Under the Hood

This document explains how the simulation engine part of the Hurricane Landfall game works. If you understand this document, you should be able to explain nearly all the questions players may ask about why something happened in the game, and be able to point out various details of game events relevant to whatever teaching points you're using the game to address.

The Big Caveat

Before we start, it's very important to note that this is a game, not a model. We're not trying to make accurate predictions about urban dynamics and natural hazard response; we're trying to teach players some ideas about how urban development and natural hazard interact.

A serious game has two parts: the serious part (whatever it is you want the players to learn) and the game part (interacting with an engaging and challenging -- a.k.a. 'fun' -- system of abstract rules and tokens). There are times when these two parts come into conflict, and you have to make tradeoffs. For example, in the early stages of game development, we considered that since future emergency managers were one of our target audiences, maybe one of the roles should be the emergency manager. The problem there is that realistically speaking, an emergency manager has very little political power. Either we would need to give that role a much bigger voice than it has in the real world, which would distort some of the teaching points, or we would have to give that role a whole lot fewer votes than all the other roles, which would make the game a whole lot of not-fun for whoever ends up in that role. So we decided that it was better not to have the emergency manager as a role and to bring in hazard as one of the metrics players might care about for their city.

There are some very sophisticated and realistic simulations of urban dynamics out there. They operate on vast amounts of data and are very complicated and difficult to understand. They're highly non-linear, which makes it hard to be sure that a given action is related to a given outcome. This game is does not use such a model. Instead, it just produces results that look right.

In a game context, persuasiveness is more important than realism. In fact, there are some areas where we intentionally steered away from realism. During the design process, we had some suggestions that we could use more realistic symbolism for the various buildings. Perhaps even use GIS to pull in real-world data or aerial photography of a real city to base the game on. We thought about it, and the problem is that then the game would look more realistic than it actually was. Players would see things like the locations of individual firehouses or highway exits and assume that they had an impact on the game, when such things wouldn't even exist in the underlying game representation. This is unpersuasive in the worst way: the game looks real, but it doesn't act real, so the player concludes that it has no bearing on reality and dismisses it entirely. By keeping the representation abstract, we keep the player's expectations matched to the game's reality, dealing with it on its own terms.

Point being, there are many ways in which this game does not reflect the reality of how cities function. Generally speaking, we did it that way because it made the game work better as a learning experience, whether it was to emphasize a teaching point, or to make the game more engaging, or simply because it made an otherwise intractable issue manageable. Keep that in mind when thinking about how the game works.

The Map

Let's start with the map itself, the environment of the virtual city where everything happens. The town of Disaster Landing is inspired by a real-world Gulf Coast town in terms of its geography, but we have done our best to genericize it and give it its own issues. There are some things that are intentionally a little vague (such as the distance to Big City, which way is north, and exactly how big the town is) because we want it to represent the problems of many Gulf Coast communities, not just one in particular. (We've been somewhat gratified to hear from various users that we were relatively successful in this goal).

The first thing to note about the map is the hexes. As far as the game engine is concerned, a hex is a single location. Everything within a hex is in the same spot, effectively. There are 19 positions within a hex where a landuse can be located, but they all have exactly the same elevation, hazard ratings, transit connectivity, and so on. The only way they differ is in where they are shown on the screen.

(This is, frankly, a design flaw, and if we ever get the time and money, we will go back and change the code so that one landuse equals one hex exactly, but this would be a fairly serious undertaking, and we haven't had the resources as yet. Our apologies for that.)

The hexes are, as mentioned, filled with different landuses (abbreviated LU). There are 9 of them: house, hotel, store, condo, industrial, openspace, road, water, and 'none'. The first four are associated with the four roles and represent different patterns of landuse within the city. The industrial LU only shows up on the northern island and represents the big industrial park up there. 'Road' only matters for the highway, other smaller roads are too small to count. Water should be self-explanatory. 'Open space' and 'none' both represent undeveloped land with nothing on it; the difference is that open space cannot be developed.

There are two other elements of the map that are significant: transit and hazard. Transit just indicates whether or not the highway runs through a hex. (The transit toggle controls display of the highway & ferry.)

The hazard layer is what shows when you click the 'hazard' toggle and it overlays all the hexes with various colors. This is the flooding hazard for each hex: red for high, orange for moderate, and yellow for low. (There is no green for very low -- it's a barrier island, and the highest point is probably all of twenty or thirty feat above sea level. If a category 5 hurricane were to hit the island, the whole thing would be inundated.) This is one of the places where it's important to note that everything within a hex has exactly the same properties.

And that's it for the map. The graphical display shows different types of terrain (beach, grass, swamp), but those are really to give the players an intuitive sense of how the island is structured. The only way in which they matter to the engine is that terrain reflects elevation which reflects flooding hazard.

Ecology and Hazard

Along with the growth rates for the four primary LU types, there are two measures of the city's state ('metrics') shown on the Info tab of the game UI: hazard and ecology.

Ecology represents the health of the local natural environment. It's calculated in a very straightforward way: it's simply the weighted average of all the non-water LU in the game. Undeveloped land (LU = "none") gets a weight of 1; open space gets a weight of 1.5. Everything else gets a weight of 0. A completely undeveloped map would have an ecology score of 1, and total build-out would give a score of 0. Now, the map starts with some development and some open space, so in practice it'll never reach those extremes. In addition, some proposals add a bonus to the ecology score. Ecology scores at the end of the game range from about 0.9 on the high end to about 0.4 in the most overdeveloped case.

The hazard score is the expectation value of the number of LU that would be destroyed in the event of a hurricane. See the section below on the hurricane for more details on hurricane damage. The scoring algorithm simply goes through all the hexes, counts the number of each LU type in the hex, figures out the odds are of each LU type being destroyed by wind and/or flooding in that location, and tallies up the totals. The raw score is then scaled so that it ranges from 0 to 1. (Actually, from about 0.2 in the best case to about 0.9 in the worst, because we want to preserve the player's sense that the world is unbounded, and things could always be worse or better.)

