Jump to content
270soft Forum

Release: President Forever 2012 Alpha v. 1.0.7


Recommended Posts

See here.

The major problem I've found at the moment is the AI which does not appear to do anything except some barnstorming.

I repeated my strategy from v1.05 to wait until 3 weeks before the election to create negative ads, I deploy them all in the last 10 days... I get 'Obama campaign in meltdown' or something similar, he goes to -15 momentum and I sweep the states by 58 to 42%. Both times I didn't even spend all my money and had 15 million left

The only positive momentum Obama seems to get is from events or endorsements. I don't get the impression he uses many ads (or he uses them very ineffectively)

I do like the addition of surrogates very much. Adds a new dimension to the game to be able to have people campaign, fund-raise and spin when you are not able to.

Also like the barnstorming/debates breakdown of how stats contribute, and being able to upgrade organization strength as well as footsoldiers ('08s problems with these: i) organization was national and not sufficiently granular - huge jump between 2->3 establishment, ii) footsoldiers were something you maintained and lasted a month... result was that you would get a huge number when a big block of endorsers decided (usually together) and a month later they would all disappear at the same time)

I also like how your ability to 'buy' endorsers is more limited and based on your issue positions... in '08 you could be a far leftist and get the NRA with sufficient CP 'influence' spamming + national momentum.

I like how Leadership, Experience etc are modelled as issues... potentially could demographic traits also be modelled that way? Women (Clinton), Born-Again, Southern, Mormon, Black and other identity politics obviously have real impact in a US election - extreme example of somebody bad at these = Mitt

Can't wait to see what you do with primaries

A couple things I have thought about changing from 2008 game primaries:

-in '08 momentum didn't really work properly... it appeared there was far more (or it had more effect) positive momentum than negative. In that game candidates' polling generally only goes up over time as undecideds reduce. The only thing that change the dynamics significantly are: 1/ a primary victory, and 2/ other candidates dropping out.

This is a problem especially for the '12 campaign scenario because we have seen huge up-swings (at the expense of other candidates) and down-swings all along the campaign many of them before any primaries started and others between and unrelated to primary outcomes. Bachmann, Cain, Newt, Santorum have all had rapid rises and falls (typically doubling their polling within a week) 5%->25%->5%, two of them (Newt and Santorum) have had multiple surges, something I have never seen happen in the '08 game.

-negative issue focus... the effectiveness of a negative attack should be diminish if you have the same position on an issue as you're attacking an an opponent on. Currently it is possible to attack Obama on healthcare in right-wing states as Romney even though you are also Center-Left. That -2 should become a -0, and only full power if you were Center-Right, two positions away

-Endorsers: rather than giving momentum should influence a block share of votes. I think this should be similar to the effect when you are endorsed by a competitor (get a % of their votes as well as momentum). The momentum even from the most powerful endorsers are usually useless. In a normal campaign a +2 momentum nationwide will usually only get you +0.5% by the end of that week, nothing like the +10% we actually see from significant endorsements. It makes Governors and former Presidents look pathetic and PIPs a waste of time (except to bribe competitors with) if all you get is +0.5%. It is a much better strategy spend a few PIPs bribing each of your 2% bottom-tier rivals who'll add more to your polling (and potentially delegates) than all the endorsers put together.

-This could maybe work with the momentum issue above, and help model swings better... as well as one-off endorsers you could have different factions which could back one candidate at a time but also withdraw and change which candidate they back... maybe through an endorser influence meter, but which keeps going and does not stop at 100, if any candidate is more than say +50 points ahead they obtain the full benefit, otherwise it is shared through some formula

-this would represent groups who control blocks of voters and whose support does swing an election... in '08/Republican terms this could be 'Tea Party' most obviously (whose switch between Bachmann/Cain/Newt explains their rise and falls), but maybe also 'Evangelicals', 'Fiscal conservatives', 'Military establishment', 'Neo-conservatives' etc

