Naghite, while a large difference in housing levels may not automatically lead to crime, I would be very surprised if this didn't make crime more likely in the slums at given levels of city sentiment. What I mean by this, is that a city of 1000 with 500 in luxury palaces and 500 in small tents, might be expected to have a higher level of crime, all else being equal, than 1000 in intermediate housing such as small casa.

It also seems highly likely that the propensity for crime to appear would be directly connected to a house's level, thus it would be expected that shacks (which have food) would be less likely to have crime than tents (which don't). On the other hand, shacks would be more likely to have crime than small casas.

I think it would be worth testing the level of crime in an all-tent city compared with one with half the population in higher level housing (of varying level) at different levels of city sentiment.

One thing I have noticed is that the tents most recently built were first to get the crime bar when city sentiment dropped. This might be worth a look as well.

It might be interesting to build luxury palaces and grand insulae in order to find out at what level of city anger the insulae spawn crime (if at all), and compare this to small tents in the same situation. This might give some idea whether it is housing level, or housing level compared to city average, or a combination of both.

[This message has been edited by goonsquad (edited 05-05-2005 @ 02:28 AM).]

When I played an all military scenario just for the fighting, my tactic was to build enough small tents to supply my barracks, acadamy, iron mines, weapon workshops, prefectures, engineers and a theatre with actor colony. I did not build wells or supply food.

If I recall correctly, this never led to crime. I'm fairly sure about this because I wouldn't have used this tactic so often if it would lead to major crimes .

This supports the theory that the difference in housing leads to crime.

GS

I understand what you are saying, that is a good test - and I will do that once I have a comprehensive list of all that I want to test.

EmperorJay

I do not see how never getting crime with low level housing supports the theory about a difference in housing levels leading to crime. I created a test city with nothing but lots of small tents without water or food and another section of Luxury Palaces. There was not even the hint of crime over 50 years. As GS pointed out, that may have been only because city sentiment was good for other reasons, so I will test this some more.

More ideas please, the more the better

Did you change anything regarding wages and taxes in that test city?

In the Valentia challenge I built a block of all small insulae (no other houses in the city), Tax 25%, Wages 38, no unemployment, people extremely pleased and no sign of crime. As soon as I started a new block the first small tent got a crime bar and spawned a protestor. Could it be that every house has its own crime "score", not unlike an entertainment score, and what you see in the Chief Advisor is merely the city-wide average?

I've got a Grand Insulae block (19 GI's) in the same challenge, same conditions (except tax is now 15%), no other houses, but "People Idolize You as a God". I'm going to see if a new tent spawns crime immediately as happened before. This is important in this challenge, since each year you get protestors costs 1 peace. Will update shortly.

Edit: To be able to build a single small tent without a protestor spawning, tax had to be lowered to 8% with wages at 38 (8 above Rome). If 10 small tents were placed at 9% or higher tax rate, all would have a crime bar and protestors would be spawned (1 at a time). A single tent was sufficient to lower mood to "people love you", 10 tents made it drop further to "extremely pleased". 10 tents at 8% would not spawn a crime bar or a protestor but sentiment still dropped to "extremely pleased".

Finally I tried 1 tent with wages 12 above Rome and tax 9%. A crime bar appeared for a short time but no protestor. This would indicate that raising wages more than 8 above Rome gives a further benefit, although no doubt with diminishing returns.

[This message has been edited by goonsquad (edited 05-06-2005 @ 07:58 AM).]

Emperor Jay

In my test city, I never changed wages (+2 over Rome) and although I did change the tax rate, I never raised it above 7%, so not really.

GS

Very interesting for sure. I would be willing to bet that the houses that immediately spawned with crime bars would have the crime bars diminish/dissappear once you started getting food into them (assuming no change to wages/taxes) -this observation in one of my cities was why I originally thought food was key. Since then I've revised my thoughts on this issue, however that might be a nice quick test to see if you get the same results as I did.

There are 3 different things discussed here: crime "bars" in houses, spawning of protestors (not rioters--I don't have enough experience to discuss them) by houses, and City sentiment. All appear to be affected by wages, taxes, and unemployment, and may be affected by other things (such as how recent was the last festival and the gods' happiness).

