HOW TO: Design a Mod

[dh_epic]

I find a lot of people have a great imagination, and can usually come up with something that sounds cool enough that they'd get at least a FEW other people saying "hell yeah! I want that too!"

But adding something can have a real danger. There seems to be two main problems when it comes to adding stuff:

1) OVERCOMPLEXITY: people add something that's much like a game in of itself. They imagine a whole new system with dozens of buttons, and complex interacting rules, with another dozen exceptions where the rules don't apply. This might be fun and it might even work, but it's so brutally complex that you'll never make it. Or if you do, you'll end up with something that only the most hardcore of players will have the patience to learn and thus enjoy.

2) FUNCTIONLESS FLUFF: Sometimes people add a feature that really does nothing to the actual *game*. Let's say someone adds a 'crime' element to the game, with a random 'criminal' unit popping up that you need to kill from time to time. What's one criminal unit going to do? Maybe pillage a tile before the player takes care of it. This doesn't add a new strategy or choice to the game. It adds fluff in the name of realism.

One of the most insightful things I heard about Civ 4 is that it's a series of extremely simple systems that interact to create a complex game. In this new SDK era of modding, I think people have to keep that in mind. You should be able to explain the basic system in but one or two sentences: Religion spreads from city to city via trade routes and missionaries. The owner of the Holy City gets more gold the more that the religion spreads. Or Specialists and Wonders generate Great People Points. When a threshold of GPP is reached, a Great Person pops out capable of researching a tech or giving some other special bonus. If it's easy to explain, then it's probably easy to implement, balance, and understand.

 

[Kael]

2) FUNCTIONLESS FLUFF: Sometimes people add a feature that really does nothing to the actual *game*. Let's say someone adds a 'crime' element to the game, with a random 'criminal' unit popping up that you need to kill from time to time. What's one criminal unit going to do? Maybe pillage a tile before the player takes care of it. This doesn't add a new strategy or choice to the game. It adds fluff in the name of realism.

Thats a really good insight, I couldn't put my finger on what I didnt like about some (usually very creative) ideas until you said that. I agree 100%.

We have a lot of people saying that one of your goals should be to keep your design as simple as possible. Is there anyone who disagrees with this? Does anyone think that the game is to simple and adding complexity makes it better?

 

[Jeckel]

I've never taken any programming classes, but I've softcoded on mushes and muds for about 6 years, creating various systems from character creation to weather and I've found that first you need a generic concept that can be stated in a sentance or three that gives the basic idea of the mod.
"DnD Mod is a small mod that limits civs to 12 hero units and one city and pits them against each other and the wandering barbarian monsters."

Once you have that you need to determine what areas need to be modded. By the above DnD example, need new unit graphics and decide if unit limiting can be done in xml or python.

When you have these two things my advice is to jump in and start messing with the code. Afterall the worst that can happen is a game crash, just remember to keep back up copies or some sort of record of what you change so you know where to look for errors.

Originally Posted by dh_epic
2) FUNCTIONLESS FLUFF: Sometimes people add a feature that really does nothing to the actual *game*. Let's say someone adds a 'crime' element to the game, with a random 'criminal' unit popping up that you need to kill from time to time. What's one criminal unit going to do? Maybe pillage a tile before the player takes care of it. This doesn't add a new strategy or choice to the game. It adds fluff in the name of realism.

This is a good point. The example pilliaging criminal is empty and almost just annoying because it has no real effect. It needs to have some sort of impact on a game. The player needs to interact with it in some way. Even if the effect is "bad" and it makes the player "angry" that is better then indifferense and bordem and its often that angering aspect that keeps you playing for hours and hours. Like, how bout if your city had X unhappy people for Y turns, then Z population would leave your city as Riot units and begin pillaging the squares around the city and any that are killed won't return to the city after N turns of rioting. This then has function, if you don't keep your city happy or policed then it will be negatively effected, and also fluff, instead of numbers being changed behind the scenes there are onscreen actions that can be seen. On top of that it has a way for direct player actions to effect the outcome, ie they can attack the Riot units and destory them or just gaurd specific resourses limiting or even nulifing the pillage effect.

It goes back to what has been said about System vs. Feature, a feture generally have a very liner cause => effect stucture(unhappyness => criminal unit => tile or two pillaged), where as a system tends to have multiple causes, effects, and sideeffects(unhappyness and unhealthyness taken into acount for spawning, decisions given to player, outcomes variety and interactable).