-If you set eg the Tea Party at 15% of Republican Primary vote share, that would explain candidates like Santorum zooming into second place after Bachmann withdrew and after Newt was hit by serious negative momentum (the effect of suddenly overtaking rivals in 'endorser points')

-That kind of mechanic is probably not applicable to general elections

  • Upvote 1
Link to post
Share on other sites

Two other general points:

1/ regional issue profiles. Issue profiles should remain generally the same but it should be possible to raise the profile of a particular issue in a target region, or have it so that some regions weight certain issues more strongly by default (eg, a state with a very polarising issue, for example jobs might be weighted more heavily in Michigan/Nevada than say Connecticut, in the UK devolution is much more of an issue in Scotland than Kent)

2/ Moderate/centrist positions should provide less power than solidly right-wing/left-wing positions. The best candidate is usually a mix of centrist with a few center-left and center-right positions, in the game it hits the 'ideal' spot for more states without many strongly negative (-2) positions so there is no trade-off. Less moderate positions hit the ideal spot in fewer states but should generate more enthusiasm... I think that would follow reality in that perfect moderates are not the automatic winners in a lot presidential democracies, certainly not in the USA and France, their strength is in having more a wider 'potential' vote because they alienate fewer people

don't know whether any of this can be fitted in to '12 but these are some things that bugged me which I would like to see in the game in the longer term

Link to post
Share on other sites

@Dr ABC,

Thanks for the feedback - there's a lot there, but a few notes:

The computer player AI is not complete - computer players should start becoming more competitive once gameplay mechanics are more settled.

In theory, demographic traits could be modeled in a similar way to experience, and so on. One way would be to create a new issue, and then give candidates 'positions' on that issue that reflect their demographic trait(s), and then lock the positions, with varying 'centers' on the issue varying by region.

The 'if attacking someone with the same position, don't have same bonuses' idea is a good one - noted. The question is how to do this without making the mechanics too complex. One simple way to capture this would be to have a 'hypocrisy' news story that had some chance of occurring if attacking someone with same position as self.

Blocks require modeling the electorate in a different way. This would allow for several other more realistic elements to come into play. We'll see.

Thanks again!

  • Upvote 1
Link to post
Share on other sites

Blocks require modeling the electorate in a different way. This would allow for several other more realistic elements to come into play. We'll see.

Thanks again!

Wow, this sounds exciting. As someone who was previously giving a hard time about the release date, I for one would love to see a Pres Forever 2012 that incorporates as many realistic elements as reasonably possible. The product so far has been very improved from Pres Forever 2008. But the ideas Dr.ABC posted are great.

Link to post
Share on other sites

A few more details (again, these might not be implemented, but feedback welcome on it). The game currently uses a relatively simple "committed-leaning-undecided" model, where each candidate has a certain starting percentage of each, and it is relatively difficult to move committeds than the other categories, and leanings than undecideds, to another candidate.

To more accurately model the dynamics of, say, the 2012 Republican primaries, there would be favorability numbers attached to each candidate. These numbers would then vary by 'blocks', which would be exclusive and exhaustive (probably) groups specified in, say, the percentages.xml file. For example, on the Republican side, there might be the 'Tea Party' block, the 'Evangelical' block, the 'Financial Conservatives' block, and so on. Each block would then have a certain average favorability rating for each candidate generated, based on the block's issue positions and then modified by campaign efforts to increase or decrease favorability of various candidates. There could also be customized favourability adjusters to account for factors not easily captured in issue positions - the suspicion among some Social Conservatives about Romney based on his past positions, for example.

If things were done this way, then you could see the bouncing around of a block between candidates while more or less ignoring more ideologically distant candidates, as well as rapid increases and collapses in support when favorability numbers reached thresholds vis a vis other candidates.

Again, this may not be implemented.

Wow, this sounds exciting. As someone who was previously giving a hard time about the release date, I for one would love to see a Pres Forever 2012 that incorporates as many realistic elements as reasonably possible. The product so far has been very improved from Pres Forever 2008. But the ideas Dr.ABC posted are great.