Growth

After each round of the game, the city grows. The rate of growth of each of the four urban LU types is controlled directly by the four growth rates shown on the Info tab. These growth rates all start out at 0.5 (50% growth each round) and are modified by various proposals.

The growth model is brain-dead simple and totally unrealistic, but it produces results that are persuasive. It's what is called an agent model, and this is how it works.

During the growth phase, the game engine goes through the map and examines every LU. For each of them, it pick a random number between 0 and 1. If the number is less than the corresponding growth rate, that LU spawns, creating a new LU of the same type (a 'growth agent'). The agent then wanders around the map. At each step, if the current hex has an unoccupied slot (an LU of type "none"), there's a chance it will settle there.

The rules for movement are fairly simple, and implemented by giving each hex a list of outgoing links to other cells. The agent moves by picking a random link from the list. If there are more links between two cells, the odds are increased that an agent will move in that direction. A hex links to each of its adjacent cells except for those completely filled with water, industrial, or open space. The number of links depends on the sewer provisioning status of the current cell and the target: standard number if they are the same, fewer links if moving from provisioned to non-provisioned, more if moving from non-provisioned to provisioned. Extra links are added if the target cell has a highway running through it. There are also a couple links across the channel, where the ferry is; more are added if a bridge has been built. If the cell has available empty slots, it adds links to itself; an agent that picks a self link will settle in the current hex.

The net result is that areas with one kind of LU will tend to get more of that LU in the nearby vicinity. Growth will happen preferentially along the highway, and within "city limits" -- the zone where the city provides water and sewer services, rather than the builder having to pay for them. Mostly buildings grow near other buildings, but sometimes they will show up further away, sometimes even off in the middle of nowhere. It's nothing like how it happens in the real world, but it generates results that look plausible.

Events & Proposals

So that's the baseline dynamics that the game engine works with. On top of that we layer the events of the game and the decisions of the players. Each round begins with an event, and then the players respond to those events by passing proposals. The events are static. Every game will have exactly the same three events in each round. However, the events generate problems, and the problems that occur depend, both directly and indirectly, on what proposals have been passed.

The rest of this document will discuss the dynamics of the events that begin each round and the proposals that can be passed in response. Note that after the first round, not all proposals will be available in each round. Some proposals are targeted toward addressing particular problems, and so will only show up if that problem occurs. Other proposals are generic, and will always show up (unless a similar proposal has already been passed in a preceding round).

Round 1: Windstorm

The first round begins with the "windstorm" event, which downs some powerlines and knocks over a bunch of trees. This event doesn't really have a problem associated with it; it's the first round of the game and is really designed to introduce the players to the mechanics of the game and to allow them to make changes that will have impacts later on.

Because all games start out the same, the same proposals will always show up in Round 1.

R1 Proposal: End Turn Now

This proposal is always available in every round, and has the same effect every time: if it is passed, the round immediately ends. Half of the remaining budget, both social (work) and economic (money), rounded down, is added to the budget for the next round.

One lesson that can be learned from this proposal is that just because you can do something doesn't mean that you should. Sometimes the best option is to reserve resources for the future, even if that means some present opportunities are lost.

R1 Proposal: Acquire Open Space

This proposal converts undeveloped land (LU "none") to open space in hexes along the beach and/or along the wetlands. Each empty LU in the selected zone has a 50% chance of converting to open space; if the "required" option is selected, it increases to 90%.

Transfer of development rights (the TDR option) separates the right to develop land from a particular parcel of land and allows landowners to transfer unused development density to another location. If this option is chosen, then during growth between rounds, links are added from the selected hexes to a downtown hex, increasing the density of growth in that region. That's really a side-effect; the primary effect of the TDR option is to trade monetary cost for work cost.

Acquiring open space results in a small (0.05) drop in the condo growth rate, to reflect the fact that many important open space areas are also prime locations for the development of high-end housing.

The open space proposal is generic, but does not appear in future rounds. Opportunities to acquire open space are often highly situational. Aside from the obvious ecological benefits, adding open space has a lot of indirect effects on hazard and growth because it affects where development can occur. This is a subtle but powerful proposal.

R1 Proposal: Hazard Education Program

This proposal increases the social budget in subsequent rounds by 2 points (3 if the "increase budget" option is chosen). It also decreases wind hazard by a small (0.05) amount, to reflect the fact that an informed populace can contribute to the reduction of airborne debris, and decreases the growth rate of houses by 0.05, to reflect the fact that some homebuyers may be scared off by well-publicized hazards.

This is an 'invest in the future' proposal. It has few effects now, but provides more flexibility in the future.

R1 Proposal: Highway Improvements

It is a well-documented but poorly-known fact that roads induce growth. In the short term, increasing road capacity decreases congestion, but it also allows drivers to make trips that would otherwise be too inconvenient, and results in a long-term increase in traffic ("induced demand") that often exceeds the increase in capacity. In addition, the decreased travel cost means that many areas are now no longer "too far away" from homes, workplaces, and commercial activities, resulting in an increase in development in the areas accessed by a new highway.

This proposal boosts the growth of condos (luxury homes owned by vacationers and other commuters) by 0.2 and all other LU by 0.15. If the bridge option is chosen, those boosts are increased to 0.3 for condos and 0.25 for the other LU. A bridge also significantly increases links across the channel during growth, and thus the amount of growth that will happen on the northern island.

Although the first-order short-term effect of road improvements is to increase the evacuability of the town, in the long run the second-order effect of increased growth can overwhelm that benefit and even make things worse. Thinking about the long term means realizing that, over time, small things add up.

R1 Proposal: Municipal Sewer Reform

This proposal has a number of different (and possibly quite indirect) effects depending on which options are chosen.

