Recommend
5 
 Thumb up
 Hide
8 Posts

ZhanGuo» Forums » Variants

Subject: Automa Solo Variant rss

Your Tags: Add tags
Popular Tags: [View All]
Joel Oakley
United States
Brandon
Mississippi
flag msg tools
badge
Avatar
mbmbmbmbmb
I developed rules for an automa player. You can find them here:

Automa Solo Variant Rules

I appreciate any feedback you can offer to improve the rules (either in clarity or in the actual rules of play). Enjoy!

Original Design Notes:

My overall approach to designing the automa was based on two ideas:

(1) I did not want to use any components not included in the standard base game of ZhanGuo.

(2) I wanted the automa to emulate some of the interaction you would have when playing against a human opponent while making automa turns relatively simple to execute.


As far as point (1) was concerned, it was obvious to me that a large deck of cards with both color (ivory, orange, brown) and number (1 through 120) should provide ample information that could be randomized and make decisions for the automa throughout the game without requiring any additional components. It was simply a matter of deciding exactly how to utilize all the information on the cards.

Regarding point (2), I wanted the automa to interact with you both by playing cards to the Court Area and by tucking cards under his player board to gain unification markers (competition for end of round bonuses). Additionally, I really wanted some reasonably challenging competition for the palace locations, governor majorities, and wall scoring tiles. More specifically, since the Tasks play such an important part of the game in multiplayer games, I wanted the automa to pursue them in a somewhat human manor, both in terms of the pacing and in the specific placements of governors, walls, and palaces.

To address the pacing issue, I thought about how I typically play out my games. I usually tuck at least 3 cards (sometimes more) in the first round of play, with a gradual decrease of cards tucked in each successive round. To have the automa player emulate this style of play (building on the board slowly at first and more rapidly toward the end of the game), I made use of the ones digit on the cards in comparison to the current round number. As it is currently, the automa will typically tuck about 4 or 5 cards in the first round and around 2 to 3 cards in the final round -- this is probably more than a typical human player will do, but it is balanced out by the way that the automa takes court actions (each court action for the automa will place a structure on the board).

For further information about point (1), my main idea was to use the ones digit on cards to decide whether the automa was going to do a court action or a unification (card tuck) action. To simplify the game play in accordance with point (2), I decided that the automa would not have to bother with resources (officials or workers) or with unrest. Thus, it was really only necessary for the automa court actions to be those of installing governors and building walls or palaces. Since there are 3 colors of cards and 3 court actions that the automa should take, the solution to maintaining point (1) was quite obvious -- each color corresponds to a specific court action (ivory = governors, orange = walls, and brown = palaces).

Through some play testing of various approaches to the automa deck, I ended up settling on two decks. One larger "Decision deck" would determine whether the automa would take unification action or a court action (as well as what type of court action). The other small "Action deck" would be the standard 6 cards used by human players and would be used either to take the unification action or court action; additionally, it would serve as a way of deciding the location in which to perform the action. The main advantages of using two decks in this way is the following: it allows one to maintain an even mixture of cards used to take actions while letting the automa have preferences for certain court actions (in my rules, I used the ratio of governors : walls : palaces in the components as a guide for the card ratios -- this could still be tweaked through more play testing).

I am sure I could go into much more detail on the various methods I tried, but in the end, this two deck approach seemed to be the most natural and the most robust. There is probably still some room for tweaking things such as the exact card ratio for the Decision deck as well as the difficulty of the automa (some different difficulty levels already exist in the current rules).

Design Update Notes:

After more extensive play testing of the original rule set, there was a major overhaul to the automa system. The following is a justification of these changes.

The primary issue I had with the old system was the variance in the actual challenge presented by the automa. This seemed to boil down to 2 main factors:

(a) Although it was originally designed to build around 12 total structures on the standard difficulty, the exact amount that it would build was highly dependent upon the makeup and order of the Decision deck. It could vary quite a bit from play to play; I definitely saw 15+ structures in some plays and 10 or fewer in other plays at the same difficulty level.

(b) The automa originally randomly selected between the two task sections. Sometimes it would end up rushing only one section, and other times it would build in opposite sections in such a way that it did not really interfere with the human player's plans.

To address problem (b) first, I decided to try a two-color automa player that would focus one color on the upper section of tasks and the other color on the lower section of tasks. This only added a minor amount of automa management and was a marked improvement over the old system.

However, problem (a) was still an issue with my trials of a two-color automa system that relied entirely on the Decision deck and Action deck to decide which automa color would build (some decision cards would cause one color or the other to build, and some would cause both to build). In some trials, one automa color would outscore the other automa color by around 100 VP. It was clear that the Decision deck was too unreliable as a source for deciding which automa color would build.

To address the problem (a) and the new wrinkle created by the two-color automa, I came up with a cube pool system that determined the exact amount (with very little variance in most plays) of total structures that each automa color would build. For the now standard difficulty of the game, each automa color will build 13 structures (and rarely anything other than 13 exactly). Moreover, as mentioned above, each of the two colors will focus on one of the two task sections as it builds. The result is a much more consistently challenging automa player and one that can have its difficulty adjusted with more precision.