As goonsquad said, the type of house affects protestors and crime "bars". My guess is that for each house, an "instantaneous sentiment" is calculated and compared to thresholds which determine whether a protestor is spawned and whether crime rises or falls.

City sentiment changes slowly (I think twice/month). It may move toward some sort of average of the "instantaneous sentiments" of the houses, or may be a somewhat different calculation--it would be hard to tell.

Does City sentiment affect whether a house spawns a protestor or its crime "bar" rises? My guess is no. Does the crime "bar" of a house affect whether it spawns a protestor? My guess is no. Do the crime "bars" of all of the houses affect City sentiment? My guess is no, but not a very strong no.

Do city-wide differences in housing affect on these things? My guess is no, although it is a weak no for effects on City sentiment, based on my feelings for what programmers would have done rather than on specific observations.

[This message has been edited by Brugle (edited 05-06-2005 @ 01:29 PM).]

I have now done a little testing in this area. Mostly I have confirmed what was already known, or thought to be known. There is a lot left to check, but here's some intermediate observations.

When the city mood starts to fall, houses that are of a lower level (no food) compared to the rest of the city may get crime bars associated with them. The instant a crime bar appears on a house, a protestor also appears. Assuming no lower level housing (say all small casa or LI), then no house is likely to get crime until the city sentiment has fallen to annoyed. At annoyed, all the houses of the lowest level will get crime bars, and one protestor will spawn for the whole city. If a city has a mixture of levels, like small tents, small casa and Luxury Palaces, every small tent in the city will get a crime bar regardless of when it was created once the mood falls to annoyed. At angry, all non patrician houses with food will also get crime bars. And eventually patrician housing will also get crime bars if you continue to let city mood fall.

I believe that houses are divided into 3 categories .... no food, with food, patrician (food and wine). I need to test more whether city sentiment is based on the lowest level of housing, or the average. Also I suspect that the wage/tax rate when you build the house is significant, as opposed to how long ago the house was built.

City sentiment is like city health. There are broad ranges (not related to city health percentages) and it takes 5 full months to move from one state to the next. It will therefore take years for your city to fall from Idolize you as a God to annoyed. Your city will not spawn crime bars until the mood falls to at least annoyed, while individual houses may still spawn crime bars. If your city gets annoyed, the houses already with crime bars will increase more.

In order to eliminate crime bars within a house, the wage/tax ratio must be severely decreased - once the crime bars disappear, you can then increase taxes up again.

Your city sentiment can be either rising, neutral or falling. There is essentially no in-between. Protestors may spawn if your sentiment is falling (regardless of it current level), is annoyed or below or if you build a house that has wage/tax rate above its means while it is being built/occupied (further testing on this last point required).

To sum up

Does city sentiment affect whether a house spawns crime bars or a protestor? YES, if it falls below indifferent.

Does the actual value of city sentiment at indifferent or above affect crime bars/protestors? NO, wage/tax/unemployment affect city sentiment and can cause crime/protestors as well.

I need to do more testing.

More to follow when I having something valuable to add.

Good work Naghite. I don't understand your statement about city mood swings, I remember having had large changes over considerably less periods than a full year. City mood doesn't change instantly, but just twice a month IIRC, is that what you mean? The broad ranges you mention can be quantified, and in fact the game engine works with exact quantities, which however are not conveyed to the player.

Anyhow, city sentiment is not the issue at hand, crime is, and protestors. It seems from your observations Brugle is correct, crime bars and protestors have little correlation, so the question is, if a high crime bar does not indicate a high risk of a protestor, what does it indicate, rioteers?

To answer Joshofet's question, one possibility is robbers. In my early C3 days I never had a riot, but an occasional city would suffer from 1 or 2 muggings before I stabilized the mood.

I felt I needed to really understand city sentiment before I could continue with crime itself.

Joshofet,

All my tests were run on essentially steady state city environments. I just happen to have steady state cities of all LI housing, all casa level housing, a mix of small tents/casa level and one map which is not steady state, but approximates it well with all levels of housing. In a steady state environment, it takes 5 full months (10 checks at 2 a month) to move through the levels (idol is one check only, love, ext please, very pleased, pleased, indiff, annoy, upset, very upset angry ect...). Sudden changes can happen when population changes significantly, housing evolves/devolves across the thresholds, venus or possibly other reasons unknown to me. If you maintain unemployment exactly and your populations precisely, it takes 5 months to go through the levels. Going from 5 to 6% unemployment will instantly drop your maximum mood level to love from what I have seen. I will try and quantify these effects later.

