Beginning the PTW transition process for GOTM

We have a great deal going on this month in addition to smackin' Impis and being grateful that the Romans did not maintain control of the world and require us all to still speak Latin.

One of the first things we would like to do, is to faciltate the process of discussing the impact of the Single Player features that can be included in a game of Civ3 that might be played in the GOTM.

Unfortunately for us all, many of the technical issues related to game play in PTW have been lost and confused in a number of other features that were designed and tested only for implementation into the multiplayer game modes.

Here is a link to a review article that attempts to look at a number of the features of the PTW expansion pack and to provide some basis for discussing how these features might impact game play when confronted by a knowledgeable and experienced pool of players as we have here in the Games of the Month.

PTW Review of Single Player Game Features

This article is an exclusive advance discussion tool for use by members of the CivFanatics gaming community.

Moderation of discussion of this article will keep the topic focused on issues related to Single Player game play and will not generally allow the discussion to drift into issues related to multiplayer, short games, or scenario support.

This is also not a PTW bashing or PTW fanboy thread, so make sure that your contributions focus on discussion of how the PTW features may effect single player game play and you will avoid the indignity of the red ink and the

That is a great review and analysis of PTW differences (and a lot of work), thank you! :

I have only experimented a bit with PTW so far. I did a number of starts trying to get a feel for differences. Mostly tried it on deity/large/pangaea and played only a couple of times past 10AD.

A couple of things I did not see in the review article which I'd like to mention:

1) The price for buying a rival's workers seems to have gone up by a factor of 4. (Typical asking price now 120g vs. 30g.)

2) It seems to me that the AI is greatly improved in its approach to initial expansion. In all the starts I played the AIs developed a stronger core of cities in less time than with 1.29. I can't quantify this unfortunately. But in the starts I tried it seemed quite noticeable. Another way to describe it would be that I think the PTW AI would do noticeably better in its QSC scores than the Civ3 AI
In my testing I was comparing starts to a huge/pangaea HOF attempt I'd recently finished to guess whether the same maximum scores could be acheived. My initial rough guess is that I could get about the same score with PTW against 8 rivals on huge/deity/pangaea as I could get in Civ3 with 15 rivals, the difference coming from the more effective AI expansion.

Am I right to understand that the intent is to, at least for some time, make it possible for each GOTM to be played with either Civ3 or with PTW? I think that will present some challenges:

1. Score comparison. I think that the potential maximum score (milked or early victory) may be noticeably different on any given map with PTW. If I'm right about the AI expansion logic, conquest will probably usually take longer. 100K culture may be possible sooner due to Internet wonder. Depending on map and difficulty, other victories might be harder or easier with PTW. (E.g. at lower difficulties, a faster AI buildup could actually be a boon to the player, the AI providing more support to the player. At higher difficulties the player may have a more difficult challenge. And of course the different units and improvements can have some impact.)
Eventually I think that differences in these areas can be accomodated well in GOTM, thanks to Aeson's score calculator. As experience with PTW develops it should be possible to weight the factors already in his algorithms differently for PTW to take care of this. Until that experience is obtained imbalances seem likely.

2. Timeline comparisons. Early game development could take very different paths as a result of different AI behaviour. This may or may not matter. I guess at worst we'd find people just comparing mostly to others using the version they're using and treating the other spoilers as an interesting comparison. Doesn't seem too bad as a worst case.

3. Bugs/exploits. The glitches you've discovered and written about so far sure do add up. I hope there's a patch very soon to address some of those issues. I'm a bit nervous that with the number of glitches apparent so far, that suggests a real possibility of exploits yet to be discovered. One suggestion I'd make in this regard: There are debatable exploits and then again there are exploits which almost no one would defend. In the past I think they've all been treated about the same, i.e. until an exploit is explicitly banned it is accepted in submissions - the first person to find and use it tends to be allowed. I'm in favor of the GOTM staff reserving the right to disallow a submission which uses a previously unknown exploit if in their opinion it is over the top. People could always write in advance to ask permission if they find something new.