We have a lot of people saying that one of your goals should be to keep your design as simple as possible. Is there anyone who disagrees with this? Does anyone think that the game is to simple and adding complexity makes it better?

I both agree and disagree with complexity in mods. Take FfH, or any of the other mods that rewrite the game, these are very complex and intricate endevors and many many people take on the learning curve above and beyond the basic Civ4 for one because it is well done and exicuted. Its basic premise required alot of complexity, that complexity required alot of time, and the production team put that time into it. If these points are met then complex mods are fine.

On the flip side many complex mods don't require the complexity or the modders didn't put in the need time and in these situations it doesn't work out well. These leads back to designing before modding, laying out what needs to be done and what doesn't and how that translates into code.

Now it is good advise to start simple(whatever simple is for your current coding profishensy) and build upon that. But the complexity of the final outcome is usually more dependint on coding knowledge, the desired result of the mod, the time put into it, level of realism, ect.

I would say, Start simple and keep it effecient.

As to the greater question of a simple game vs a complex one, I personally like the middle ground, I enjoy a good amount of micromanagement, but it has to build around an acual game. I don't mind doing some math when it relates to my future actions, but I don't like having to micromanage current actions(like figuring combat odds when I'm battling units).
I like how Masters of Orion 2 was set up, but three went to far with regions and buildings and different sizes and all the screens you had to click through just to do stuff. They got way to complex with the UI and the game mechanics.
Civ4, I think came out with a nice balance in the vanilla, less micromanagment, more gameplay, but I could stand for some more micromanagment, I miss acual rioting cities from unhappyness, gave a reason to go into the city screen.
This really is a preference and modders are going to find a market for mods from the most simple ot the most complex if it is done right and has a half decent premise, so all you future modders keep at it and sick with the number one rule always

Make a mod that you enjoy playing.

If you do this then you always win.

 

[Zurai]

One thing that I don't hear mentioned enough with respect to game design (warning, this doesn't apply quite so much to Civ4 mods specifically as it does to design in general):

Player rewards.

What is a player reward, you ask? Simple; a player reward is any reaction by the computer that gives a player positive feedback about actions they have taken.

OK, that's a nice textbook definition. So what, exactly, does that mean? Basically, player rewards are what keep the player wanting the play the game. A great example from a recent game: in Heroes of Might and Magic 5, whenever one of your units completely destroys a stack of enemy units, your entire army gives a visible and audible cheer. This gives the player a sense of accomplishment and shows them, without any annoying messages, that he just did something that he was supposed to.

Player rewards can take almost any form. They can be visible - particle effects, victory dances, a character pumping his arm. They can be audible - cheers, a fanfare, "DING!". They can even be tactile in console games with rumble controllers. Points are player rewards; so are powerups dropped from defeated enemies, experience levels, cutscenes, and so on.

The important thing to remember with player rewards is that it is very difficult to use too many and very easy to not use enough. You want the player to have a constant sense of accomplishment and pride/pleasure while they are playing; the more rewards you can feasibly fit in, the more likely they are to enjoy playing your game. The best rewards take multifaceted forms. For example, take Diablo 2: killing an enemy gives you a satisfying death animation, a death sound, increases your experience, usually yields money and items, and can result in cut scenes and additional audio rewards. Clearly, killing enemies is what the designers want you to do, and equally clearly killing enemies is what people enjoy doing. The correlation comes from the rewards the player gets for doing the action.



So, how does all this apply to mod-making in Civ 4? Not usually directly. Most of the player rewards are already handled by the engine - things like tech completion popups, unit acknowledgements, experience from winning fights, cities growing, and the like. What we, as mod designers, can do is to minimize the things that players do that do not yield rewards. Try not to add too many spreadsheet-like screens. Try to avoid adding new features unless you also add new rewards for using those features. Do your best to consider, as you design, what things you can include that will make the player want to keep doing whatever it is you're designing.

Always keep in mind that player rewards will not make a bad game design into a good game; however, lack of player rewards will always ruin a good game design and make it a game that is, at best, not fun to play long-term.

 

[Kael]

Great point Zurai and something I haven't considered. I will be thinking about that as I work on my mods.

