Editor Questions

[Timskorn]

Been working on an updated 1939 campaign and am currently testing it (takes a long time to see if D-day works properly!). So some questions.

1. Will there be a way to test certain events without having to play through an entire game (ie, changes made to D-day code)?

2. Is there anything to override an AI's willingness to operate its units? I thought using the Garrison effect would do it, but it still looks like Russia operates (and moves around) units too much instead of letting them dig in. Or is it possible I am seeing 'extra' units the AI happens to be shifting around (ie, units that are not part of any AI strategy, or part of the Garrison scripts)?

3. Is there a way to tell the AI to use their Siberian units specifically for something? The editor picks units generally with #Size, but not specifically (all tanks, for example). Or am I missing something? I'd like to add a Soviet offensive script utilizing most of their best units at the time.

I'm currently in the middle of testing the latest version. The Soviets are loaded with troops, it's the winter of 1943 and I'm about to take Moscow. I added a couple scripts for the US to land at Algiers, but I'm still waiting to see if it occurs. Also, will the US automatically DOW Vichy France in order to land there?

[Edwin P.]

3. Is there a way to tell the AI to use their Siberian units specifically for something? The editor picks units generally with #Size, but not specifically (all tanks, for example). Or am I missing something? I'd like to add a Soviet offensive script utilizing most of their best units at the time.

Excellent point - I would really like to tell the AI what type of units to use for specific plans. Too often, as is the case with the Siberian transfer the AI will send one tank unit to Moscow and another to Stalingrad. A most ineffective use of amor.

[Timskorn]

Exactly. This is especially true for Russia, which has the daunting task of driving an Axis human player back. With units being dropped here and there, the most they can do is stall. This is, of course, against experienced players. I feel that my Russian AI right now would put up a tough fight against a casual player. They produce more units and managed to keep me from Moscow until Dec. of 1943.

If I can get the North African landings to work, and D-day to be pulled off, I'll be stretched. The US has already sent a bomber and fighter to the UK, on top of the UK's bomber and fighter and have been hammering my production. I've been forced to divert 2 of my 3 AF's to France to stem the bleeding.

Edit: Sweet! I got the US to land in North Africa! They DoW'd Vichy France and landed two Corp and 1 Army to take Algiers in March 1943. Later I turned FoW off to see how D-day was progressing, and I see a problem. The US fails to ship their HQ's to the UK, nor did they send any to support the amphibious landing. Hubert, is this just a bug right now with the US or do I need to include something in the code?

[ April 20, 2006, 04:57 PM: Message edited by: Timskorn ]

[Edwin P.]

Cool.

One item I did today is to refine 4 scripts to counter the Spanish Strategy - i.e. Conquer Spain and then send Italian Fleet through to Atlantic.

Activation_1 Events (x2)

1. Turkish Alarm - 25% Turkey is alarmed - and switches 40% to 100% towards Allies

2. Turkey's Generals Pledge to Defend Turkey - 25% Turkey switchs 40% to 50% towards allies.

Thus you could have a 100% to 150% swing towards the Allies for Turkey.

Diplomatic Plan

3. Added UK Diplomatic Script so that if Turkey is >75% pro-allied UK will purchase Diplomatic Chits to influence Turkey.

Unit Event

4. USA receives 4 fleets from Pacific Fleet if Gibraltar is Axis controlled.

The key is that the events are unpredictable - Turkey will not always move towards allies or join allies, nor will transfer of Pacific Fleet to Atlantic always occur.

Of course, now I have to write Turkish and Soviet Plans to take advantage of the Turks joining the Allies.

I also added a Unit Script for a Turkey HQ unit to appear part of the time.

#UNIT= 0 [10][0][Fevzi]

Mustafa Fevzi was the commander of the Turkish Armed Forces during the WWII period. I rated the Fevzi HQ a 5.

Of course, another option would be Mehmet Kazım Orbay, about 10 years younger and a 4 star General.

[ April 20, 2006, 05:05 PM: Message edited by: Edwin P. ]

[Blashy]

Is all of this Historical though? I'm all for improving the AI but going out of historical context is not really improving it (imho).

Would Turkey really switch this dramatically? Hubert already has Turkey scripted if Axis attack Spain, my guess is he did some research on the subject and came up with an appropriate number.

And should Gibraltar be taken, would USA really have sent more ships to the Atlantic? They had their hands full in the Pacific and the slightest relent on their part could have been disaster because the Japanes were persistant beyond comprehension, give them the slightest breather and it could have doubled their "gutso".

In the end it is your call, but so far you have been doing a good job of staying the historical course.

Once it is all said and done I will gladly test your campaign.

[pzgndr]

Some answers.

1. Yes, try the scripts in one of the later campaigns. Like use 1943 for the D-Day testing. Use 1941 for Siberian Transfer testing. Etc.

2. This is a known AI quirk, and hopefully we'll see reduced op moves in the future. There's no scripting effect to fix; it's a generic AI issue.

3. Unless Hubert adds more detail to the planning scripts, no. Without knowing what the situation may be whenever a script triggers, I'd be reluctant to overscript such detail. There may be other generic AI enhancements that could help improve overall performance.