  • Upvote 1
Link to post
Share on other sites

Hooray! :D

Quick question: Will the game include the same candidates from the P4E2008 Beta test? Or the candidates featured on the blog?

POLLWONK

Currently enabling primaries, which is a big step. I’m guessing either the end of this week or next week the next version will be released.

  • Upvote 2
Link to post
Share on other sites

In the final version, I don't know which candidates the game will include. We will start adding candidates, probably largely along the lines of the ones featured on the blog.

Hooray! :D

Quick question: Will the game include the same candidates from the P4E2008 Beta test? Or the candidates featured on the blog?

POLLWONK

Link to post
Share on other sites

A few more details (again, these might not be implemented, but feedback welcome on it). The game currently uses a relatively simple "committed-leaning-undecided" model, where each candidate has a certain starting percentage of each, and it is relatively difficult to move committeds than the other categories, and leanings than undecideds, to another candidate.

To more accurately model the dynamics of, say, the 2012 Republican primaries, there would be favorability numbers attached to each candidate. These numbers would then vary by 'blocks', which would be exclusive and exhaustive (probably) groups specified in, say, the percentages.xml file. For example, on the Republican side, there might be the 'Tea Party' block, the 'Evangelical' block, the 'Financial Conservatives' block, and so on. Each block would then have a certain average favorability rating for each candidate generated, based on the block's issue positions and then modified by campaign efforts to increase or decrease favorability of various candidates. There could also be customized favourability adjusters to account for factors not easily captured in issue positions - the suspicion among some Social Conservatives about Romney based on his past positions, for example.

If things were done this way, then you could see the bouncing around of a block between candidates while more or less ignoring more ideologically distant candidates, as well as rapid increases and collapses in support when favorability numbers reached thresholds vis a vis other candidates.

Again, this may not be implemented.

Hi Anthony

Regarding the favorability ratings, didn't we have that in a way already in '08?

There is a stat if you go into player view, which is called 'electability' and I always assumed was based on national issue distances and appeared also to be affected by momentum... I wondered what that did and if that came into use in gaining undecideds

Obviously if that could be generalised to sub-categories/blocks (not just national average) it would open up many more possibilities -

in a primary, taking a stance which might be negative 'national' for favorability, would be worthwhile if it improves your favorability in a specific upcoming state... when Iowa is coming up you stress your social conservatism and ethanol subsidies, even though it plays badly in New York/California, when it's Florida stress Cuba or maybe support for NASA etc.

More generally Romney tacked to the right to win the Republican primaries though it damaged his national 'moderate' image, and has set up specific problems eg on immigration with Hispanic swing voters in the general

This would also work just as well in legislative races with winner-take-all seats, if you are the NDP and have an opening to break through in say Quebec, you might want focus on appealing to francophones more than the general population

I think the problem is that if blocs are mutually exclusive identifying what the sub-categories should be.. in reality every voter has varying mixtures of affinity to different groups, and only a minority are really loyal to one ideology or faction.. and the strengths of different factions can vary over time

I recently played Democracy 2 again (notice you have an ad linking to them on the front page) and they have a system for multiple alliegeances, shifting groups sizes, and using strength of support and motivation (turn-out) in overlapping groups to influence re-election results - I don't know the mechanics and I'm not particularly happy with it (don't think the game is either realistic or interesting) but there might be something in there

Link to post
Share on other sites

Another point about issue profiles, in a previous post I said that individual states/regions should be allowed some variation in issue profiles. After consideration I don't think even that would be enough

The problem is that even with an issue at -2 (issue distance penalty cap) and that issue was highest profile (level 4, purple), it wouldn't be much of a problem as long as you are average on the other issues. Instead of advertising on that key issue (which all the media is obsessing about), you can ignore it and create ads on two or three +0/+1 issues which are low/medium profile and still end up generating the same momentum. You'll be spending 3x as much money but it's doable for the odd unfavorable state.