I said that a protestor occurs the instant a crime bar appears. This I have observed many times at least. I never said that protestors could not appear at a later date. Until I have run protesting houses at the tinderbox level for more than 25 years I won't know for sure. However, after a few years of tinderbox level, like wolf suggested, I have seen thefts. My gut feeling is that protestors, theives and rioters could all appear with prolonged crime bars, but I haven't really started my testing in this area. I would also guess that riots only happen with a large number of upset houses, possibly needing annoyed city sentiment /shrug.

I agree that at indifferent or higher city mood levels, mood does not effect crime, but to restate, the things that do cause crime (tax/wages/unemployment only so far .... ) are the same things that also affect city sentiment. A falling city sentiment seems to mean that any relatively isolated house that does not recieve food will get crime bars, and hence at the spawning of those bars a protestor (the protestor is actually one tick game later, so instant building fluctuations when building in desirable/set up areas do not trigger protestors although the crime bar may appear briefly). I need to test this more, but I'm leaning towards thinking of recently built houses as spawning crime due to the tax/wage levels when they were built. Falling sentiment levels are much higher than new building levels, as far as I can tell. Here is an example, consider a city of 0% unemployment for many years with all very stable large insulae in two blocks, population 5290. Everything is perfect and city sentiment starts at indifferent

wages are +1 and tax is 11%. City sentiment slowly rises one level on the 6th month until idolized is reached. Tax is then raised to 12% still idolized, 13% still idolized or possibly dips to just love for years, 14% sentiment slowly falls one level every full 5 months until angry. At any point you can decrease the tax to 13 or 12% and the sentiment will cease decreasing and stop at that level. Crime however will not cease if it has started. It is equally true for city sentiment, if the tax is increased from 11 to 12 or 13 percent the sentiment stops increasing from its level at 11 percent. 12 and 13% are neutral tax numbers for city sentiment with respect to 11 and 14% at +1 wages for this city.

These neutral numbers are why increasing wages more than +8 have can be shown to have essentially no effect on city senitment. A city with +8 wages and +15 tax which went to idol. Increasing to +25 tax and was found to be neutral by purposely lowering the mood to love and then increasing the tax to 25 % (it remained at love for years). Increasing the wages by up to +64 (the limit for current wages) still did not make the sentiment go back to idol. Lowering the tax was the only way to increase the sentiment to idol again.

The neutral numbers may be relevant to crime prevention in terms of new housing .... or they may not.

To this point, average health or better seem equivalent. Below average health or lower may have an effect on sentiment - not tested.

Edited: City population was halved

[This message has been edited by Naghite (edited 05-11-2005 @ 02:32 PM).]

See, that's the problem with giving preliminary results, they are usually wrong

I apologize Joshofet for a few things I had wrong. When you change taxes very slowly by one each time and watch the results, city sentiment changes take 5 months to move through a level. I guess from my work on city health (where large increased health coverage doesn't speed the process up, but population changes can), I made an incorrect assumption. Thinking about what you said, I tried changing the tax rate a large amount, and you were correct. The city mood then changed a level much quicker.

It also seems that my "tax rate when built" theory is garbage I agree with others that most recent built certainly works much better.

Below average or worse health makes no difference on crime or city mood as far as I can tell. Went down to bad for longer than I'd have thought possible without a change in either.

need more testing

Are there thieves in CIII? I know about thieves and tomb robbers in Pharaoh, that can steal from tax offices and monuments, but I do not remember seeing them in CIII.

When city mood becomes very low, you may get the message that a tax collector has been robbed, accompanied with a nice movie indicating that the polite thieves have not really harmed the poor fellow, but I have nver been able to actually spot either the thief or the offended tax collector anywhere on the map.

Protestors and rioters are actual figures in the game, that can be attacked and removed by soldiers and fire marshalls, theft in CIII is directly connected to overall city mood, not to local crime levels in my experience, but I never investigated in detail.