 

[Commander Bello]

We have a lot of people saying that one of your goals should be to keep your design as simple as possible. Is there anyone who disagrees with this? Does anyone think that the game is to simple and adding complexity makes it better?

Yes, Sir, here!

There have been several posters in this thread stating that one should avoid complexity. I have to disagree with those statements.

Please let me explain.
Complexity and being complicated - at least in German language - are different terms - "Komplexitaet" (complexity) and Kompliziertheit (being complicated). Complexity is nothing bad per se, as long as it is not too complicated to understand. I think, this is a very crucial point.

Let me give you an example. Many people really miss quantitative ressources. The idea behing it is that those ressources would allow for a more "realistic" economic system, would allow for a solution of the old problem of "industrial" centers somewhere in the middle of a hill region and so on. No need to go too much into details.
Anyway, personally I think, that this system could easily be implemented (spoken from a very general point of view of course) because it is not complicated to understand - it just displays something all of us know from our day-to-day live. To make it work, would nevertheless require very complex algorithms, but this is just, what computers have been designed for - the numbercrunching.
As long as you can explain such a system by some sentences: "You have some deposits of a ressource, which will give you x, y and z amounts of that ressource per turn. The total of it will be distributed between your cities automatically. If you want to build something in a given city A and that city has a shortage due to this ressource missing, you will be notified and can chose from a list of cities which will redirect that ressource to the city A. The downside is that the redirecting city will have to postpone the completion of its current building task".
Not very difficult to understand. (I admit that this example has been drastically simplified to fit into this posting, but the explanation in principle would stay the same)

How the automatic distribution works, would nevertheless require some complex rules. But these are things the players doesn't have to know necessarily (of course, if he would be interested into the understanding, it should be possible for him to determine those rules).

Well, not to make this one become too long: Complexity is neitehr good nor bad by itself.
As long as it is easy to understand and to cope with, AND if it adds the feeling of more realism and / or more fun, I would go for it.

 

[Commander Bello]

Ah, I just read Zurai's great posting. I couldn't agree more with what he explained about the rewards. Especially, as Civ4 doesn't offer the rewards he mentioned (or not many of them).

Where have they left the palace or throne room, you could slowly build up?
Where is the reward to create and add more and more of the spaceship parts? I want to have a little animation, showing how the cabin fits to the main storage rooms, I want to listen to some electrical sound "zzzzzt" ...."bssss", I want to see the spaceship sailing in the orbit, with the moon in the background. I want to see a shuttle docking to it.....
Where are the funny advisors from Civ2? "Give me more soldiers, Sir!"

Nothing of those things adds directly to the gameplay, but they add much to the experience.
From time to time it is just fun and relaxing to lean back, have a proud look at your spaceship being half part completed, or waiting for this last piece of the tower of your palace. You discovered a tech as first? No need to have a free tech all the time - a nice addition to the palace perfectly serves the same purpose - the player gets a feeling of satisfaction, of really having achieved something.

 

[Shigga]

Complexity and being complicated - at least in German language - are different terms - "Komplexitaet" (complexity) and Kompliziertheit (being complicated). Complexity is nothing bad per se, as long as it is not too complicated to understand. I think, this is a very crucial point.

I agree. ANNO 1503 is a very good example: The multiple chains of production are highly complex but easy to understand bc they followed a very natural logic.

 

[Rashad]

The multiple chains of production are highly complex but easy to understand bc they followed a very natural logic.

At least in terms of a Civ4, I feel "complex" and "more is not better" and "the economy of choice" are all tightly bound together.

"To have options we need to have limits, we can’t do it all so some things need to be sacrificed to gain others."

This can apply to units, buildings, tech research, even what squares you choose to improve, or how.

For example, for your Crusades Mod, you might decide that the player might like to have the choice of a dozen different mounted cavalries (same gameplay statistics) in varied Templar skins, and I have to admit, the option does sound quite appealing, but you have to ask yourself "Why should the player choose one over the other ?" If you cannot answer that yourself, then the player most definately wont be able to, and the units are probably redundant.

But what does it hurt to have a selection of different sprites ?

The economy of choice. By dilluting the field with redundant choices, you are obscuring the direct choice that a player has to make, and in effect reducing the enjoyment they can gain from making a successful choice.

To use an example people are likely familiar with (many intimately so), FfH1 and 2.

