On BoardGameNews Kris Hall posted an article entitled What would Knizia do? which asides from discussing Nefertiti, also suggested Reiner Knizia as a paragon of game-design sensibility to aspire towards. I took mild exception, describing my goals for game decisions a little differently to those perceived for Knizia’s designs (see below). This resulted in the following comment sequence in the ensuring thread. I’ve concatenated the three comments below as they parse reasonably as a single argument:

My most common internal question during design is: How can I make this decision point more nuanced, more subtle and less obvious? Obvious decisions are non-decisions. Early decisions should only make later decisions more difficult and less obvious. Ideally every decision in the game should be a challenge, a challenge both to determine that the decision is present in the first place as well as to decide on a good answer for that decision.

I’m not interested in making it less obvious that a decision is to be made. The game says to build something, you build something. That’s fine. I am interested in players not being sure what their decisions really mean for long-term implications. I’m interested in making the decisions have many facets, have implications in multiple areas over multiple time scales and for those implications to be soft-edged, un-obvious and un-intuitive, for many of the things decided by implication to be unclear in prediction but obvious in retrospect.

Yeah, I know the classic pattern is “have to do N things but can only pick one”. That’s not my favourite pattern. I’m not so fond of the “I have to do XXX but I don’t really want to!” pattern (eg Ra). I much more like decisions which all seem variously reasonable, where it seems reasonable to do any one of them and to not do the others, but they are in fact all importantly different in many different ways.

Yes, there is a danger is that decisions are so indistinguishable that they are insignificant. So far I’ve dealt with this in terms of commitment, decisions committing players to directions they can’t fully comprehend at the time.

Of course as the game comes to a close the effects of decisions become more obvious, intuitive or calculable as their lifespan and range becomes obvious. However earlier in the game the idea is for the decisions to largely be attempts to assess and influence the similarly cloudy decisions of the other players. each player committing to a suggestive posture in attempt to influence the other players to commit to resonant postures, thereby communally setting a larger direction for the game as a whole.

That last paragraph is poor. I don’t yet know how to clearly articulate the concept I’m attempting but it has to do with the goal of games not being mechanical exercises, not discretely calculable sequence models, but densely flavoured organic melanges, gestalts that more emerge from among the player’s intersections than being a discrete logic system that the players externally engage as manipulating clients.

I’ve been concerned about transactional density in Space Race. I’d like to have players (near) constantly having to pay each other for little actions and the sum of the current game state. Pay for this, pay for that, etc with all debts/payments recurring for every turn of the game after their inception.

This is a problem. There could easily be 20+ transactions per player in various directions. Of course many of those would more or less cancel. PlayerA plays PlayerB $17 and PlayerB pays PlayerA $14, for a net effect of PlayerA paying PlayerB $3. Of course the loop could be larger than that with all the other players involved in various degrees as the money sloshes about.

The obvious choice is simply polish all that detail out; figure out a way to remove all the accounting detail and yet let the game continue smoothly and with rich decisions. However, I’d like to not do that. I’d like to keep the net effect of all those hordes of micro-transactions without having to have the game support the overhead of hordes of micro-transactions. I intend the game to be severely cash strapped, so a few dollars or more draining out in one direction, turn after turn, is a big deal.

My first thought was that each player would have an income track containing markers for each other player. A player’s marker on a track would represent an obligation to play the track-owner that sum of money at the end of each round. As debt events occurred in the game players would synchronously move their markers up and down on each other’s tracks to represent the new balance of payments. However even this is too fiddly. A five player game would have 20 tracks (4 per player), each containing 4 debt markers (one for each other player), with two markers being adjusted for each debt-causing event. Further all this complexity doesn’t clearly represent or collapse debt cycles (eg PlayerA owes $5 to PlayerB who owes $5 to Player C who owes $5 to PlayerA for a net sum of $0). It also doesn’t clearly represent what the net sum of a player’s debt posture is. Each player would need to add the sum of their debts against the sum of their incomes. Fie!

The current thought is to simply have an income track for each player, centred on zero, say, running from -20 to +20. As debt events occur in the game players would move their income markers up and down as appropriate. Thus if PlayerA incurred a $5 debt to PlayerB, then PlayerA would move their income marker down by $5 and PlayerB would move their income marker up by $5. In a 5 player game this would result in 5 tracks, each with a single marker. All end-of-round payments would be to and from the bank. The net effect would be the same, money in sum moving identically to the more detailed system, but with the bank’s implied capital pool acting as a sump to simplify net transaction resolution.

As thoughts from the shower go, this one seems a keeper.