The problem is that weaknesses (or strengths) linked to specific issue positions aren't that relevant because it is affordable to overcome them

Would it be possible to increase the effect of issue profile?

Instead of +0/+1/+2/+3 bonuses have a formula which weights issue positions, eg use the square of that issue profile

example conservative rural state; for simplicity just 6 issues

Ethanol high +0

Abortion very high -2

Immigration medium +2

PIPA/SOPA low +0

Environment low +0

Foreign policy medium +1

average issue position +0.16

Total of squared issue weights for this state is (3^2 + 4^2 + 2^2 + 1^2 + 1^2 + 2^2 = 35)

Once reweighted by issue, you get

9/35 * 0

+ 16/35 * -2

+ 4/35 * +2

+ 1/35 * +0

+ 1/35 * +0

+ 4/35 * +1

weighted average issue position = -1.2

Locally high issue profiles (and changes in profile) can make issues dramatically more important than they would be normally

On the national stage, you could it the same way but use the square of the average of regional issue profiles as a weight. This should mean that varying issue profiles or tailoring your positioning is important in specific states, but won't make a huge difference on the whole (except of course, that you need to win a specific sequence of primaries to win a party nomination)

Maybe that could be generalised to endorsers and voter blocs if they are included. An endorser or faction may have only one single issue they're interested in which would be 'very high' and the rest 'low', this extreme would result in a weighting of 16/(4^2 + 17 = 33) 48%. On the national level assuming they average out that same issue would be 1/18 = 5.5%

Add the character attributes/demographic characteristics as 'issues', and reweight them as well. Then you could have endorsers like:

The Economist Magazine: R on Budget, C on other issues. Issue profiles: Budget and 'Issue Familiarity' Very High, rest Low

Focus on the Family: FR on Abortion, Stem-Cell, Same-Sex Marriage. Issue profiles: Very High on those issues, High on 'Integrity' and 'Born-Again/Evangelical'

Just thinking this could also have the effect that unusual or cross-party candidates could do very well.. a conservative Southern Democrat winning endorsers like the Family Council above, against a social moderate fiscal conservative Republican. Could be mitigated by adding 'Republican', 'Democrat', 'Independent' to the characteristics that can be weighted. If a particular party is listed and has a high profile it would make upsets less likely

Is this something that would be possible?

  • Upvote 1
Link to post
Share on other sites

"Regarding the favorability ratings, didn't we have that in a way already in '08?"

Yes, but it was secondary, and an attribute of the candidate (see below).

"I think the problem is that if blocs are mutually exclusive identifying what the sub-categories should be.. in reality every voter has varying mixtures of affinity to different groups, and only a minority are really loyal to one ideology or faction.. and the strengths of different factions can vary over time"

The advantage to making it mutually exclusive is that it is conceptually and practically simple.

Another way to go is to create a database of voters in each region, and then assign each various attributes (age, sex, ethnicity, and so on). Then, one would need a way to connect a specific profile with a specific set of issue positions that voter would have. This gets ... really complex.

Regardless, the basic idea of making a favorability rating primary to voter affiliation is simple and offers lots of possibilities in terms of modeling voter movement. Favorability of a candidate would be an aggregate of voters' assessments of that candidate, not an attribute of the candidate, and so would vary based on region or sub-set of a region's voters (as in party primaries).

Hi Anthony

Regarding the favorability ratings, didn't we have that in a way already in '08?

There is a stat if you go into player view, which is called 'electability' and I always assumed was based on national issue distances and appeared also to be affected by momentum... I wondered what that did and if that came into use in gaining undecideds

Obviously if that could be generalised to sub-categories/blocks (not just national average) it would open up many more possibilities -

in a primary, taking a stance which might be negative 'national' for favorability, would be worthwhile if it improves your favorability in a specific upcoming state... when Iowa is coming up you stress your social conservatism and ethanol subsidies, even though it plays badly in New York/California, when it's Florida stress Cuba or maybe support for NASA etc.