I enjoyed the mods immensely. I thought the flavor was well done. The art, music, and quotes all served to further the immersion. Each unit and system was easy to understand, and contributed something to the end goal.

So everything should have been like roses and onions, right ? Wrong.

Although the buildings, units, and heros all contributed something to the mod, the role they played was not unique. For instance, a common choice in Civ is "OK, I need to build a unit.", and I gazed at a list of a dozen or so units. To narrow it down, I decide "I want the unit to defend my town.", but I'm still looking at 6 different choices. Unfortunately, I cannot determine what difference, if any, there is between the choices (at least not in a timely manner), let alone answer the question of why I should select one over the others.

Choice is not always a good thing, especially if the consequences of the choice are not apparent to the player. (or worse yet, contradictary).

 

[Kael]

Yeah, oddly enough sometimes the best thing that can be done to improve a mod is the removal of components instead of the addition of them. A difficult process because as designers we are often in love with the minutia of our mods.

As to Commander Bello's point Im glad someone spoke up for complexity. His comments about the number of resources reminded me of the "Sliding Scale" that usually marks the limit between simple and complex in my mind. The Sliding Scale means that instead of being on or off there are multiple stages for whatever mechanic you are discussing. For Commander Bello it was resources, you wouldnt simply either have or not have Oil, but you would have various degrees of having it, each with different effects (I hope Im not paraphrasing to badly).

We recently came across this in regards to alignments in our mod. Right now you are either Good, Neutral or Evil. There is no real sliding scale. It wouldn't be to difficult to add a players alignment as an integer value and modify it based on actions, and then determine which of the three major alignments you belong to based on your placement on that scale. But what did that complexity give you? Now you have a system to manage, something the players will need to understand, but does that change make the mod better?

Sliding scales aren't bad in and of themselves. They are everywhere. Units hit poitns are a sliding scale (they arent simply alive or dead), production is a sliding scale (it would be simpliar to allow a city to create a building each turn). There are to many to mention.

But in my opinion they should be avioded unless they give a very tangible benifit to the player. To Commander Bello's point about the amount of resources. That could be a really cool function for a mod if it is fully developed. I think that merely putting in resource numbers in would add complexity without much benifit, but if it is the part of a larger, more interesting econmic model it could be a great feature.

 

[Shigga]

Unfortunately, I cannot determine what difference, if any, there is between the choices (at least not in a timely manner), let alone answer the question of why I should select one over the others.

The problem you describe here is not a matter of how many choices there are but a lack of proper documentation and the absence of a self-explanatory quality.

 

[Commander Bello]

[...] I think that merely putting in resource numbers in would add complexity without much benifit, but if it is the part of a larger, more interesting econmic model it could be a great feature.

About just having some numbers more, I absolutely agree with you, Kael.
My idea goes further in my mind and I agree that it has to fit into a complete economics model (with all attention put to keep it simple and understandable).
On such a model I am currently working and putting my thoughts together.
I have to admit that I am not finalized with it yet, therefore I stayed that general as above.
As soon as I have put it onto paper, I will offer it as a proposal for the RaR4Civ-mod, and you may feel free to use it wherever you want.

But this will last for some time, so I won't go into any details which may very well change.

 

[Karhgath]

When talking about complexity we are talking about the functional complexity of the system. Maybe an idea looks really complex behind the scene, but can be understood logically with simple rules. There are no catches, or tricky exceptions, or gray-areas.

For example, let's take the following example. We want to add the idea of supply in civ4:


Units have supply that depletes when they are outside of your borders and when out of supply, units start losing HP. They can resupply when they can reach your trade network and the resupply amount comes from a supply treasury. Forts extends the resupply range of your trade network.


This is straightforward. If you take out the functional system of the above idea, here's what it would look like:


Each units have X supply.
They have a bar indicating the amount of supply they have, like HP.
Explorers and special units have usually more supply.
When moving in own territory, they do not use supply.
When moving in friendly territory, they use 0.5*X amount of supply.
When moving in neutral territory, they use X amount of supply.
When moving in ennemy territory, they use 2*X amount of supply.
When a unit reaches 0 supply, they start losing HP at the same rate as supply.
You have a supply treasury.
Each turn X supply is added to your supply treasury following the formula ABC.
Roads connected to your trade network, or that can reach your trade network, unencumbered by ennemy units and cities, can resupply units that are idle on that road for X amount. This amount is taken from your supply treasury.
Forts, if controlled by the player, and connected to a road that can ressuply, increases the range of the resupply to 3 tiles around the fort.
An indicator on each units indicates if a unit can be resupplied.
An indicator on Forts indicates if they could be used to resupply or not, whether you control the fort or not.