The base effect (no options and status quo provisioning) is to decrease water exposure everywhere on the map by 0.1 and to apply a bonus of 0.1 to the ecology score. These represent the beneficial effects of having a good drainage system.

Changes to the provisioning area indirectly affect growth patterns, as described in the section on growth patterns. Growth agents are reluctant to leave the provisioning area when wandering and are drawn into it when from nearby. Making the provisioning area smaller thus makes the city more compact, while increasing it will tend to generate sprawl. Liberal (more area) provisioning also imposes an upkeep cost of 2 money in future rounds.

If the "improve facilities" option is chosen, a sewage treatment plant is built, increasing growth rates for condos and houses by 0.1. This option also prevents certain problems from happening later.

If the "restrict septic tanks" option is chosen, hexes outside the provisioning area get a relocation link to the downtown area during growth (further encouraging a compact city). This option also prevents certain problems from happening later.

Like acquiring open space, this proposal is subtle but powerful because it affects where development happens. One of the major teaching points of the game is that where people are located is a major determinant of what happens in a disaster. If a hurricane hits an uninhabited island, that's not a disaster, it's just an event. Where the city will provice water and sewer service might not seem especially important, but because it has a major effect on where it's cheap to build new houses, it greatly affects how many people live in high-hazard zones.

R1 Proposal: Retrofit Buildings

Hazard is composed of two elements: exposure and vulnerability. All four LU types start out with a vulnerabilty of 0.5 both to wind and to water. This proposal decreases those vulnerabilities. (See the Round 2 event section for more details on how hazard works.)

The base form of the proposal decreases hazard vulnerability by 0.05 for all LU for the chosen hazard(s). If the "low income grants" option is chosen, hazard vulnerability is lowered by an additional 0.05. In addition, each of the "Target Houses/Hotels/Stores" options lowers vulnerability for that LU in particular by another 0.05. (Condos are not included as an option as they represent newer buildings that do not need retrofitting to bring them up to code.)

R1 Proposal: Revitalize Waterfront

This proposal just straightforwardly adjusts growth rates. In the base form, it increases hotel growth by 0.1 and store growth by 0.2; with the "public facilities" option, it increases hotels by 0.2 and stores by 0.1.

The "relax regulations" option has two side-effects beyond decreasing the cost of the proposal. First, it increases the growth rates of all four LU by 0.05. Second, it increases sprawl. If regulations have been relaxed, then during the growth phase the odds of an agent settling are decreased, resulting in more diffusion and mixing. The agents thus spread out further, making the city less compact. (This is the "dumb growth" condition.)

R1 Proposal: Tree Trimming Program

The tree trimming proposal is intended to be a simple and obvious solution to the problem presented in the scenario. The facilitator can walk the players through debating and passing this proposal to introduce them to gameplay without significantly affecting their available decisions for the rest of Round 1.

This proposal adjust the global wind hazard (the exposure probability) downward by 0.1, from 0.5 to 0.4. Upkeep costs decrease the monetary budget by 1 for all future rounds.

Round 1 Resolution

At the end of the round, the game shows a summary of the proposals that have been passed and any changes in metrics due to player actions. An intentionally vague amount of time then passes, during which growth happens according to the process described above. Growth always results in a decrease in the ecology score and an increase in the hazard score.

Round 2: Hurricane

The round 2 scenario is a category 3 hurricane striking the island. The hurricane always hits with the same intensity, but some of the consequences can be prevented depending on what choices the players made in the previous round.

The scenario specifies that all problems the players do not address will be handled by other authorities, though not as effectively. The players are very unlikely to have enough budget to address all the problems even tangentially, so the emphasis this round is on prioritizing the different problems and deciding what's most important.

R2 Problem: Destroyed Buildings

The hurricane always causes physical damage to the buildings in the town. Although in reality the damage would range continuously from none to complete destruction, in the game we regard them as either okay or destroyed. (Although, as indicated in the problem description, 'destroyed' also includes 'severely damaged'.)

Hurricane damage is not deterministic, and can vary between two games even if the players make exactly the same decisions. This can even lead to different outcomes in round 3 and at the end of the game.

The engine handles hurricane damage as follows: each LU on the map has a rating for wind exposure (constant across the map), water exposure (depends on the hex the LU is in), and wind vulnerability (depends on the LU type) and water vulnerability (also depends on LU type). The engine generates a random number between 0 and 1; if that number is less than the product of wind exposure and wind vulnerability, the LU is destroyed. Then it does the same for water exposure and vulnerability.

Destroyed buildings show up as black Xs on the map. The fraction of buildings that are severely damaged or destroyed actually much higher than would be realistic in a real hurricane. We chose to exaggerate the damage of the storm for emotional impact: increasing the number of buildings icons that suddenly change to big black Xs gives the players a feeling for the devastation of this kind of storm. With a more realistic number, it would seem like not such a big deal.

Destroyed buildings are treated as an average of all four LU types when the hazard score is being calculated. This could theoretically cause the hazard score to shift up or down slightly, but is unlikely to be noticeable during play.

R2 Problem: Flooding Downtown

The hurricane also always causes flooding in the downtown area. If the players passed the Revitalize Waterfront proposal, disruption of downtown businesses can be avoided. (Disruption of downtown businesses is primarily a warning sign of future vulnerability.)

R2 Problem: Road Damage

Damage to the roads at the southern end of the island happens unless the players passed the Highway Improvements proposal in round 1. This is a problem that will frequently seem more urgent to the players than it actually is; the road is impassible, but there is very little development in that region.

R2 Problem: Wetland Contamination

Wetland contamination can be avoided in two ways: either by the purchase of open space in the border zone surrounding the wetlands, or by restricting septic tanks as part of sewer reforms. This problem is another warning sign, indicating possible future problems where development can threaten the marsh ecology. If this problem occurs, it applies a -0.05 penalty to the ecology score.