Alan Paull’s otherwise excellent game, Confucius. can seem overly chaotic as the Emperor’s Reward cards start being played late in the game. Ben Keightly and I have discussed this on #bgdf_chat:

[2008-10-01/11:30] <clearclaw> I am building a bureaucratic force of extraordinary magnitude – http://www.boardgamegeek.com/article/2692463

[2008-10-01/11:31] <cocadieta> Neat, that was the game I played last night.

[2008-10-01/11:32] <cocadieta> For my last action I got to choose whether J R or Dave won.

[2008-10-01/11:33] <clearclaw> Hehn.

[2008-10-01/11:33] <clearclaw> I’d love to talk more but have to run. Back in ~45 probably.

[2008-10-01/11:39] <sedjtroll> So does Confucius often come down to such a kingmaker decision?

[2008-10-01/11:39] <cocadieta> It’s easy for it to.

[2008-10-01/11:40] <cocadieta> This is the first time I’ve seen it quite so clearly, but there was nothing terribly special about this game (other than three new players).

[2008-10-01/11:40] <sedjtroll> Oh. lame

[2008-10-01/11:46] <clearclaw> No, in this case that’s a positive quality

[2008-10-01/11:46] <clearclaw> The entire game is about creating tied situations.

[2008-10-01/11:46] <clearclaw> A kingmaker position is merely an instance of a tied situation.

[2008-10-01/11:47] <cocadieta> That’s how I looked at it. Two players won. That I had to ‘pick’ between them isn’t really important.

[2008-10-01/11:47] <clearclaw> Quite.

[2008-10-01/11:48] <cocadieta> The money isn’t too lucky but the bonus cards are.

[2008-10-01/11:48] <clearclaw> If there had been an incentive created by one of the players for you to go one way or ther other it would have been more interesting. This was just the degenerate case.

[2008-10-01/11:48] <clearclaw> Aye, the Emperor’s reward cards are swingy

[2008-10-01/11:48] <clearclaw> What complaints I have centre there.

[2008-10-01/11:49] <cocadieta> In last night’s game, I had a 30% chance of pulling a card worth 8 points to me (!!!)

[2008-10-01/11:49] <clearclaw> They’re hard to remove cleanly

[2008-10-01/11:49] <clearclaw> Wowzers.

[2008-10-01/11:49] <clearclaw> I’ve not looked hard at the distribution.

[2008-10-01/11:50] <cocadieta> The situation was: Hoju was about to resolve and I saw my position there go from a very safe second to an impossible-to-fix third.

[2008-10-01/11:51] <cocadieta> I ran my boats out to Africa and picked up a bonus card. If it was the Hoju bribery card or one of the two wilds, I could have knocked someone out and gone back to second. The VPs in the region were 8/8; I had two markers and a gift invested in the region.

[2008-10-01/11:51] <cocadieta> Brutal, but then again, them’s the breaks.

[2008-10-01/11:52] <clearclaw> Yow. Those forced minister cards are (overly?) brutal.

[2008-10-01/11:52] <clearclaw> They haven’t determined one of our games yet but they’ve come close.

[2008-10-01/11:52] <clearclaw> They make the navies far more viable than they appear at first glance,.

[2008-10-01/11:53] <cocadieta> I’m happy to accept them as swingy, and also I realize I have to revise my line that the bonus cards are worth about a point or two each.

[2008-10-01/11:53] <clearclaw> I think I’d prefer the game without the Emperor’s Reward cards, I see them as something to be polished out, but they are tightly married to the rest of the system.

[2008-10-01/11:58] <cocadieta> I like them though I know why you aren’t crazy about them. I wonder what it would do to the game to keep them face up and let players pick them.

[2008-10-01/11:58] <clearclaw> Or perhaps a narrow draught pool. THAT would be interesting!

[2008-10-01/11:58] <clearclaw> (and face up when drawn)

[2008-10-01/11:59] <clearclaw> (unless drawn blind?)

[2008-10-01/11:59] <cocadieta> There are always two cards available, that kind of thing?

[2008-10-01/11:59] <clearclaw> Yes.

[2008-10-01/12:00] <clearclaw> Aside: In our last game we had a player do 15 fleets at once.

[2008-10-01/12:00] <cocadieta> I like the sound of that. I don’t mind the chaos and the huge swinginess of the cards–I think that’s perfectly in line with the insane swings that e.g. the students can have on the game–but the blind draws and secret holdings are maybe a little much.

[2008-10-01/12:00] <cocadieta> That’s great!

[2008-10-01/12:01] <clearclaw> Aye. I do like the draught pool idea.