Finally a small detail: the PTW review in the writeup section for Ansar Warrior doesn't mention what seems to me their most important characteristic - 3 movement instead of 2.

Great article Cracker

I haven't seen anywhere that you intend to allow players to choose between PTW and Civ3 to play as SirPleb thinks, but then I have not read that much here in the GOTM forum...

FWIW, I agree with SirPleb that there are enough changes to the game to make score comparisons very difficult. Too many to list all of then here, but they are in my head for future discussions. Two biggies are the increased cost of workers (sometimes a major part in my games), and the changed diplomacy (or it al least seems changed).

Also, I agree about the cost of Gallic Swords being overpriced, although (if it were possible) I think I would go for 45 shields. I have had more than one heated discussion about it

Originally posted by SirPleb
1) The price for buying a rival's workers seems to have gone up by a factor of 4. (Typical asking price now 120g vs. 30g.)

Click to expand...

Yes SirPleb, I omitted that section of the review only because I was still working on some testing issues that have now been resolved. Many of these change issues are not well documented in the software release, so yo uhave to do some focused testing to resolve how the changes work and how they effect game play.

The change to worker value in PTW definitely seems to have been implemented by a hardcoded value multiplier that gets applied to the shield production costs of a worker.

In Civ3v1.29 the AI valued workers at somewhere between 27 and 34 gold units per worker. The AI does not seem to recognize the difference in value between slave workers and workers of its own nationality that work at full speed.

In PTWv1.14 the AI values workers at somewhere between 115 and 130 gold units per worker. This change makes it harder to initially buy away the workers from AI civs and essentially cripple the AI's ability to compete. The AI still does not seem to use its workers well, but at least now the AI does not throw totally throw away this resource.

Through testing, we have developed a method that will reduce the chances that the AI will sell you an undervalued worker in Civ3v1.29. Essentially we can change all future GOTM games so that it will cost between 120 and 139 gold units per worker to purchase any of the AI's early workers. Part of this implementation will make it impossible for the human player to disband any of these purchased or captured workers but they will still perform all other functions identical to the existing slave workers (join cities, build colonies, worker jobs, etc.). The AI will continue to undervalue workers from the human civ and will probably only offer you 27 to 33 gold units per worker for the workers that you have grown and produced yourself.

(in the background there is continuing program bug that relates to worker/citizen nationality that seems to come unstuck when you have workers or settlers produced from cities/towns with mixed nationality. We have not yet been able to test and see if this nationality bug is fixed in PTW).

So, the bottomline is that the cost to purchase slave workers from the AI can and probably will be increased in the Civ3v1.29 games as well as the PTW version of the games beginning with Gotm17-Carthage. (Note that this change is not active in the QSC-C2e & r test games that are currently testing in the SG forum).

Feedback, Comments??

anarres,

I want to respect your opinion that you think the differences between Civ3v1.29 and PTWv1.14 may be too great to facilitate meaningful score comparisons, but then strongly and absolutely disagree with you based on the results of the detailed tests that I have been performing.

Since mid December, I have been focused on peeling back some of the smoke that surrounds the PTW implementation so that we can make an informed implementation of a quality Single Player game experience out of what we have been given to work with.

We need to look at each of the PTW features and understand that well over half of the new features were implemented to support the Multiplayer Game modes targeted at rapid online play for games that will generally last only 4 to 6 hours. When we begin to clearly recognize these features and to set them carefully aside or to quantify how they impact game play, we can find that we have fairly clean solutions that can be implemented to make the Civ3v1.29 games virtually equivalent to the PTW games.

There are several implementation details that cannot be exactly matched, but these occur late in the game and impact gameplay in such a way as to pose little value to the players in the upper thrid to half of the player pool. The impact of these features will make it easier for the players in the lower sections of the player pool to survive in games and post a victorious result even if their score for this result will still be well below the score of the players in the higher ranges of the pool.

cracker,

I didn't mean to imply the task of meaningful score comparison was impossible, just very difficult. As mentioned before, I am very impressed with the amount of work put in here, and I look forward to the results of what you are doing.