Round 2 Proposals

Proposals in round 2 are added in response to various problems. The proposals Rebuild Downtown, Improve EOC, and Build Seawall are added in response to flooding downtown, which always happens even though some of the effects are avoided by the Waterfront proposal in round 1. The Highway Improvements proposal is added in response to the washed-out highway problem (i.e., only if it was not passed in round 1). Hazard Education Program is likewise added only if it was not already passed in round 1. The Ecological Monitor proposal is added in response to wetland contamination, if it occurs. Finally, Property Buyout and Building Code Update are added in response to the destruction of buildings, as is the Historic Preservation proposal unless Retrofit Buildings was passed in round 1.

R2 Proposal: Hazard Education Program

This proposal is exactly the same as the proposal of the same name from round 1. It is available only if it was not previously passed.

R2 Proposal: Highway Improvements

This proposal is almost exactly the same as the proposal of the same name from round 1. The only difference is that the cost of the Bridge option has been decreased from +4$ to +3$. This proposal is available only if it was not previously passed.

R2 Proposal: Build Seawall

This proposal builds a seawall along the northern edge of the downtown area, decreasing the water exposure hazard for the four northernmost hexes from moderate to low. This will typically cause the hazard score to drop slightly, and helps prevent the downtown area from flooding again, but has negative repercussions (hinted at in the proposal description) that will show up next round.

R2 Proposal: Building Code Update

Building codes are an example of a fixed way of doing things that can be very difficult to change without an external impetus. The hurricane opens a window of opportunity for making this kind of significant structural change. If "dumb growth" or sprawl condition has been created by previous player actions, this proposal cancels it. It also decreases growth rates by 0.1 for all LU.

(This proposal would decrease vulnerability of buildings built after it has been passed, but since there's no future hurricane in the game, that effect has not yet been implemented.)

R2 Proposal: Ecological Monitoring

This proposal adds a +0.1 bonus to the ecology score. It also prevents some future problems.

R2 Proposal: Historic Preservation

The primary effect of this proposal is to prevent problems related to LU imbalance in Round 3. The idea is not that the preservation program itself has this effect, but that it acts as a proxy: the type of town that would pass a historic preservation program is the type of town that would also pay attention to the composition of the community and work to maintain a healthy mix of landuses.

This proposal also immediately rebuilds 5% of the destroyed buildings, representing investments in preserving designated heritage buildings.

R2 Proposal: Improve EOC

Improving the EOC reduces hazard, dropping wind and water exposure by 0.05 each. (Arguably it should be vulnerability, not exposure, that is reduced, but they are basically interchangeable in this context, and it was much easier to reduce exposure, so we took a shortcut and did it that way.)

The BSA option is an adaptive response that creates dynamic ties between government and the private sector, giving hazard management officials greater flexibility and resources in responding to new and unexpected hazards. Thus, there are some future problems prevented by this option.

R2 Proposal: Property Buyout

This proposal takes the 20 highest-hazard destroyed LU and converts them into open space. In the case of hexes having the same hazard rating, LU in hexes designated as open space are preferred over those that aren't, and ties are otherwise broken randomly.

If the "Condemn" option is used, the future social budget is reduced by 1, to represent the effects of lingering resentments.

This proposal is an example of the 'tough decisions' sometimes needed to solve serious problems: the natural human instinct is to put everything back just the way it was before the disaster, but that does nothing to change the situation that cause the disaster in the first place. Solving the problem involves moving people out of harm's way, and that can be very costly.

R2 Proposal: Rebuild Downtown

When passed, this proposal immediately rebuilds any destroyed LU in the downtown area (the three hexes indicated by the proposal's graphic), converting them back to their previous LU. The mixed-use option forestalls future problems.

Round 2 Resolution

As with the end of round 1, at the end of round 2 the game shows a summary of the proposals passed and resulting changes in scoring metrics. Buildings destroyed by the hurricane are then rebuilt. 20% of the destroyed LU is too damaged to be rebuilt and is converted to empty LU. The rest is converted from "destroyed" back to whatever it was before the hurricane. Then we have another intentionally indeterminate amount of time pass during which growth happens in the same way as before.

Round 3: Torrential Rain

The round 3 scenario is based on an extremely heavy rainfall event. The rain causes localized flooding and 'triggers' various other problems that depend entirely on the current state of the town and the decisions the players made in previous rounds. Not all of these problems are directly related to the storm, but regarding it as an event that 'exposes' them in subsequent surveys is a useful conceit that allowed us to broaden the game's focus in this round and explore the different ways that players' decisions can have unintended consequences.

Because the scenario is dynamic, the number of problems that can occur will vary, though there will always be at least one. Most games will have only 1 problem, some will have 2, a few 3, and a very few will have 4 problems to contend with. (In testing, the proportions were roughly 58%, 31%, 10%, and 1%, respectively, but that's in random play; real players are unlikely to have the same distribution.) It is very difficult to deal with 3 or 4 problems effectively.

Since the number of problems change, the budget is constructed dynamically as well. It starts at 7 work, 7 money and is modified by the problems that occur, representing the availability of external resources to help deal with different problems. The available proposals reflect the problems that occur.

R3 Problem: Flooding Downtown

Flooding Downtown is a localized problem that affects the local economy because of losses and closures to stores and other businesses. This problem is relatively easy to avoid. The downtown area only floods if the players have passed none of the following proposals in a previous round: Build Seawall, Municipal Sewer Reform, Property Buyout, Rebuild Downtown, Retrofit Buildings for water hazard, or Revitalize Waterfront.

If this problem occurs, the following proposals are added: Improve Drainage, Financial Assistance, NFIP CRS, and Retrofit Buildings (round 3 version). The Improve EOC proposal is also added if it was not passed in round 2. Economic budget is increased by 1.

R3 Problem: Septic Watertable

The Septic Watertable problem happens when storm runoff from backed-up sewers washes into the wetlands, contaminating the watertable and causing a cryptosporidium outbreak. This has a negative impact on the tourist trade.