Would you consider that complex? I wouldn't. It's logical, and can be expressed without much complexity. There are not really any exceptions and it would be hard to find faults or gray areas.

Overall, it might include a lot of parts, and be complex to implement, but it's made of simple components and rules. It's pretty straightforward. Complexity of implementation is another topic entirely.

I'm not saying anything about the validity of the above system but it's not that complicated. When you start having to think about exceptions, special cases, interpretations, abuse, etc. you need to rethink the complexity.

Now, add resupply units to the above...


Some units can resupply others.
Any other units in a 1 tile range of the resupply unit can resupply. The amount is taken directly from the supply of the resupplying unit and not the supply treasury.
The resupplying unit acts as any other unit and, if they reach 0 supply, they start losing HP.
Resupplying units can resupply normaly like other units, but they CANNOT resupply each others (this could cause infinite supply).
If more than one resupply units are available to resupply a unit, the amount is taken from the resupply unit that currently has the most supply.
If a unit can resupply normaly without using a resupply unit, they ignore resupply units.
Etc.


You can see the complexity ramping up, it starts to have exceptions, and decisions need to be made. The idea of having resupply units is logical, but much more complex.

Would that complexity be good? Would depth be increased? I don't think so. You could easily use Workers as kind of resupply units and build forts as you advance. Allow forts to be build in ennemy territory.

If you add resupply units, they will have to constantly come back to a fort to resupply and would require a lot of micromanagement. Instead of managing your army's supply, you have to monitor your army AND your resupply units. If you give them too much supply, your whole supply system becomes useless. Heck, you could probably create an infinite train by cycling enough supply units, totally defeating your new system.

That's what you have to look for when talking about complexity: needless complexity.

Also, the more complexity a system is, the more abuse is possible. If it is functionless, it's even worst.

 

[Seven05]

Well I hate to interrupt the complex discussion above... ok, not really, I'm going to anyway.

First, great article, even if you only read the opening statment and keep that in mind; you first, then the players. I know it sounds mean but after modding too many games to remember since Doom in the early 90's I've seen more mod projects and teams fail because they tried too hard to please people other than themselves. In the end they ended up more frustrated than anything else- why work on somethign that is more frustrating than rewarding?

One issue with mods that I haven't seen discussed here is difficulty. My most recent completed mod was a 4-year multi-player only project. For the first year the mod was only seen & played by the development team and a select number of testers. At the time of the initial public beta we all felt that it was challenging but still fun. Unfortunately, once we let the "unwashed masses" onto the server we suddenly had to deal with people who put substantial effort into exploiting any weakness in the mod and people who were serious "power gamers" who were able to easily overcome what we had felt was a good challenge. Obviously the exploits needed to be addressed since it was a multi-player game so that's not an issue. Difficulty was interesting though, we had severl members of the development team that felt insulted or threatened by the players who found something to be easy when it wasn't supposed to be. Despite the fact that this was really less than 10% of our player base it sparked off a number of changes that ended up pushing the overall difficulty of the game too high for casual players. This made development less fun for the team as a whole and the mod less fun for the majority of the players. It went from making a fun game to being a contest between the "good" players and the development team. Thankfully CIV has difficulty levels and isn't a multi-player only game but it's still possible to fall into the trap of trying to "beat" players because you feel insulted that they find your mod too easy.

Ok, back to the complex part

 

[Jeckel]

I personally thing Civ4's difficulty levels are kinda poorly done. The levels below Prince are way to easy and Deity is just impractically hard. Civ3 had some of the same problems and I tend to think it stems from the fact that difficulty is just modifiers on costs, coruption, ect.

Diffuculties can have modifiers on production and whatnot, but the main difference between them should be the amount of inteligence the AI uses to play the game. It shouldn't be how fast the ai builds a warrior, it should be whether the AI builds a warrior or waits just trys to squeze out another worker before guarding it self.

 

[Zurai]

Diffuculties can have modifiers on production and whatnot, but the main difference between them should be the amount of inteligence the AI uses to play the game. It shouldn't be how fast the ai builds a warrior, it should be whether the AI builds a warrior or waits just trys to squeze out another worker before guarding it self.