Here are some results.

Housing - All Small Tents (5K+ pop)
wages +0
tax 8% - idol
9% - neutral - can build more small tents without crime
10% - decline

wages +1
10% - neutral
11% neutral
12% decline

wages +2
10% - idol
11% - love
12-14% - neutral
15% decline

wages +3 to +7
15% decline

wages +8
15% neutral
16% neutral
17% neutral
18% neutral
19% decline

Housing - Large Hovel 2x2 (1) and Large Casa 2x2 (10)
wages +0
tax 9% - idol
10% - neutral
11% - neutral
12% decline to angry

wages +1
11% - idol
12% neutral
13% love
14% decline

wages +2 to +4
14% - idol
15% decline

wages +5 to +7
14 - idol
15-18 neutral
19 - decline

wages +8
15% - idol
16-25% neutral

Housing - All small tents from devolution (water being taken away from 22 2x2 Small Casa)

All identical to All Small Tents

Housing - 22 2x2 Small Casa and 8 1x1 small tents

wages +0
tax 6% - idol
tax 7% - love (crime in small tents)
8% decline no additional crime until city sentiment gets to annoyed

wages +2
tax - 7% neutral
8% - decline cime in tents

Housing - 5K pop in small tents 1K pop in small casa (small casa created after the small tents)

same as All Small Tent

Housing - All Large Insulae (5K+ pop)
wages +0
tax 9% - idol
10% - neutral
11% - neutral
12% - decline

wages +1
11% - idol
12% - neutral
13% - love
14% - decline

wages +2 to +4
13% - idol
14% - neutral
15% decline

wages +5 to +7
15% neutral

wages +8
15% - idol
16-25% - neutral

wages + 64
25% - neutral

Housing - 5250 in Large Insulae and + 2 4x4 small tents then added

wages +0
tax 7% - idol
tax 8% - love with crime in 2 year old small tents
9% - neutral
10% - neutral
11 % - neutral
12% - decline

Every decline above leads to crime in small tents once the city sentiment falls to annoyed.

Every level (almost anyway) was determined by remaining there for at least 6 months. Unemployment was always less than 5% and usually less than 2%.

A neutral city sentiment means that the city mood can be anything, it will neither rise nor fall from the value at which that current tax rate was set. Crime may rise on the other hand if it had already started previously.

Here's some more

Housing - total pop 8.2K made up of 7 4x4 Lux Palaces, mostly large/grand insulae, 80 small insulae, 30 lower levels with food and one 4x4 small tent

Wages + 0
7% idol
8% love - crime in small tent
9% neutral
10% decline

Wages +8
15% love - crime in small tents

(have to lower tax to about 6% to get rid of crime from houses with crime)

Housing - total pop 8.2K made up of 7 4x4 Lux Palaces, mostly large/grand insulae, 80 small insulae, 30 lower levels with food and NO houses without food/water

Wages +0
9% idol
10% love - rare crime in small casa from devolved buildings
11% neutral - rare crime in older small casa
12% decline

Housing 8K+ pop, housing 60% small tents and 40% Palaces

Same as for All Small Tents

Letting Unemployment be 9% in small tents gave
Wages +0
tax - 9% decline and crime

Letting Unemployment be 20% in small tents gave
Wages +0
2% - neutral
3% - love / crime
4% - decline

[This message has been edited by Naghite (edited 05-14-2005 @ 12:40 PM).]

Last results as follows

Housing - total population 10K+ consisting of
2.5K in Lux Palaces
4.5K GI
5.5K in Small Casa
700 in Large Hovels

All rates equal to All Large Insulae

similarly for some small tents in lieu of large hovels - same as for Large Insulae with tents

Summary:

The above results confirm with data, many of the existing thoughts on crime. Housing levels do affect crime, more recently built buildings are more likely to get crime and lower levels of housing (without food) are also more likely to get crime.

A mixture of housing levels (between getting food and not getting food) results in crime at lower tax for same wage.

A mixture of houses without significant high level housing is worse off for crime and city sentiment than an all tent city.

Higher level housing (Insulae) and better mixed with low level housing will get crime early, but city sentiment will not drop as compared to the same city without lower level housing. With enough high level housing, crime is also reduced to almost zero.