This problem only happens if there are at least 35 buildings in swampy wetland hexes in and near the bird sanctuary. Thus, it is indirectly prevented by proposals that prevent growth in those areas, such as Purchase Open Space with the wetland terrain focus. A combination of proposals that decrease growth (especially of houses and condos, which are most likely to expand into the wetlands area) or encourage growth within a restricted area can also indirectly prevent this problem, but are not guaranteed to do so because of the randomness inherent in growth.

The problem is prevented completely if either a sewage plant was built or septic tanks have been restricted (both options on the Municipal Sewer Reform proposal). If the Ecological Monitor proposal has been passed, the problem is not prevented but averted: contamination of the watertable still occurs, but it is detected before it becomes a significant problem. This demonstrates the adaptive and flexible nature of hazard mitigations like monitoring programs, which can solve problems other than those they were intended to solve.

If this problem occurs, the following proposals are added: Water Purification Plant, Boil Water Notices, Septic Enforcement, and Promote Tourism. Budgets are not affected.

R3 Problem: Black Swan Virus

The Black Swan Virus problem happens when a previously unseen bird-borne virus crops up in the bird sanctuary and surrounding wetlands. This has a negative effect on the tourist trade.

A "black swan" is a high-impact event that is also hard to predict. In large part, such events are high-impact because they are hard to predict. The 9/11 terrorist attacks are often regarded as an example: it's only because no-one had conceived that hijacked planes might be flown into buildings that it was possible for the hijackers to do so; the unexpectedness of the outcome was one of the factors that made it possible. Similarly, when a disease organism like West Nile Virus or Dengue fever establishes a presence in a new area where it has not previously occurred, its impact is increased by the fact that no-one is looking for it as a possible cause: doctors don't know how to treat it, patients don't know how to avoid it, public officials have no mitigation policies in place, and so on. The disease might be common in other parts of the world, but because people believe that it is a foreign problem, it is made worse until they realize that assumption is false.

The term "black swan" used to mean something impossible; by definition, swans were white. When actual black swans were discovered in Australia in the 17th century, the term came to mean something that was perceived to be impossible until it came to pass.

The "black swan virus" problem occurs only if the Septic Watertable problem does not occur and the ecology score is above 0.75. This reflects the fact that we generally don't think of flourishing ecosystems as a potential danger. It occurs reasonably infrequently, since it requires the previous passage of a couple different proposals that provide ecology bonuses to get an ecology score that high.

This problem is averted (but not prevented) by the Monitoring Program proposal or the Business Sharing Agreement option of the EOC proposal, both of which are adaptive measures that provide flexibility in responding to unexpected events.

If this problem occurs, the following proposals are added: Public Awareness Program, Mosquito Control, Breeding-Site Reduction, and Ecological Monitoring (round 3 version). Budgets are not affected.

R3 Problem: Beach Loss

Beach Loss happens if a seawall was built in Round 2. Erosion is a constant process along coasts, and although it may reach a steady state in many places, that should not be confused with the shoreline being fixed and unchanging. Changes in the contours of the underwater landscape, whether due to large storms or human intervention, lead to changes in ocean currents, which typically lead to changes in the position of the shoreline. The loss of beach is bad for tourism.

This problem can be averted by the purchase of open space along the shore in round 1. This doesn't prevent the erosion from occurring, but if there are no or few buildings near the waterline, the resulting buffer zone keeps the event from being a problem that needs to be addressed.

If this problem occurs, the following proposals are added: Shoreline Armoring, Beach Renourishment, Undercurrent Stabilizers, and Bird Watching Ads. Economic budget is increased by 2.

R3 Problem: Industrial Runoff

The Industrial Runoff problem happens if there are at least 16 buildings on the land north of the channel, reflecting the impacts of industry on adjacent land. Runoff would happen regardless, but if the land is neither developed nor protected, no-one is likely to notice or care. Because of the barrier to growth that the channel presents, this problem typically only happens if a bridge has been built or if there's a high rate of development focused on the downtown area. (Because of the randomness involved, it could happen even in a low-growth scenario with no bridge, but it is unlikely; likewise, building a bridge does not guarantee that this problem will occur.) The potential loss of the company that owns and operates the industrial site threatens the local economy, since so many locals work there.

If this problem occurs, the following proposals are added: Tax Breaks, Pollution Regulations, and Remediation Assistance. Budgets are not affected.

R3 Problem Class: Landuse Imbalance

Most cities have a mixed composition, with a variety of different landuses and economic activites that support and reinforce one another. If that mix gets too far out of balance, various feedback loops can kick in that will make the city gravitate toward a different composition. The real-world dynamics of this process are much more complicated than the game dynamics (which depend solely on the growth rate scores), but the following problems are illustrative of the way that a city's nature can change because its composition has changed.

To figure out if there will be a landuse imbalance problem, we first calculate growth ratios for each of the 4 player LU. This is simply the count of the number of LU of that type at the beginning of round 3 divided by the number of LU of that type at the beginning of the game. Then we calculate the imbalance ratio by dividing the biggest growth ratio by the smallest growth ratio. A landuse imbalance problem occurs if the imbalance ratio is greater than 1.215 (a number determined empirically to give good results in terms of gameplay). Which imbalance problem it is depends on which growth ratio is the biggest.

All of the landuse imbalance problems will be averted by passing either the Rebuild Downtown proposal with the Mixed Use option, or the Historic Preservation option. These proposals don't necessarily act directly to prevent the problem, but passing them is indicative of the kind of attention to a community's makeup and structure that will keep imbalances from getting too big. So they're proxies for an attitude and an approach that prevents the problem, but they also make sense as proposals in a game context.

In addition to the other proposals, Development Plan is added if any imbalance problem occurs. Budgets are not affected by an imbalance problem.

R3 Problem: Overcommercialization

This imbalance problem occurs if the runaway LU (the fastest-growing one) is "store". An overabundance of commercial development puts a town like this one on the path to becoming a tourist trap.