[2008-10-01/12:01] <clearclaw> I’ll try and write that up later as a proposal.

[2008-10-01/12:01] * clearclaw will probably dig lightly at Faidutti.

[2008-10-01/12:01] <cocadieta> Ha, ha

[2008-10-01/12:02] <clearclaw> Undoubtedly he likes the chaos. His style.

[2008-10-01/12:02] <cocadieta> Yeah, absolutely.

[2008-10-01/12:02] <clearclaw> This variant will lead directly away from that quality.

[2008-10-01/12:02] <clearclaw> I expect it will be accused of attempting to turn the game into something that it isn’t.

[2008-10-01/12:25] <cocadieta> J C, will your ego be bruised if I type up that variant as part of a reply to J R’s report?

[2008-10-01/12:26] <sedjtroll> I think JC’s ego is impervious

[2008-10-01/12:51] <clearclaw> I am currently in the process of putting that variant into an OtherWise post which will be cross-posted to BGG as a variant proposal

[2008-10-01/12:53] <clearclaw> Outside of that yes, my ego will not be bruised.

[2008-10-01/12:53] <cocadieta> I mentioned the variant in J R’s report. If you’d like I’d be happy to credit it to you as well, it just didn’t seem worth it.

[2008-10-01/12:53] <clearclaw> I’m not overly worried.

[2008-10-01/12:53] <clearclaw> (or underly)

Ben’s post to BoardGameGeek:

I really love this game, and I think navigating its insane chaos is a very unusual challenge. I agree that the luck of the money draw really just means players need to be prepared to spend only a little or a lot on any given turn, and there are plenty of cheap as well as expensive things to buy. Poor planning will make anyone susceptible to bad card draws, but a versatile board position and gift pool should ensure that you always have something to do.

After last night’s game, my new feeling is that the bonus cards are a little too lucky. Chaos is the order of the day here, but it’s basically ‘fair’ chaos: I have, at every moment, a very good overview of everyone’s position and incentives (to borrow a word–hi, J C). That is, except for bonus cards, which are hugely variable in worth to every player, randomly drawn, and secretly held. I love the swinginess of the cards, I love the chaos they create, but I don’t appreciate their random draw.

Proposed variant: At the beginning of the game, shuffle the cards and keep them face down. Flip over the top two (three?) cards; when a player wins a bonus card, he selects one, keeps it face up in front of him, and flips a new card to replace it. I feel like this would preserve the crazy chaos of the game but continue to allow each player to have a ‘complete’ picture, letting him make more informed decisions. I also don’t think it adds much, if any, complexity to the game.

I followed that with the following post:

The more I think about this proposal the more I like it. There are a few questions:

• Does the draught pool refill immediately when a card is taken (this is significant for the cases in which players take more than one card with a single action via the navies)

• May a player draw blind?

• Size of the draught pool

I hadn’t thought these through when I proposed the variant on #bgdf_chat. My general view is that the Emperor’s reward cards as currently implemented are unnecessary complication to be polished out by development. I’ve spent a few hours looking at doing that but the cards are tied deeply into the rest of the system and can’t be removed easily without grossly affecting other key game relationships.

The more cards are in the draught pool the higher the probability that one or more of the cards will be specifically useful to any given player. Additionally the exposure of the cards in the draught pool increases their value for the players as they know what they are getting/competing for. There’s also an argument that the card values are decreased because they are also revealed to the other players, however this value change seems small.

Fleets are the only way to acquire multiple cards with a single action. We’ve had players send 15 fleets out in a single action in more than one of our games, thereby taking three cards at once. I strenuously doubt that any player will ever send out 20 fleets with a single action and thus take 4 cards at once. I’m willing to discount that contingency or cover it with a special rule that the last card is blind. 15 fleets is rare enough as to provide a reasonable outer bound.

Allowing the draught pool to immediately fill as cards are drawn allows such a fleet player to benefit from luck of the draw without any foresight by other players. If not a step backward, this is not an improvement.

Blind draws effectively recreate the (overly) chaotic system which we’re trying to address. Blind draws would be better if the drawn card were then revealed by the player, but that’s a band aide atop the problem, not an actual improvement. Of course the player drawing immediately after another player’s draw may also profit from the blind flip as the draught pool is restocked. This could be prevented by also revealing the next N cards which will be put into the draught pool as other cards are taken. I like this idea but am unsure if the complexity is justified.

