Basically, their mission is to tie up one, and hopefully, to three units. There are three basic states (although this may depend upon the rules you use):

• Neither in missile or melee range.
• In missile range.
• In melee range.

There are all sorts of variations or "sub-states", if you prefer to call them that, but I like to think more in terms of conditions. Rather than having a state for "in missile range of the Warriors" and another for "in missile range of the Bodyguard" and maybe yet another for "in missile range of the Warriors and Bodyguard", etc. I simply treat it as the state "in missile range", with conditions indicating which units that applies to. I can then have rules that list the preferable target in order of precedence.

Range Bands

One thing to note about these example states is that you can visualize them as range bands. Consider a unit that moves 6" and has a missile range of 6". The opponent has a movement of 6" also. Technically, if you are in the Melee band, you are also within the Missile and Move bands. This helps us understand that we need to set an precedence order for checking which band, and thus which state we are (or should be) in: Melee, then Missile, then Move.

So, let's think about this, We start the game in the Move band, i.e. outside of both melee (6" move) and missile (6" move plus 6" shoot) range. The first condition to check is whether the enemy moved and we are now at a different range. To keep it simple (for now), let's assume we have the first move, and the situation is as indicated in the map above.

My basic program is to move towards the closest enemy and get into a position where the unit can fire its missiles, and if possible, move back. If the enemy approaches, the unit is to back off, firing missiles as it retreats. (Note: the basic reason for firing then retreating is because the unit's missile range is 6", the same as the movement distance of the enemy. That means if the unit moves forward to fire, it put itself into Melee range automatically. Thus it wants to back out of Melee range before the enemy unit can react. Your rules may not allow such a maneuver, thus this model will not work for you. The rules I will be testing this with, Saga, do allow this sort of behavior.) One other point might be that where the unit can either move and fire at either the Warriors or Bodyguards unit, it will choose the more valuable unit, the Bodyguards.

As the rules I will be testing this with is Saga, for those unfamiliar with the rules here are the basics you need to know:

• Each turn the player gets a certain number of dice to roll.
• Each die rolled indicates what abilities the player can activate during the turn, which includes ordering a unit to take a single action.
• A unit can be ordered to take more than one action per turn by committing more dice to that unit.
• The more actions a unit takes in a single turn, the more fatigued the unit becomes.
• When a unit takes enough fatigue, it is exhausted, and it may not move or shoot until it has rested sufficiently.

This last point brings out another state required in our model: Exhausted. If a unit is exhausted, it can do nothing else before resting. This state will take priority over Melee, Missile, or Move.

Back to our Warrior unit. An ideal turn would be to move forward into missile range (6"), shoot at the enemy, and then retreat back out of the enemy's Melee range (6"), for a total of three actions. As this sequence causes fatigue, a fourth action for resting would make this a truly ideal turn. I can tell you now that committing four actions to a single unit is excessive in most Saga games, so we need to look at the next best option, which is move in, shoot, and move back in one turn, and rest on the alternate turn.

If we put all this into an ordered rule base it might look like:

State: Exhausted

rule: ...

State: In Melee Band (within any enemy unit's movement distance)

rule: ...

State: In Missile Band (movement distance + missile range to an enemy unit)

rule: if not overly fatigued
    then move forward into missile range, shoot the enemy, and retreat out of the enemy's melee range

rule: if overly fatigued
    then rest

State: In Move Band (farther than Melee and Missile bands)

rule: ...

This now covers the basic program of the unit for harassing the enemy with missile fire. We still do not have specifics about how the unit approaches or retreats, or if it fires at the enemy in any specific way (for example, to attempt to kill off a special character like a leader), but we have the basics on how it should act.

As I indicated in the articles on Battle Card Systems and Hand Management, rule bases should be built up over time, rather than trying to do it all up front, at once. I will close this out for now and try and write up the skeleton for the programmed units for this scenario (i.e. all four units) and post them. Maybe then someone else can take them up and play a game with them and provide a critique.

Messing with the rules

40mm wooden Dark Ages Warlord

I have always been an advocate of playing my solo battles "strictly by the rules". In fact, I do not like injecting new game rules in; at best (worst?) I have used dice to apply a personality to the non-player general (NPG) or to determine which course of action to take, but the thought of fudging the numbers in favor of the NPG never crossed my mind.

Recently I started playing Saga, a set of Dark Age skirmish rules (see my review on my Dale's Wargames blog) that have some really interesting game mechanics. As a part of learning rules I typically play a game or two solo, usually by playing the best you can for each side. Probably the absolute worst way to solo game, from an "excitement" viewpoint. It was while playing a learning game yesterday (actually, it is still set up and on-going) that I ran into a situation where I could play a reaction ability against an enemy unit and then I thought "wait a minute, it would be better to react to this other unit". I then realized that the other unit I was going to react to had not declared its action yet; I was reacting to an action my "opponent" (me) had been thinking about, but had not actually done. Put more simply, I was anticipating my moves as the opponent using knowledge I should not have possessed. This is always a problem with solo gaming and "playing the best you can for both sides". This creeps into our solo gaming all the time and usually manifests itself as a bias towards one side or another.

That got me thinking that I need to get back to working on solo gaming mechanics – specifically how to model what the best (?) move your NPG should take – and set aside the campaign ideas for the moment. (Don't worry, I am sure that I will come back to them.) I decided that one way to help me solve this problem was to take a look at what computer programmers are doing for AI in their games. I went out and purchased Programming Game AI By Example, by Mat Buckland.

Solo game with wooden Napoleonic figures

I've just started reading the book, but as I get to the interesting bits I figured I would report it here, and how I might apply the idea to a miniatures gaming AI. Well, it did not take long to hit an interesting bit, which is why I am writing this entry. The author was making the point about academic AI (strong) versus game AI (weak), and that the latter is really about "the illusion of intelligence". The designers of the AI for Halo discovered that their playtesters could be fooled into thinking the game's AI was more intelligent simply by increasing the hit points of the enemy. In one test session they game the enemy low hit points, allowing them to die easily, and 36% of the testers thought that the AI was too easy while 8% thought the AI was very intelligent. After increasing the enemy's hit points, 0% of the testers thought the AI was too easy and 43% thought the AI was very intelligent!

This led me to question my own cardinal rule about changing the game rules for the NPG. In the past, solo gaming had been a vehicle primarily to learn a set of rules or tactics, or to prepare for a tournament (see the work I did on De Bellis Antiquitatis Solus on the Solo DBA Development forum on Yahoo). As my face-to-face opponents wax and wane, especially as I use rules or periods that others are not interested in, I find that solo gaming must become more of an option for personal entertainment. Put another way, game solo or don't game at all. (Well, or game the popular rules, like Warhammer 40K, Warmachine, etc.)

If I did decide to break the rule by altering the rules for the NPG, what might it look like? Would I, for example, give an automatic +1 to the combat roll for the DBA NPG? What would the impact of such an action be? For one thing, it would require that I think and plan my attacks much more carefully. I could not rely as much on "the odds" as they would now be against me. That might certainly sharpen my play and allow a sub-optimal move by the NPG win out. Put another way, it takes the pressure off of making perfect moves for the NPG and allows "good enough" moves to present me with a surprise.

So, what are some of the other ways you can break the rules that lead to a tougher opponent, forcing you to sharpen your game? There are numerous ways to change or add to the rules in the NPG's favor, of course. I am looking for mechanisms that increase toughness, but don't essentially break the game for the player. Increase the challenge, but not the frustration. I would like to hear from you if you have ideas. In the meantime, I am going to keep reading the book and figure out ways to translate the concepts to the tabletop.