If this problem occurs, the following proposals are added: Restrict Permits, Relax Housing Regulations, and Raise Sales Tax.

R3 Problem: Weak Job Market

This imbalance problem occurs if the runaway LU is "house" or "condo", unless condo is imbalanced against house (i.e., condo has the maximum growth ratio and house has the minimum growth ratio). This imbalance puts the town on the path to becoming a bedroom community, where everyone commutes to the city to work and there's little to no local economy to speak of.

If this problem occurs, the following proposals are added: Raise Property Taxes, Reduce Licensing Fees, and Solicit Industry.

R3 Problem: Soaring Land Values

This imbalance problem occurs if the runaway LU is "hotel", or if it's "condo" and the slowest-growing LU is "house". When a town is very popular and there's a lot of expensive development (hotels and condos) pushing out local housing and businesses, it's on the path to becoming a high-priced resort town. This is sometimes referred to as "Aspenization", after the Colorado ski town.

If this problem occurs, the following proposals are added: Lower Housing Standards, Visitor Use Fees, and Zone Affordable Housing.

R3 Problem: Casino

The last problem is that everything is going well, and so the town has caught the attention of a casino developer, who wants to base a gambling cruise ship there. The degree to which this is a "problem" is something that many players will disagree on, but regardless of its subjective desirability, a casino will change the town. This problem is the default problem for Round 3, and occurs only if there are no other problems. (We needed a default since it wouldn't be a very interesting game if there were no problems in round 3.)

If this problem occurs, the Anti-Casino Campaign proposal is added.

Round 3 Proposals

The resolution of round 3 is complex, and problems can be solved partially or fully. Each proposal therefore has a mitigation rating for the different problems that it can affect. The players may need to pass multiple proposals to deal with a problem fully.

The Community Center and Research Laboratory proposals are always available in Round 3. If, after adding whatever proposals go with the problems for this round, the total number of available proposals is less than 6, the following proposals are added, as long as they have not been passed in a previous round: Historic Preservation, Hazard Education Program, Highway Improvements, and Tree Trimming Program. The extra proposals are added in that order, so if adding Historic Preservation brings the total up to 6, none of the others will be added.

Note: if a proposal is determined to be unaffordable before the round begins, it is removed from the list. This shouldn't happen normally, but it is theoretically possible depending on the budget effects of previously-passed proposals.

R3 Proposal: Anti-Casino Campaign

Prevents the casino from being built, solving the problem entirely. (2 points of mitigation for the casino problem.)

R3 Proposal: Beach Renourishment

A temporary solution; replaces sand without affecting the factors that caused it to erode in the first place. 1 point of mitigation for beach loss, because it does assist with keeping tourism going.

R3 Proposal: Bird Watching Ads

Doesn't address the underlying problem of erosion, but helps keep tourism going by pulling in birders to replace beach-goers. 1 point of mitigation for the beach loss problem. Does not stack with the Promote Tourism proposal for solving second-order tourism problems.

R3 Proposal: Boil Water Notices

A cheap and easy way of dealing with the cryptosporidium problem, but doesn't look good from a tourism perspective, so it's only worth 1 point of mitigation for the septic watertable problem.

R3 Proposal: Breeding Site Reduction

Breeding site reduction is normally an essential component of mosquito control programs, but because the town is adjacent to a protected wetland, the breeding sites that can be eliminated are only a small contributor to the problem. Does nothing by itself, but provides 1 point of mitigation for BSV if added to another proposal that addresses it.

R3 Proposal: Community Center

The community center basically trades off money for political goodwill, giving +1 work and -1 money to future budgets. Each of the options for the community center mitigates against one of the landuse imbalance problems, but only if combined with another proposal. Library, public space, and water park provide 1 extra point of mitigation against weak job market, soaring land values, and overcommercialization, respectively.

R3 Proposal: Development Plan

A development plan is a good general-purpose solution to structural imbalance problems. Provides 1 point of mitigation against each of the landuse imbalance problems.

R3 Proposal: Ecological Monitor

Although this proposal will avert the BSV problem entirely if passed in round 2, implementing it after the fact provides only 1 point of mitigation. "An ounce of prevention is worth a pound of cure." Otherwise identical to the proposal from round 2.

R3 Proposal: Financial Assistance

Helps with the current crisis, but does not address the underlying cause or protect against future occurrences. Provides 1 point of mitigation for the downtown flooding problem.

R3 Proposal: Hazard Education Program

This proposal is exactly the same as the first-round proposal of the same name.

R3 Proposal: Highway Improvements

This proposal is exactly the same as the first-round proposal of the same name.

R3 Proposal: Historic Preservation

Although this proposal will avert imbalance problems entirely if passed in round 2, implementing it after the fact has no effect on these problems. Otherwise identical to the proposal from round 2.

R3 Proposal: Improve Drainage

Solves the flooding issue by dealing with the underlying cause and preventing it from happening again. However, the second-order problem of the resulting economic disruption still needs to be addressed separately, so this proposal only provides 1 point of mitigation to the downtown flooding problem.

R3 Proposal: Improve EOC

This proposal is exactly the same as the round 2 proposal of the same name.

R3 Proposal: Research Laboratory

This proposal doesn't address any of the round 3 problems directly, but it broadens the local economy, making it more resilient, which provides 1 point of mitigation against the second-order economic problems that occur if the main problems are not fully solved. (See Round 3 Resolution.)

R3 Proposal: Lower Housing Standards

Fully solves the housing landuse imbalance problem by allowing the creation of low-quality, inexpensive housing. Provides 2 points of mitigation against soaring land-use values and immediately converts 5% of the condos and hotels on the map to houses. Decreases condo growth rates by 5% and increases housing growth by 5%. However, beause of the lag involved in returning to a more balanced distribution, if this is the only proposal addressing the landuse imbalance, it creates (or worsens) the second-order problem of weakened local economy. (See Round 3 Resolution.)