I think increasing the cost of workers from the AI in Civ3 versions of GOTM is a very good thing. It is almost exploitative how the AI lets you strip them of workers in the early game, which can put them at a serious disadvantage.

Originally posted by cracker
Through testing, we have developed a method that will reduce the chances that the AI will sell you an undervalued worker in Civ3v1.29. Essentially we can change all future GOTM games so that it will cost between 120 and 139 gold units per worker to purchase any of the AI's early workers. Part of this implementation will make it impossible for the human player to disband any of these purchased or captured workers but they will still perform all other functions identical to the existing slave workers (join cities, build colonies, worker jobs, etc.). The AI will continue to undervalue workers from the human civ and will probably only offer you 27 to 33 gold units per worker for the workers that you have grown and produced yourself.

(in the background there is continuing program bug that relates to worker/citizen nationality that seems to come unstuck when you have workers or settlers produced from cities/towns with mixed nationality. We have not yet been able to test and see if this nationality bug is fixed in PTW).

So, the bottomline is that the cost to purchase slave workers from the AI can and probably will be increased in the Civ3v1.29 games as well as the PTW version of the games beginning with Gotm17-Carthage. (Note that this change is not active in the QSC-C2e & r test games that are currently testing in the SG forum).

Feedback, Comments??

Click to expand...

Excellent! It sure makes sense to increase the worker purchase cost in 1.29 to match PTW. I don't think the AI still only offering the lower value to purchase workers will matter at all.

I'm in awe of the effort your are making to make 1.29 games get a feel comparable to PTW.

I'm curious if you will have two threads for each time period--1 for each version. Will you have a scoring section that lets divides the scores into each version and/or note the version that each game was played in?

Also, late in the game I disband workers and slaves to halve the build cost of hospitals and mass transit that I want to build in 1 turn and then cash rush it the rest of the way. I'm glad I'll be using PTW, because not being able to disband slaves would hurt my strategy. I'm wondering how many 1.29ers do the same. Of course, with the milking strategy becoming invalid it would have much less of an impact.

Like the others, my hat is off to you for the amount of work you've done to make the GOTM playable by both versions which will keep the player pool as large as possible.

Airfields:

LMAO at your review of them. You went even further than I did in noting how much stronger they are then building airports. They would certainly have a big impact on maps that have wide oceans between continents. In GOTM 16 I've built several airports and would have loved to just slay slaves instead to build airfields. For everyone that waits till late in the industrial age to conquer the world with transports and tanks this would make a huge impact. Why build a bunch of transports and ships to defend them when you can build one and fill it with slaves that would make instant airfields on the other continent(s) and allow you to move a large force over on the same turn?

Originally posted by RufRydyr
I'm curious if you will have two threads for each time period--1 for each version. Will you have a scoring section that lets divides the scores into each version and/or note the version that each game was played in?

Click to expand...

Ruf,
We will probably not see two different game discussin threads for the different versions. Based on testing of the issues to date, players should see very little that is different of consequence between the games. One game - One Community of Players - One big Happy Experience.

I do think we will see some minor variations in AI behavior but as these are identified accurately we have some higher level interventions that can be applied to help the AI's in the two software versions play more like there cousins in the other versions.

Originally posted by RufRydyr
Also, late in the game I disband workers and slaves to halve the build cost of hospitals and mass transit that I want to build in 1 turn and then cash rush it the rest of the way. I'm glad I'll be using PTW, because not being able to disband slaves would hurt my strategy. I'm wondering how many 1.29ers do the same. Of course, with the milking strategy becoming invalid it would have much less of an impact.

Click to expand...

Ruf,
I would also say that disbanding workers for shields is in general a waste of their value. Their primray value gain is in cities above size 12 where they are worth 80 to 40 food units per worker. If you dispand them, these food units are wasted. The Civ3 version will still be able to join workers into cities to boost population, the only thing you will give up is the ability to hit the disband button to gain 2 shields for slaves. You will syll be able to disband your own workers for 2 shields if you choose, even though that is not generally an approach that will be recommended.