Initial conclusions:

• Draught pool is three cards wide

• Draught pool is not refilled until after the player’s turn

• Emperor’s Reward cards may only be taken from the face up draught pool

• Players must keep their Emperor’s Reward cards revealed to all players

Possible extension:

• After player turns the draught pool is refilled in order from an additional set of face-up cards

Having thought about it a bit more since that posting, I like the conclusions and extension above. I have yet to play with this variant. I’ll try to get it on the table today.

I taught Wabash Cannonball to three new players this evening. It was an interesting game that ended on the 8th dividend, with me being high cash and high income for the last 4 dividends! In short the other players manipulated turn order to ensure that I never got a chance at a capitalisation action in order to sell a share to end the game. Additionally as I only owned shares of two companies (2 PRR and 2 C&O, both already in Chicago), I had no way to drive the game closed via track or Development cubes. It was most annoying. End-game scores were $187 (me), $166, $162 and $159. However that was not the interesting part.

At the end of the game the $159 player said that he thought one rule should be added to the game:

If all the players end the game with more than some amount, say $150, then they should all be declared winners. We’re all filthy rich and thus gloriously happy; we should all be winners!

Without commenting on the equation of wealth and happiness, this really isn’t such a bad idea. Really. What it actually provides is a strong incentive for the players to drive the game to a close rapidly so that there is the opportunity for a single unambiguous winner. Conversely it also gives the behind players a strong incentive to drive the game long to push the cash holdings to break the $150 ceiling and thus also win. It puts the collusive base of the temporary emergent alliances on a new footing as the end-game approaches, making what are otherwise fairly straight-forward e(V) calculations not so necessarily calculable any more. Now the game is potentially one-against-all as the winning player attempts to drive the game closed before the other players push their cash holdings over the threshold.

I find this clever: a simple, trivial, change which allows players to collude to their own self-interest to invert the scoring system of the end-game to their advantage.

The variant is unnecessary for Wabash Cannonball as this odd corner case is so unusual as to not require a special rule. If I’d owned shares of more than two companies (I’d have ended the game on track cubes much earlier). However as a base mechanism for other games it has me thinking. Provide an inversion cap on scoring that suddenly declares that all the players win and thus create a massive selfish competition in the end-game of one against everyone for the solo-win. Oh, what spice! Neat!

Subject: B: 'Ohana Proa
From: J C Lawrence 
Date: Fri, 31 Oct 2008 17:39:25 -0700
To: autorenwettbewerb@hippodice.de

J C Lawrence would like to submit 'Ohana Proa to the Hippodice
competition.  He is the sole designer and his email address is
XXXX@kanga.nu. 'Ohana Proa is designed for 3-5 players, age 10+ years
and lasts about 150 minutes.

Please find attached a description, the rules and a player aid.

--
J C Lawrence                        They said, "You have a blue guitar,
-------------------------(*)        You do not play things as they are."
XXXX@kanga.nu                       The man replied, "Things as they are
http://www.kanga.nu/~claw/          Are changed upon the blue guitar."

• Description
• Rules
• Aid

I posted this as a comment/reply on BGN, but it seems worth preserving:

My most common internal question during design is: How can I make this decision point more nuanced, more subtle and less obvious? Obvious decisions are non-decisions. Early decisions should only make later decisions more difficult and less obvious. Ideally every decision in the game should be a challenge, a challenge both to determine that the decision is present in the first place as well as to decide on a good answer for that decision.

Assuming that Capitalisation (or control of Capitalisation) remains the most attractive early action in the game (an argument I’m having difficulty supporting):

• (E7/D5/C3)
• P1-C3/3 P2/C3/3 P3-C3/2 P4-D1/1 P4-D1/2¹ P4-E2/4² (E6/D3/C0)
• P1-D1/4³ P2-D1/4³ P3-D1/4⁵ (E6/D0/C0]

Total actions performed: 9

Total actions per player: P1:2 P2:2, P3:3, P4:3

Now let’s assume that P1 recognises the action race and changes tempo:

• (E7/D5/C3)
• P1-C3/3 P2/C3/3 P3-C3/2 P4-D1/1 P4-D1/2⁶ P4-E2/4 (E6/D3/C0)
• P1-E2/5⁷ P2-E2/5 P3-E2/5 P4-E2/6 ⁸ (E1/D4/C0)
• P1-E2/7⁹ (E0/D4/C0)

Total actions performed: 11

Total actions per player: P1:3 P2:2, P3:2, P4:4

Interesting.