Crime was most often observed soon after crime bars spawned on a house. The lowest level of housing was usually the first type to get crime, with the newest of that lower level usually being the first to get crime.

Crime is related to city sentiment in two ways. Crime normally starts within non-food fed houses when city sentiment first starts to fall. Secondly, whenever a city senitment falls to annoyed, the lowest level of housing gets crime bars.

One must lower the taxes considerably below normal idol limits in order to eliminate crime once it has occurred. Afterwards, the taxes can be increased again.

Unemployment greatly effects both crime and city sentiment, keep it low.

I am pretty sure (not 100%) that protestors can spawn at any level of the crime bar, but very rarely.

I have never seen thieves, just the message about them.

The bests way to stop crime
1) As Goonsquad suggested, keep taxes low while building until those houses have food
2) evolve all houses to get food as a rule of thumb
3) keep unemployment low
4) pay high wages

Does crime in a house matter? Apparently not (other than spawning a protestor once), if city sentiment is OK. Rioting (like emigration) is apparently caused by poor enough city sentiment (not by crime in houses).

I added a few small tents to my no trade Valentia and let it run. The crime bars for the tents quickly rose to the maximum ("This whole area is a tinderbox of unrest! Crime is endemic, and anything might happen") and Peace increased by 4 instead of 5 at the end of that year, but I noticed no other effects in several years.

Therefore, a city with some tents can have a high tax rate without problems, assuming that other factors (such as a high wage rate and low unemployment) keep city sentiment reasonable. (Perhaps this was obvious to some players, especially those who routinely use scattered tents to provide labor access to industries, but it was news to me.) Even when the final city will have no poor houses, it may be useful to have some tents for a long time during development.

So I realize I'm like a year and a half too late on this, but ...

I was messing around with Lutetia last night to do some experiments with housing blocks (specifically palace blocks); so I wasn't really playing the game very seriously; just screwing around.

My tax rate and wages were set at the normal rates (7% taxes, wages same as Rome). Most of the plebian houses I set up were fairly well serviced. Food stocks apparently got very low for the plebian graneries, though (where I had neglected them), while the patrician housing still had maybe 1/2 - 3/4 granery full of food and I wasn't haven't any problems supplying them. As a result of all this, most of my plebian housing evolved back to tents (because of the food shortage); however, there were a few insulae left (there had been quite a number of casas and insulae before, but as the food shortage got more severe, more and more housing devolved).

I ended up having quite a number of problems with crime in the low-income tent areas, which I mostly ignored at first since I was mostly just trying to test stuff with the patrician housing. It should be noted that with the exception of the small number of insulae, about the only housing in the city was tents and palaces, so there were extreme disparities in wealth.

Obviously, all of this is just observational evidence, but it would seem to me that wealth disparities are more likely to produce crime because I didn't particularly see what could have been causing the crime otherwise. While, I had ignored the plebian housing to some extent, they still had more than adequate services (and indeed, as I suggested, there were quite a number of insulae before I let them all devolve due to the food shortages). Unemployment was only at 1 or 2%. The Gods were either happy/pleased and a few were perhaps indifferent. Everything else that I could think of that might cause crime seemed to be in decent shape; maybe not excellent shape, but I had never had problems before under the same set of circumstances (minus the extreme wealth disparities).

As can be gathered by reading some of the posts in this thread, crime is very complicated and nodody seems to be certain exactly how it works. The following is my understanding of it-

You have an overall city satisfaction level, which if it falls to certain levels, leads to slowed immigration, then no immigration, then emigration, then riots. This level is determined by wages, taxes, unemployment, maybe overall city housing level and maybe overall city religion coverage. It can be temporarily raised by having a festival or receiving a Venus blessing, and temporarily lowered by being cursed by Venus but the most efficient method of raising it is by paying wages higher than Rome. 8dn above Rome seems to be the breakpoint, after this returns diminish very rapidly.

Aside from this you have crime levels in individual houses, the factors of which are overall city satisfaction level (as described above), variance of individual house levels from the city average, and length of time since a house was first occupied. This is indicated by a red crime bar. It is this individual rating that causes protesters to appear and thefts from teasury to occur (not 100% certain on this last one). Protesters appearing in new housing or low level housing are really only a problem if you are trying to reach a peace target in the shortest possible time, as any year in which a protester appears causes peace to rise by one less than otherwise. You can test this by looking at the crime overlay, and can overcome it by raising wages, reducing unemployment to about 10%, and lowering taxes.