R3 Proposal: Mosquito Control

Partially solves the tourist problem created by BSV, providing 1 point of mitigation. Both strategies (fogging and bio-larvicide) have an ongoing upkeep cost of 1 money. Fogging imposes a -0.05 penalty to ecology.

R3 Proposal: NFIP CRS

Participating in the National Flood Insurance Program's Community Rating System lowers insurance costs for businesses, helping economic recovery, and provides an incentive for the community to become more resistant and resilient to flooding in the long term. Provides 2 points of mitigation against the downtown flooding problem and decreases hazard score by dropping the water exposure hazard by 0.1.

R3 Proposal: Pollution Regulations

Taxing polluters will help protect the ecology, giving an ecology bonus of 0.05, but exacerbates the economic problem caused by industrial runoff, adding -1 point of mitigation. This proposal often partially negates the effect of other proposals, reflecting the mixed messages sent by helping with the runoff problem on the one hand while imposing fines on the other.

R3 Proposal: Promote Tourism

Advertising can help undo the bad publicity caused by the cryptosporidium outbreak, but does nothing to solve the underlying problem. Provides 1 point of mitigation against the septic watertable problem. Does not stack with the Bird-Watching Ads proposal for solving second-order tourism problems.

R3 Proposal: Public Awareness Program

This is a cheap and easy way of dealing with the BSV problem, but it doesn't look good from a tourism perspective, so it's only worth 1 point of mitigation.

R3 Proposal: Raise Property Taxes

Slows housing and condo growth by -15%, adds +3 money and -3 work to future budgets (revenue increased at the cost of political goodwill), and provides 1 point of mitigation against the weak job market landuse imbalance problem.

R3 Proposal: Raise Sales Tax

Slows hotel and store growth by -10%, adds +2 money and -2 work to future budgets (revenue increased at the cost of political goodwill), and provides 1 point of mitigation against the overcommercialization landuse imbalance problem.

R3 Proposal: Reduce Licensing Fees

Solves the weak job market problem (2 points of mitigation) by providing economic incentives that draw new businesses. Upkeep costs impose a -1 money penalty to future budgets.

R3 Proposal: Relax Housing Regulations

Solves the overcommercialization landuse imbalance problem (2 points of mitigation) by making it easier to build residences. However, this also generates sprawl. Increases house and condo growth rates by 10% and sets the "dumb growth" condition (see the waterfront proposal from round 1).

R3 Proposal: Remediation Assistance

Each of the options (assisting with cleanup or improving the drainange system) provides 1 point of mitigation against the industrial runoff problem; choosing both options provides 2 points of mitigation. In the long-term, assisting with drainage improvement is better, since it will prevent the problem from recurring.

R3 Proposal: Restrict Permits

Restricting sales tax permits for future businesses decreases growth rates for hotels and stores by 10% each, imposes a -1 work penalty to future budgets, and provides 1 point of mitigation against the overcommercialization landuse mitigation. Restricting existing permits also provides 1 point of mitigation and a -1 work budget penalty, and weakens the local economy a little (+0.5 to the second-order local economy problem). Restriction of both types of permit simply combines these effects.

R3 Proposal: Retrofit Buildings

Same as the round 1 retrofit proposal with the water hazard focus targeting stores. Provides 1 point of mitigation against the downtown flooding problem, or 2 points if the low income grants option is chosen.

R3 Proposal: Septic Enforcement

Provides 2 points of mitigation against the septic water table problem, but imposes a -1 penalty to future work budgets due to resentment of the thorough enforcement required.

R3 Proposal: Shoreline Armoring

Protects buildings, but does nothing for the shore, and may in fact speed beach loss. Provides no mitigation of the beach erosion problem.

R3 Proposal: Solicit Industry

Solves the weak job market landuse imbalance problem (2 points of mitigation) by pulling in another industry to replace it. Also incurs a -0.1 ecology penalty unless the green option is selected. Has a small negative impact on tourism (+0.5 to the second-order tourism economy problem). Will compensate for the loss of local industry due to unsolved industrial runoff problem, in the rare case where both events occur.

R3 Proposal: Tax Breaks

Provides one point of mitigation against the industrial runoff problem, but incurs an upkeep cost of -2 money for future budgets, and does not prevent the problem from recurring.

R3 Proposal: Tree Trimming Program

This proposal is exactly the same as the first-round proposal of the same name.

R3 Proposal: Undercurrent Stabilizers

Provides 2 points of mitigation for the beach erosion problem, solving it completely. Note that in real life, this technology is still controversial; there are studies that suggest it is an effective long-term solution, but there are also many professionals who regard it as snake oil. The important thing from a pedagogical point of view is that this is a solution that addresses underlying causes (the underwater currents that cause erosion) rather than just papering over the symptoms with a short-term fix. This fits well with a number of the teaching points we wanted to make, so we chose this solution as the one that solves this problem most effectively.

R3 Proposal: Visitor Use Fees

Adds some fees that will decrease the town's popularity with tourists. Decreases hotel and condo growth rates by 5%, adds a +1 money bonus to future budgets, and provides 1 point of mitigation against the soaring land values landuse imbalance problem. Has negative consequences for tourism, obviously; +1 to the second-order tourism economy problem.

R3 Proposal: Water Purification Plant

Solves the septic watertable problem entirely (2 points of mitigation). The only downside to this proposal is that building a sewage treatment plant in round 3 is considerably more expensive than building it in round 1 as part of the Municipal Sewer Reform proposal.

R3 Proposal: Zone Affordable Housing

Increases the growth rate of houses by 5% and provides 1 point of mitigation against the soaring land values landuse imbalance problem.

Round 3 Resolution

At the end of round 3, the game checks to see to what degree each of the problems was addressed and provides text describing the outcome depending on whether the problem was solved (2 or more points of mitigation), partially solved (1 point of mitigation), or unsolved (0 or fewer points of mitigation). It then checks for second-order problems and synergies, where a side-effect of a proposal or a partially-solved problem or the interaction between such elements can cause a problem, and produces outcome text describing that effect.