Let’s assume that P4 wins one of the early share auctions with his cash lead and thus has more he can develop:

• (E7/D5/C3)
• P1-C3/3 P2/C3/3 P3-C3/2 P4-D1/1 P4-D1/2 P4-D1/3 (E7/D2/C0)
• P1-E1/5¹⁰ P2-E2/5[^111] P3-E2/5¹² P4-E2/5 (E3/D2/C0)
• P1-D1/6¹³ P2-D1/6¹⁴ (E3/D0/C0)

Total actions performed: 12

Total actions per player: P1:3 P2:3, P3:2, P4:4

Now let’s assume that P1 is cash poor, diluted on both sides and that P2 in similarly incented towards a fast dividend:

• (E7/D5/C3)
• P1-C3/3 P2/C3/3 P3-C3/2 P4-D1/1 P4-D1/2 P4-D1/3 (E7/D2/C0)
• P1-D1/4 P2-D1/4 (E7/D0/C0]

Total actions performed: 8

Total actions per player: P1:2 P2:2, P3:1, P4:3

This is fascinating. The players have no choice but to focus on Expansion next turn as there’s just no opportunity for any more Development. What interesting rhythms!

Ooops. I just noticed that one of the side effects of nesting the game-project categories is that the RSS feed URLs for the categories have now changed to represent the nesting. Arghh. Sorry guys. If this is a real problem please comment below and I’ll stick in some Apache rewrite rules to handle it.

A more tactically interesting turn:

• (E7/D5/C3)
• P1-E2/2 P2-E2/2 P3-E2/2 P4-C3/3 (E4/D5/C2) (no change)
• P1-D1/3 P2-D1/3 P3-C3/5 (E4/D3/C1) (no change)
• P1-E2/5¹ P2-C3/6 P4-E2/5² (E2/D3/C0)
• P1-E2/7³ P3-E2/7⁴ (E0/D3/C0)

Total actions performed: 12

Total actions per player: P1:4 P2:3, P3:3, P4:2

Note that the round ended due to exhausted actions at the same time as it ended for total action points. This may be deceptively convenient, in which case the reins will need to be drawn in a bit. More interesting is the turn-penalty for being late in the turn order. How curious! Is this the coat-tail-riding forced-alliances of Preußische Ostbahn, or is it just bad to have cash? How very appealingly curious. It might be time to build a small AI to model a few basic player incentive assumptions and see how this pattern plays out.

Somehow I seem to keep returning to the action selection and turn order mechanism in Muck & Brass. The current Pampas Railroads and Wabash Cannonball is a little unstable when employed for the very cash and share-sensitive Muck & Brass. I’m not convinced it needs changing, but I’m also not convinced it doesn’t. An unpleasant kettle.

I’ve been looking at a variation on the previously discussed and dismissed model, but losing the sliding concept and using a fixed tie-breaker for collisions.

• There is a grid of action squares, one row per player. 7 columns long
• Players start with a marker in the first column of one the rows of the chart, ordered by increasing personal cash going down (this could change, see below)
• The player whose marker is furthest to the left and closest to the top (tie breaker) has the current turn
• There are three action tracks (Expand, Develop, Capitalise) but likely with a different distribution: 7/5/3(?)
• On their turn a player selects an available action, moves the cube on the track in the standard manner and does the action (not optional?)
• Upon completion of the action the player’s cube is moved forward on the action chart per the cost of the action (Expand: 2, Develop: 1, Capitalise: 3?)¹
• The next/same player (leftmost/top) now takes their turn, etc etc
• When two actions are exhausted or two players reach or pass the end column of the chart dividends are paid in the normal fashion and the action tracks reset

Possibly the turn order tie-breaker, after the first round, can instead of cash be the reverse order of total action points used in the last round; the implications aren’t obvious to me.

An example opening round using the format Player-ActionCost/ActionTotal followed by a summary of how many of each action are left is listed below, one line per re-ordering:

• (E7/D5/C3)
• P1-E2/2 P2-E2/2 P3-E2/2 P4-C3/3 (E4/D5/C2)
• P1-D1/3 P2-D1/3 P3-C3/5 (E4/D3/C1)
• P1-D1/4 P2-C3/6 P4-D1/5 (E4/D1/C0)
• P1-E2/6 P3-E2/7 P4-E2/7 (E1/D1/C0)

Total actions performed: 13

Typical total actions per round in Wabash Cannonball: 8

Typical total actions per round in Pampas Railroads: 9-10

Total actions per player: P1:4 P2:3, P3:3, P4:3

I’ve made little attempt to make action choices logical. This is just a thought model. Most noticeable is that the rounds is longer (more actions done in the round). This may be acceptable, albeit at a cost in game length. My surface sense is that the tactical choices in this ordering are interesting and rather tweaky.