[This message has been edited by goonsquad (edited 01-23-2007 @ 09:15 AM).]

This thread caught my attention while I was looking for information about protestors. Although it is somewhat dated, it is the most recent 'crime' thread found in a search so I'll pitch in here so that everything is in one place.

The principal mistake made by earlier posters is to conclude that crime starts because city mood is falling (defining 'crime starts' as the appearance of red 'crime bars'). That's a bit like saying that the sun only comes up when it's light. Crime bars indicate poor sentiment in individual houses. Since overall city sentiment is an average of all the individual household sentiments, falling sentiment in some houses pulls down that average. Falling city sentiment is an effect, not a cause - it falls because there is discontent, it does not cause the discontent. Indeed sentiment elsewhere in the city may be completely unaffected.

Crime bars are not really crime bars at all. They are indicators of negative sentiment - more correctly they indicate the degree by which sentiment in a house falls short of the 'median' figure. As joshofet posted, the hard numbers are not conveyed to the player, but examination of the Buildings List shows that median figure to be 50, representing indifference.

No house will have a crime bar, no matter what wages/tax/unemployment combination, if its sentiment is 50 or above. The moment it falls to 49 a short bar appears, and the overlay reports 'There is only the occasional crime in this area' (right-clicking reports 'This is a peaceful neighborhood). Once it falls to 40 the bar increases in height. The overlay continues to report the same message but the right-click message becomes 'There is some crime here, nothing serious'.

The bar continues to rise (and the messages become more alarming) at 30, 20 and 10 points until finally sentiment falls to zero (when you get the 'tinderbox of unrest'/'threat of riots' warnings). So the height of the crime bar is a direct indicator of the level of sentiment in a house (and nothing else). Whether or not city sentiment also falls depends on whether there are enough houses with poor sentiment, relative to the total housing stock, to cause a significant difference when averaged out.

When a vacant lot is first placed it is given a neutral sentiment value of 50 (Brugle's 'instantaneous sentiment' theory upheld). Note that this is only 1 point higher than the crime bar threshold. At each 2-weekly sentiment review the game gives any vacant lots a new value based on the difficulty setting as follows:

Very hard no change
Hard 60
Normal 70
Easy 80
Very Easy 90

This occurs at every review, so a save made at Hard can be reloaded on Easy and all the vacant lots will be upgraded to 80 (at the next review). This reveals both a defect and an exploit - the defect is that if a lot becomes occupied very quickly (before the next review) it will not get the sentiment bonus applicable to the difficulty level (this would only happen right by the entry point) and the exploit is that you can place lots, switch to Very Easy, wait until they become occupied (and a sentiment review has occurred in the meantime) then switch back to the difficulty you were playing. You will retain the bonus.

Once the tent is occupied its sentiment is free to rise or fall according to game conditions (except in a very small city, where it is controlled). Most of the various influences (tax, wages, unemployment, festivals, Venus effects, maybe others) are city-wide and affect all houses equally. I haven't quantified this but in a test a city with 10% tax, 96% registered (don't know if that matters - it does in Pharoah), wages +5 and unemployment 4%, sentiment in all houses rose by 2 per review until the maximum 100 was reached. If the net result of these influences is negative then inevitably all houses (even palaces) will develop crime bars sooner or later unless something is done.

There is, however, another (unquantified and unknown) negative effect which is only applied at the month end (not the mid-month review) and I'm fairly sure it relates to differences in housing levels. I'm guessing it occurs when there is a difference above a certain threshold between one housing level and another. Let's say the threshold is 10 (I don't know) so small tents would get this effect if there are Grand Insulae on the map, large tents if there are Small Villa, etc. I also think that the greater the difference the greater the effect. That is why it is seen first in small tents. The reason it is seen first in the most recently placed is that the first visible indication is when sentiment goes below 50, and new tents start from a lower value than established ones and so get there more quickly. In my test city (max housing Grand Villa) my small tents go down 7 points at a time (actually 5, but they should have gone up by 2) but a large shack stays happily at 100. I'll investigate further and report.