What about the differences in barbarians?

There is no question that barbs are more of a pain in PTW since you can't depend on them mindlessly attacking the nearest unit. I know GOTM#16 would have been much more difficult under PTW.

Originally posted by LKendter
What about the differences in barbarians?

Click to expand...

Lee,
My testing of barbarians, so far, shows that they may be different but not better. I covered this briefly in the article (hope you've read it). I have not done a compiled code overlay to see exactly what is changed form v1.29 to PTWv1.14.

My major impression of the barbs is that they are more of a nuisance and tend to make a greater effort to go for soft targets such as pillaging, workers, and settlers. This means they can do more economic damage to you and don't just donate free hit point promotions to your units who stand by on defense.

The counter side is that barbarians seem to lose strategic mass and that some random decision making will let your units survive when they otherwise might have died. If you plan to play the games in PTW you will not be able to deal with the barbs in the same way you could in Civ3v1.29 but the question wil lremain to be addressed in the first several games as to whether they are tougher or just different. I think when we get a few of the expert players playing the game our feedback will be far more accurate than it has so far been from the existing PTW player pool.

My initial assessment is that the difference in PTW barb behavior does not impact scoring and game progress for the higher level players but that the mid to lower ranges may experience some added challenges. You will have to hunt down the barbs before they damage you but this emphasizes the attack side of the equation where the player will get an atatck bonus on most difficulties. The impact of the barbs on the AIs will be stronger because the AI's will be less capable of learning and responding. So the two effects will balance each other to keep the game progress equivalent even though the barb events may be different. In the big picture this will probably have no more of an impact than the effects of the Rndom number generator that sometimes sacrifices a tank to a spearman.

One of the major differences in PTW relates to the upgrade paths for the ancient units of warriors-swordsmen and archers-longbows.

This new feature has been implemented to obsolete the ancient units from the late game play to address complaints that somw players had about seeing these units in the age of tanks and bombers.

Another reason for this change seems to be to eliminate much of the strategic risk and strategic decision making in the early game so that human players will be more evenly matched in Multiplayer games.

For the purposes of single player games in the GOTM games, these upgrade paths have been converted to a "game preference" just like culturally linked starting positions, preserve random seed, and allow respawning of AI players.

This new prefference setting is title "Upgrade Ancient Age Offensive Units". This preference setting effects whether the Mace and Rambo units will appear in the games.

To help ease the transition to PTW and further explore how the presence of the "feature"/"Preference" may impact game play, the prefrence for "Upgrade Ancient Age Offensive Units" will be set to OFF during Gotm17-Carthage and during Gotm18-Celts.

If you would like to help with further testing and data analyisis of the impact this preference to potetnially allow "Upgrade Ancient Age Offensive Units" please contact me by email at gotm@civfanatics.net

This is a more general comment about the GOTM special maps.

There was a problem with a couple of the southernmost lands up against the anartic shelf. In 1.29 when an AI builds a city here, you cann't assign workers to cities tiles but must resort to using the governor. I have not tested this in PTW as the autogenerated maps leave open sea next to anartic shelf. If PTW also does not allow assigning citizens to work picked tiles. it might be good to make the land next to the anartic shelf mountains or a type that does not allow city building in the most bottom tile.

== PF

Completed testing of the basic fixes that we will try to implement to make these features balanced and functional parts of the Game.

Outposts - These will be disabled in Gotm17 and Gotm18 by eliminating the ability to build outposts on any terrain type. This makes them a non-ssue for the short term because they will not be in either the Civ3v1.29 version or the PTW version. These features seem to be a leak from the elimination type of MP short game.