Casino

If the gambling cruise is established in the town, it bring both economic traffic and a noticeable increase in crime in the waterfront area. Because the casino has such economic power, it becomes a significant force in local politics and decreases the influence of the town's businesses and residents.

Flooding Downtown

Partial mitigation through economic assistance helps businesses survive the disruption of flooding, but leaves them vulnerable; the town's economy becomes fragile and is likely to suffer serious damage the next time a flooding event happens. Partial mitigation through reducing vulnerability solves the underlying problem, but does not prevent some businesses from folding due to the disruption of the flood, resulting in some permanent damage to the local economy. Complete mitigation deals with both aspects of the problem, while no mitigation at all results in significant damage to the local economy.

Industrial Runoff

If no mitigation occurs, the parent company running the industrial park pulls out entirely, causing serious damage to the local economy and the loss of many jobs. If mitigation is partial, the parent company doesn't pull out, but it does decrease its investment in this facility, resuling in the loss of some jobs and some damage to the local economy. If the Solicit Industry proposal was passed and the industrial runoff problem left unsolved, the parent company pulls out, but is replaced by a new industy, cancelling the effects and leaving things mostly unchanged.

Beach Loss

Erosion can be halted by the use of undercurrent stabilizers. Without them, erosion may still threaten the beach, but the associated tourist trade can be preserved, protecting the tourist economy. Partial mitigation of this problem results in some damage to the tourist economy, and no mitigation results in complete loss of beach tourism and damage to the town's economy.

Black Swan Virus

Partial mitiation of this problem results in some damage to the tourist economy, and no mitigation results in complete loss of bird-watching tourism and damage to the town's economy.

Septic Watertable

Partial mitiation of this problem results in some damage to the tourist economy, and no mitigation results in complete loss of bird-watching tourism and damage to the town's economy.

Overcommercialization

If left unchecked, overcommercialization results in the town losing its unique character and sense of place and transforming into a tourist trap. Partial mitigation slows but does not stop this problem, and future action will be needed to prevent a complete transformation. This problem can be mitigated by second-order problems that affect tourism negatively: up to 2 points of tourism hit counts as 1 point of mitigation, and more than 2 points counts as 2 points of mitigation.

If unmitigated, after the end of round 3, half the condos convert to hotels and half the houses convert to stores. If partially mitigated, 20% of condos convert to hotels and 25% of houses to stores.

Weak Job Market

This problem is just like overcommercialization, but the town threatens to convert to a bedroom community. If unmitigated 90% of hotels and stores convert to condos and houses, respectively, while partial mitigation results in only 50% conversion.

Soaring Land Values

Just like the other two landuse imbalance problems, but the town will transform into a ritzy resort town. This problem can be mitigated by unresolved black swan virus or septic water table, both of which make the town less appealing. If either problem is only partially mitigated, it counts as 1 point of mitigation against this problem, and if it's completely unsolved, it counts as 2 points. If this problem is unsolved, 90% of houses convert to condos and 80% of stores convert to hotels. If partially solved, 50% of houses convert to condos and 40% of stores convert to hotels.

Second-Order Problems

The local economy rests on the two legs of the industrial park and the downtown commercial area. Mitigating either the industrial runoff or flooding downtown problems only partially generates a "hit" against the local economy, and if either of these problems is left unsolved, it generates 2 hits. Likewise, the tourist economy rests on beach tourism and wetlands tourism. Partial or no mitigation of the beach loss, septic watertable, or black swan virus problems leads to 1 or 2 hits against the tourism economy. (Note that black swan virus and septic watertable are mutually exclusive.)

The total 'hits' against the various economic sectors are summed to determine whether there are any second-order problems caused by the interaction of these problems.

Other factors that contribute to second-order problems:

• The Bird-Watching Ads and Promote Tourism proposals do not stack; if both were passed, 1 tourism hit is added to compensate.
• If Lower Housing Standards is the only proposal addressing the soaring land values problem, the lag time before it becomes effective creates 1 hit against the local economy.
• The Visitor Use Fees proposal generates 1 hit against tourism.
• If existing sales tax permits were restricted, it generates half a hit against the local economy.
• The Solicit Industry proposal generates half a hit against tourism. It also cancels the effect of an unresolved industrial runoff problem, as the new industry basically takes the place of the old industry.
• The Water Park option of the Community Center removes half a hit against tourism.
• A Research Lab removes 1 hit against the local economy.

If there are 1 or fewer hits against the local or tourist economy, there's no effect. Between 1 and 3 hits means that that sector of the economy has become weakened and fragile, and a major disruption could lead to the collapse of that sector. If there are 3 or more hits, that sector of the economy collapses and is lost: most businesses that deals with that sector close their doors.

If both sectors of the economy are weakened, the town moves onto a downward spiral of decay, where losses in one sector reinforce those in another sector.

If both sectors of the economy are lost, the town suffers total economic collapse: there are no jobs to be had, no profits to be made, and no reasons to stay in the town, which will swiftly decay into poverty and emptiness.

End of the Game

And that's the end of the game. After resolving the round 3 problems, the game calculates how the ending ecology and hazard scores compare with 10,000 randomly simulated games (where proposals were just picked at random) to give the players an idea of how the game stacks up in comparison to the spectrum of possiblities.

The question is then asked: are you happy? Does your city look like what you wanted it to? Are you satisfied with the tradeoffs you made and the outcomes that resulted? This is a lead-in to a general discussion about the game, the player's experiences, and whatever lessons the instructor wants to emphasize.

Because "what a town should look like" is a very values-laden problem, we don't want to say that the players "won" or "lost". Though nearly all players will want a high ecology score and a low hazard score, they'll differ in how much they value it, and how it compares to other aspects about their town. So we just provide a space to reflect on the outcome and discuss it.