As for the actual occurence of crime, the manual suggests that the higher the crime bar the more likely it is that the occupant will commit a crime (indicated by the presence of a protestor). I don't see this - one tent spawned a protestor as soon as sentiment went below 50 but I waited 15 months for another from another tent (by which time the bars had been at full height for some time). The manual also indicates that if city sentiment is very good the worst you'll get is a 'mugging' (an unreported crime, not the same as a tax collector robbery) - that seems to be true. My tents were angry enough to riot but did nothing because overall sentiment was 'people love you'

I really ought to know better. Like Naghite, I've found thst preliminary results tend to be wrong

Sentiment does not change equally for all houses. Nor is my theory about disparity of houding levels fully upheld. In a city with all Large Hovel and one small tent, chsnages in sentiment at different tax/wage levels are always 1 point less in the tent than they are in the hovels. The once-monthly negative effect I mentioned amounts to an additional three negative points in the tent (in this particular city) once it gets a crime bar.

I have more theories about this (to do with c3_model.txt), but I will refrain from further speculation until I have satisfied myself that I won't be changing my mind as soon as I look at another map,

Erm, Okay. So basically, try to keep everyone happy, cause you don't really know about this one.
Pity rioting is only enjoyed by the really bad players, but I've been bad enough to get the tax collector robbery. If you go to disaster, there is a "criminal" with the same picture as the protester. I think you can get this without the tax collector robbery, but I don't know. Rioters are basically an army if I remember back to a very long time ago. They are similar to natives/gladiator revolt, but they like burning stuff. Woo, sounds fun, I'm going to make a riot!

Hi pigsnoutman

Yeah, I guess we don't really know everything, but we know a fair bit. The manual gives a fair bit away but, as usual, conceals the specific numbers.

Crime, so far as C3 is concerned, amounts to three possible events:

'Petty' crime, which is purely conceptual. You have houses in your city with sentiment below 50 and criminals are about (not visibly). You may see a protestor, and this protestor may be dealt with by a prefect. Either way, the crime he is about to commit is so trivial you won't even be told about it, and it has no impact on your city. It's just a warning shot across your bows.

'Serious' crime is manifested as tax collector robberies. This can also happen in a city with houses a little below 50 sentiment, but it's more usual when sentiment is somehat worse. Again, you may see a protestor, and he may be dealt with by a prefect, in which case the robbery won't be committed - this time. But a robbery will be committed sooner or later (unless your city is swarming with prefects).

'Anarchy' - riots - is the most serious. This will only happen when you have pretty bad conditions.

The level of sentiment in an individual house reflects the severity of the crime that house would like to commit. However, they are to some extent dissuaded from doing so if the rest of the city is fairly happy.

Because of the way sentiment works it is likely that your tents population will be less happy than the rest of your city, and that is where crime is likely to appear. Because overall sentiment is an average of individual household sentiments a small number of absolutely furious tents in an otherwise happy city might not weigh the average down very much. On the other hand, a large number of slightly unhappy tents might weigh the average city sentiment down much more heavily. What appears to spark actual crime (ie robberies) is a sufficiently angry tent in a sufficiently unhappy city.

I'm not aware that anyone has fully quantified this, but over at CBC I took a look at one contest entry where mood was only just 'pleased' at 60.75. That contestant had a number of tents with zero sentiment (tinderbox of unrest) but no robbery was committed so long as sentiment remained above 60. If sentiment fell to 'indifferent' a tax collector robbery took place almost immediately.

I'm pretty sure that if the tents were a little happier they would not commit crime in an indifferent city but would do so in an annoyed one.

So crime depends not just on the unrest in the house, but also on the overall sentiment of the city. So while I don;t really know, I'd say that you will not get a robbery in a city that is at least pleased, and you'd only get one in an annoyed city if there is at least one house ranked as a 'tinderbox of unrest'. Probably a very annoyed city (I don't remember how the next level of annoyance is actually described) would get a robbery with a less angry tent in it. Riots would take some provoking.

Thanks, that was a useful summary. I did have some fun starting riots, which just took 0 wages and 25% tax. Poor people, no wonder they got annoyed!