Airfields - We will leave these on in the PTW version even though there is no substitute in the Civ3 version yet. In the Out-of-the-Box PTW version, airfields can be built on any terrain type including jungles, forests, and mountains. We will implement a restriction to allow airfields only on the flat, non-soggy terrain types (desert, plains, grassland, tundra). You will be able to clear forests and jungles and then build the airfields. The base number of worker turns required to build an airfield will be increased from 1 (which is instantaneous) up to 288 turns. This means it would take about 48 turns for a single standard worker to complete an airfield in the modern age or 24 turns for the industrious workers. You will be able to stack workers to speed the process and only the first worker will die of a construction accident.

Radar Towers - We will leave these on in the PTW version because they may come too late in the game to be of significant value to the upper end of the player pool and may in fact be too late to help the other players do more than just continue the game. The base worker turns required to build a Radar tower will be increased from 1 up to 48 which will effectively increase the turns required for a single standard worker to complete the task from instantaneous up to 8 turns.

Wow, that was a ton of work cracker, and a very nice analysis, thanks!

One important error I noted however --
You said Berzerks don't upgrade from Archers? They sure do! In the first release they did not but in v1.14f the upgrade path does go from Archers to Berzerks to Rambos. (And you did say your testing was based on 1.14. If you already knew this and were referring only to a gap with no 40 shield unit, that's not clear as written)

For an example, see the mass upgrade of archers to Zerks in:
http://forums.civfanatics.com/showthread.php?postid=660144#post660144

Note that if you *start* or load a game with version before 1.14 then upgrade to v1.14, the rules for that upgrade path are in the save and you cannot do the archer->Zerk upgrade. We were hit with that in RBE5, started just before 1.14. We had to build all our berzerks from scratch.

Thanks,
Charis

Yeah, Charis is correct. They upgrade to zerks in 1.14.

Grey Fox,

The primary point is that the new game preference controlling the multiplayer unit upgrades of warrior-swordsmen and archers-longbows through the additional units of Mace and Rambo is a radically game altering change.

This change increases the reliance on unit upgrades that the AI cannot seem to understand and implement with any sort of planning.

These two changes also reduce most of the need for strategic risk taking in the ancient age. The Longbow is virtually eliminated from the game except in the rare cases when you edit the map to deny Iron.

The upgrade features that would use the Mace and Rambo units have been consolidated into a game preference that can be turned on and off to fit the definition of the scope of the game.

The preference that controls the Mace and Rambo units is turned off for the Gotm17 and Gotm18 games.

Here are a few additional changes that have been implemented into the GOTM to upgrade Civ3v1.29 to play like the functional parts of PTW Single Player:

The improvement of the "Stock Exchange" has been added at a cost of 200 shields and maintenance of 3 gold per turn and increase your uncorrupted tax output by 50%. You must have a Bank in the town and already have discovered the technology of "The Corporation" in order to build a Stock Exchange.

The small wonder of Wall Street has been relocated to require that you have already built five stock exchanges before Wall Street can be enabled.

The great wonder of "The Internet" has been implemented exactly as featured in PTW; costs 1000 shields and coincides with the discovery of Miniaturization. This wonder gives you a Research Lab in every city on the continent and has not been extensively tested for secondary impacts on culture and pollution. You will note that the architect for the Internet Building may have gone to the same school as the architect for the UN.

The improvement of "Civil Defense" has been implemented exactly as featured in PTW; costs 120 shields to build, has a maintenance cost of 1 gold per turn, and requires a barracks prerequisite in that city. This improvement adds a 50% defensive bonus in that city. Civil Defense is available with the discovery of Radio.

PTW implements an improvement called Commercial Docks that cannot be exactly duplicated, but all civilizations in your game have been provided with the ability to build "Ports-of-Entry" as equivalent game features. Ports of Entry function differently from Commercial Docks but can produce an equivalent impact on game play. Ports of Entry are small wonders that increase the uncorrupted tax output in selected coastal cities by 50%. Each civilization only has access to build 4 of these Ports of Entry and they require a progressively increasing number of harbors in your empire. You will need 2, 5, 8, and 12 harbors respectively to enable the Ports of Entry. Technically you could build all 4 of these features in the same coastal city, but this exploit is not allowed for the human player and can be detected in the game replay.