Yet Another Update:

An update is now available, the main change addressing a potential concern that the automa focuses too heavily on one type of structure (say a lot of governors for example). Modify the setup and use of the Decision deck in the following manner:

1) To build the Decision deck, shuffle together 9 ivory, 6 orange, and 7 brown cards.

2) During the Play Cards Phase, you only reveal a Decision card when Qin is supposed to build (i.e., if a red, grey, or white official is drawn from the Decision cube pool).

There are a couple of other very minor changes, but the Decision deck modification is the most important aspect of the update.
4 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Joel Oakley
United States
Brandon
Mississippi
flag msg tools
badge
Avatar
mbmbmbmbmb
After some more play testing, there is a significant overhaul in the works. It makes the game feel more like a 3-player game (hooray!) and should reduce the variance in the strength of the automa player.

The main new idea is that the automa will effectively play as two players (but still only one player board and one "turn" between your turns). This means that one automa color focuses on the upper task section while the other focuses on the lower task section. I think I have all of the adjustments made, but I need to read through the rules again and test them a few more times before releasing the new rule set.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Scott Olson
United States
Chicago
Illinois
flag msg tools
Avatar
mbmbmbmbmb
Thanks for doing this, I'll try it out!
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Joel Oakley
United States
Brandon
Mississippi
flag msg tools
badge
Avatar
mbmbmbmbmb
MadStork wrote:
Thanks for doing this, I'll try it out!


I would love any feedback you have, but...

I would recommend waiting at least a couple of days. The new automa rule set that is coming very soon is a big improvement!

Edit: the file has been submitted -- hopefully it will be approved quickly and available in the next day or so.

There are two big modifications:

1) The automa now plays 2 colors to emulate a 3-player experience -- due to the way I have implemented this, it really does not add much in terms of time and management of the automa.

2) The decision to use a card for unification action (tucking) or court actions (building walls, palaces, governors) has been handed over to a cube pool system (cubes drawn from a bag). This provides a much more consistent experience than the old system. For example, with the old rules, the automa would build around 12 to 13 combined structures on the board, but it could easily vary from around 10 to upwards of 15. With the new system, each automa color will build 13 combined structures almost every game (in standard difficulty).

2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Dale Buonocore
United States
Warren
Pennsylvania
flag msg tools
mbmbmbmbmb
Kudos, Joel, and nicely done -- many thanks for doing this. When I eventually circle my way back to this great game, I'll definitely take the plunge and give this Automa a try...!
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Dale Buonocore
United States
Warren
Pennsylvania
flag msg tools
mbmbmbmbmb
I've just started learning how this Automa works, and I have a question: if you choose to pass on a Unification Reward which you could take, I'm guessing that Qin does not automatically claim that reward (thus giving up all of his Unification markers of that type for no benefit)?

By the way -- your updated tweaks look excellent.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Joel Oakley
United States
Brandon
Mississippi
flag msg tools
badge
Avatar
mbmbmbmbmb
Elad wrote:
I've just started learning how this Automa works, and I have a question: if you choose to pass on a Unification Reward which you could take, I'm guessing that Qin does not automatically claim that reward (thus giving up all of his Unification markers of that type for no benefit)?

By the way -- your updated tweaks look excellent.


Great question. I have been playing that he does give up his tokens when you choose to forgo the reward, which is what I intended (i.e., Qin always takes the reward slot if he has the most or if you pass on the reward), but the rules are admittedly ambiguous about this point at best.

However, since having him keep his tokens when you forgo a reward would lead to a more challenging game (tighter competition for future rounds as well as potential extra points for Qin from card tucking without sufficient markers available), I think future versions of the rules will have something akin to the following clarification:

Quote:
Qin only claims the reward slot and gives up his unification markers of a certain type if he actually has the most markers (or wins the tie breaker for the most) of the corresponding type. If you have the most unification markers of a certain type (or you win the tie breaker for the most) and you do not take the reward, either because you choose not to take it or because you cannot take the full reward, then Qin neither claims the reward slot nor gives up his markers of that type. In such cases, neither you nor Qin claims the reward slot.


That will mean another adjustment to the two human players rules modifications. In this version of the game, Qin would take the reward slot whenever possible unless he is in 3rd place in the unification marker count. Even if both human players do not take the reward, Qin does not take the reward slot if he is in last place in the unification marker count.



1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Dale Buonocore
United States
Warren
Pennsylvania
flag msg tools
mbmbmbmbmb
Many thanks for your response.

Just finished my first play through. Much to my delight, things gradually fell nicely into place, and flowed more and more smoothly as I got a handle on the Automa's mechanics and decision-making processes -- and by the end Qin was showing himself to be both a challenging opponent ("right" Qin racked up a lot of points from walls at the end, coming back from way behind to thump me!) and a blast to play against. I'm now looking forward to incorporating those updated tweaks and taking on Qin again tomorrow -- especially now that I more fully understand what I'm doing and should thus be able to give him a better match.

Well done. This is a worthy Automa indeed.
1 
 Thumb up
1.00
 tip
 Hide
  • [+] Dice rolls
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.