More generally Romney tacked to the right to win the Republican primaries though it damaged his national 'moderate' image, and has set up specific problems eg on immigration with Hispanic swing voters in the general

This would also work just as well in legislative races with winner-take-all seats, if you are the NDP and have an opening to break through in say Quebec, you might want focus on appealing to francophones more than the general population

I think the problem is that if blocs are mutually exclusive identifying what the sub-categories should be.. in reality every voter has varying mixtures of affinity to different groups, and only a minority are really loyal to one ideology or faction.. and the strengths of different factions can vary over time

I recently played Democracy 2 again (notice you have an ad linking to them on the front page) and they have a system for multiple alliegeances, shifting groups sizes, and using strength of support and motivation (turn-out) in overlapping groups to influence re-election results - I don't know the mechanics and I'm not particularly happy with it (don't think the game is either realistic or interesting) but there might be something in there

Link to post
Share on other sites

"Would it be possible to increase the effect of issue profile?"

Yes, it is simple to increase the effect of issue profiles on momentum. The importance of a region's issue profile is already a multiple, so a profile of 4 makes the momentum 4x the power of a profile of 1. Perhaps what you're saying is that an issue's importance can be much more than 4? Perhaps instead of a ^2 system, a profile of 1 to 10, where 10 is 10x the impact of 1? (Using exponential systems is usually less intuitive from a user's perspective.)

Another point about issue profiles, in a previous post I said that individual states/regions should be allowed some variation in issue profiles. After consideration I don't think even that would be enough

The problem is that even with an issue at -2 (issue distance penalty cap) and that issue was highest profile (level 4, purple), it wouldn't be much of a problem as long as you are average on the other issues. Instead of advertising on that key issue (which all the media is obsessing about), you can ignore it and create ads on two or three +0/+1 issues which are low/medium profile and still end up generating the same momentum. You'll be spending 3x as much money but it's doable for the odd unfavorable state.

The problem is that weaknesses (or strengths) linked to specific issue positions aren't that relevant because it is affordable to overcome them

Would it be possible to increase the effect of issue profile?

Instead of +0/+1/+2/+3 bonuses have a formula which weights issue positions, eg use the square of that issue profile

example conservative rural state; for simplicity just 6 issues

Ethanol high +0

Abortion very high -2

Immigration medium +2

PIPA/SOPA low +0

Environment low +0

Foreign policy medium +1

average issue position +0.16

Total of squared issue weights for this state is (3^2 + 4^2 + 2^2 + 1^2 + 1^2 + 2^2 = 35)

Once reweighted by issue, you get

9/35 * 0

+ 16/35 * -2

+ 4/35 * +2

+ 1/35 * +0

+ 1/35 * +0

+ 4/35 * +1

weighted average issue position = -1.2

Locally high issue profiles (and changes in profile) can make issues dramatically more important than they would be normally

On the national stage, you could it the same way but use the square of the average of regional issue profiles as a weight. This should mean that varying issue profiles or tailoring your positioning is important in specific states, but won't make a huge difference on the whole (except of course, that you need to win a specific sequence of primaries to win a party nomination)

Maybe that could be generalised to endorsers and voter blocs if they are included. An endorser or faction may have only one single issue they're interested in which would be 'very high' and the rest 'low', this extreme would result in a weighting of 16/(4^2 + 17 = 33) 48%. On the national level assuming they average out that same issue would be 1/18 = 5.5%

Add the character attributes/demographic characteristics as 'issues', and reweight them as well. Then you could have endorsers like:

The Economist Magazine: R on Budget, C on other issues. Issue profiles: Budget and 'Issue Familiarity' Very High, rest Low

Focus on the Family: FR on Abortion, Stem-Cell, Same-Sex Marriage. Issue profiles: Very High on those issues, High on 'Integrity' and 'Born-Again/Evangelical'

Just thinking this could also have the effect that unusual or cross-party candidates could do very well.. a conservative Southern Democrat winning endorsers like the Family Council above, against a social moderate fiscal conservative Republican. Could be mitigated by adding 'Republican', 'Democrat', 'Independent' to the characteristics that can be weighted. If a particular party is listed and has a high profile it would make upsets less likely

Is this something that would be possible?