The reason why this is never done is that it is a massive waste of time and resources (== money). Why develop a complex, intelligent AI only to restrict it to only the highest difficulty settings that most people will never even play on?

 

[Kael]

The reason why this is never done is that it is a massive waste of time and resources (== money). Why develop a complex, intelligent AI only to restrict it to only the highest difficulty settings that most people will never even play on?

There may be some middle ground here where the AI acts appropriatly (and that part could use a lot of tuning) and adds in some randomness. The amount of randomness could vary depending on the difficulty level so that lower difficulty opponents play more haphazardly and higher opponents are both more predictable and more efficient.

 

[griffin71]

Why develop a complex, intelligent AI only to restrict it to only the highest difficulty settings that most people will never even play on?

There may be some middle ground here where the AI acts appropriatly (and that part could use a lot of tuning) and adds in some randomness. The amount of randomness could vary depending on the difficulty level so that lower difficulty opponents play more haphazardly and higher opponents are both more predictable and more efficient.

But from the fun-of-playing point of view, I think it's one of the attractions of a game to find out how the AI thinks. For that reason I'm strongly in favour of more intelligent AI over the option of changes in production advantage. One of the major upgrades in this area was that the AI in Civ3 no longer comes up with a single attacker every second turn, but that suddenly a small army is knocking on the door.
So intelligence of the AI also affects gameplay, and not necessarily is limited to the highest levels.

In addition, more intelligent AI can often be accomplished of a set of small rules added to the main rules followed. E.g. it may be wise to have an extra focus on culture to get a valuable resource within your territory.
For the same reason, more than normal defenders in an outpost of your lands that provides you with the special resource, may be a very good choice. Such intellgence is not too hard to implement.

 

[dh_epic]

Let's use complexity in the most value-neutral sense -- complexity is neither good nor bad. Some things are just more complex, and some things are just less complex. More stuff --> more complexity.

When do you have too much complexity? When do you have too little? It depends. There are tradeoffs.

If you make something too complex, you risk making it difficult to learn. And if it's hard to learn, you shrink your audience. Drastically. / But if you make something too simple, then the player feels limited. They run out of things to do. They get bored. They move on. Simplicity can drastically lower the 'replayability' of a game.

If you make something too complex, you can make it extremely hard to balance. That's the toughest part about creating a strategy game. And if something isn't balanced, you end up in Civ 3 territory where people quickly identify 'the best strategy' and follow that every time. This ultimately removes choice. / But obviously, if something is too simple, then you also remove choice. Imagine Civilization 4 with 12 technologies -- much fewer strategic choices. A much shorter game. Easier to get tired of.

Complexity has the added dimension of AI in Civilization 4. If you add all kinds of cool stuff, it's hard enough to learn for the player. Imagine the AI. Regardless of how much people in this community love their computers, their compuers are not HAL9000 (and maybe that's a good thing). They're dumb. Sometimes it's impossible to get them to do even the most trivial of tasks -- managing Zone of Control with any kind of efficiency is one of those things.

-----

I maintain, one of the most insightful comments I ever heard from Sid, or perhaps Soren, or another key developer was something to this effect:

The Civilization series is a bunch of simple systems that add up to a complex and deep game.

Think about it.

- Citizens work tiles, which generate food, commerce, and production.
- At a certain threshold, food grows your population.
- At a certain threshold, production completes a building or unit.
- Commerce can be converted to research, which, at a certain threshold, gets you a new technology.

These are simple mechanisms that all add up. And there's a certain unity. Civilization is about tiles and icons, and it's about "filling boxes up" until they hit a threshold. This makes the game easier to learn.

This suggests, to me, that the best way to do 'complexity' isn't to think of 'what would be the most detailed, super realistic economic model' but to think of several smaller mechanisms that add up to create the broad sense of economic realism.

 

[DaviddesJ]

For that reason I'm strongly in favour of more intelligent AI over the option of changes in production advantage.

Everyone wants a more intelligent AI---that isn't the question. The question is whether the AI should be crippled (so it plays worse than its best) at low difficulty levels. Since the AI isn't so good even at its best, it seems to me a bad idea to make it play worse than its best.

Such intellgence is not too hard to implement.

It always seems that the people who say it's easy to make better AIs have never tried.