Link to post
Share on other sites

"Would it be possible to increase the effect of issue profile?"

Yes, it is simple to increase the effect of issue profiles on momentum. The importance of a region's issue profile is already a multiple, so a profile of 4 makes the momentum 4x the power of a profile of 1. Perhaps what you're saying is that an issue's importance can be much more than 4? Perhaps instead of a ^2 system, a profile of 1 to 10, where 10 is 10x the impact of 1? (Using exponential systems is usually less intuitive from a user's perspective.)

Hi

I see, I didn't know that issue profile was a multiplier in '08, I thought it acted as a a bonus to ad / barnstorming strength

In any case, I think the current range and even 1-10 is probably insufficient.

I thought exponential because it would allow the greatest diversity of situations - in actual elections out of all the issues raised over the course of a campaign there are only a few that dominate in any given week. I think it is odd that say, in the wake of the Keystone/Gas Price increases, Solyandra, Planned Parenthood controversies that have marked '12 (whatever issues these would refer to in game terms) you could just ignore all of them and talk about a bunch of low profile issues and still win primaries... Or that Romney could win South Carolina purely by spending/talking enough about his business tax proposals while Gingrich campaigns on conservative hot button issues

I don't know if having exponential multipliers and re-weighting would be that counter-intuitive, if you were simply presented with the % share of media attention next to each issue in a region. As it is I don't think most players know the exact mechanics of the game (I don't and the help file is not very detailed)

Link to post
Share on other sites

I see what you're saying, but a '% share of media' is a little different. That should act through the media, which in the game as currently works would be the news and interviews.

I can see tying issue profiles more carefully to media attention, perhaps. So, if there's an event which triggers a news story on an issue, that issue's profile is temporarily increased a large amount. Once the news story ends, the profile falls back to where it was before the news story.

However, the idea of adding 'news percentages' is interesting. It could go beyond the top 3 headlines the game currently shows. We'll see.

Hi

I see, I didn't know that issue profile was a multiplier in '08, I thought it acted as a a bonus to ad / barnstorming strength

In any case, I think the current range and even 1-10 is probably insufficient.

I thought exponential because it would allow the greatest diversity of situations - in actual elections out of all the issues raised over the course of a campaign there are only a few that dominate in any given week. I think it is odd that say, in the wake of the Keystone/Gas Price increases, Solyandra, Planned Parenthood controversies that have marked '12 (whatever issues these would refer to in game terms) you could just ignore all of them and talk about a bunch of low profile issues and still win primaries... Or that Romney could win South Carolina purely by spending/talking enough about his business tax proposals while Gingrich campaigns on conservative hot button issues

I don't know if having exponential multipliers and re-weighting would be that counter-intuitive, if you were simply presented with the % share of media attention next to each issue in a region. As it is I don't think most players know the exact mechanics of the game (I don't and the help file is not very detailed)

Link to post
Share on other sites

Now that primaries are being added, is anything being done to address the fact that Obama will have a 1 year head start campaigning against the GOP candidates? This was a problem in 2008 in scenarios with an incumbent. In another thread I suggested a cap on what he can do before a certain date.

  • Upvote 1
Link to post
Share on other sites

This will be addressed in game testing. We'll see how things feel, but one thing to keep in mind is that an incumbent does have various advantages if unchallenged in his party's primaries.

Now that primaries are being added, is anything being done to address the fact that Obama will have a 1 year head start campaigning against the GOP candidates? This was a problem in 2008 in scenarios with an incumbent. In another thread I suggested a cap on what he can do before a certain date.

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...