[Edwin P.]

I did not see a script for Turkey being affected by an Axis attack on Spain. But perhaps I was looking in the wrong place.

Historically speaking, Turkey wanted to stay neutral during WWII and played off the Allies and Axis against each other. They hated the Russians (and though I understand why for Game Play reasons Turkey favors USSR, historically they would have been non-cooperative allies). They were also paranoid about maintaining the Turkish state in light of their recent conflict which saw - Turkey lost most of its provinces, Turks being ejected from Greece (and Greeks being ejected from Turkey). The Turkish Army command didn't much care if Germany attacked minor nations or Russia. They would have reacted to a German attack on Spain - by either moving towards the Axis or Away from them, whichever way they felt would have presevered the Turkish Republic, and during the war they vastly increased the size of their armed forces so as to deter an attack by any side and purchased weapons, at most favorable terms from both Germany and the UK!

As for the event there is only a 25% x 25% that both occur - i.e. 6.25% and I may reduce the event to only one and/or reduce the percentage chance. The key in my view was to give an opportunity for the AI to counter an attack on Spain/Portugal.

Also note, that the Transfer of USA ships from the Pacific only occurs 25% of the time if the Axis takes Gibraltar prior to Pearl Harbor and a peace treaty is signed with Japan. Thus I believe this to be historically valid.

{

#NAME= Peace With Japan

#POPUP= Empire of Japan to Withdraw from China, USA Lifts Trade Embargo

#FLAG= 1

#TYPE= 0

#COUNTRY_ID= 3

#TRIGGER= 25

#DATE= 1941/12/07

#DESTINATION_RESOURCE= 3,25

#DESTINATION_RESOURCE= 6,23

; Is Gibraltar Controlled by Axis?

#CONDITION_POSITION= 55,29 [1,1] [0,9] [1]

#UNIT= 9 [10] [0] [Arizona]

#UNIT= 9 [10] [0] [California]

#UNIT= 11 [10] [0] [Enterprise]

#UNIT= 11 [10] [0] [saratoga]

}

Any futher critiques would be greatly appreciated and are most welcome. Perhaps I should reduce the chance for peace with Japan to 10%?

Why - Japan was reluctant to agree to USA terms as it would mean a loss of face; however, the USA might have reduced its terms (i.e. accepting Japanese control over Manchuko) if it saw the Axis becoming stronger in Germany and wanting to ensure that there was peace in the Pacific.

It would be most useful if events of all types could be linked to an AI level.

[ April 20, 2006, 05:39 PM: Message edited by: Edwin P. ]

[Blashy]

I would even put it at 5% chance, as you state on Japan, it was ALL about honour.

Fair enough on Turkey, seems you have your history intact .

Anytime you need a tester .

[Edwin P.]

I agree on Japan, and here is the revised scripts for Turkey - note that the impact of the Turkish Chief of the General Staff has a larger impact than the attitudes of the politicians as the Turkish Constitution as written gives them the primary responsibility for protecting the Turkish state (and they have used this several times to remove the government).

; Axis Declare War on Spain resulting in an additional

; random 10-20% increase in Turkey activation towards Allies:

{

#NAME= Axis DoW On Spain (Turkey->Allies)

#POPUP= Turkish Leaders In Ankara Alarmed By Axis Assault On Spain

#FLAG= 1

#TYPE= 0

#COUNTRY_ID= 43

#TRIGGER= 25

;10-20% activation increase towards Allies

#ACTIVATION= [10,20] [2]

; Spain declared war on by Axis

#CONDITION= 38 [1] [1]

}

; Axis Declare War on Spain resulting in an additional

; random 20-30% increase in Turkey activation towards Allies:

{

#NAME= Axis DoW On Spain (Turkey->Allies)

#POPUP= General Mustafa Fevzi Çakmak Vows to Defend Turkish Independence

#FLAG= 1

#TYPE= 0

#COUNTRY_ID= 43

#TRIGGER= 25

;20-30% activation increase towards Allies

#ACTIVATION= [20,30] [2]

; Spain declared war on by Axis

#CONDITION= 38 [1] [1]

}

I reduced the % change, to force the allies to spend resource to push them all the way towards joining the Allies.

[ April 20, 2006, 07:52 PM: Message edited by: Edwin P. ]

[Blashy]

Alot of this stuff will IMO also improve the HvH play.

[jon_j_rambo]

Blashy, lets play a game!

ICQ# 234-993-018

[Timskorn]

Thanks for the reply pzgndr. The problem with using another scenario (ie, 1943 for D-day) is it won't take into account everything leading up to it. But in all actuality, my #1 problem has been the US failing to transport HQ's overseas. I can get D-Day to occur but without those HQ's, it'll never amount to anything. Nor any North Africa landings they pull off.

As far as what I'm modding, I'm not trying to deviate too much from the standard 1939 scenario. Trying to fix some AI issues, make the AI tougher and eventually adding some variable scripts so it isn't as predictable. I'm sure Hubert's patch will do most of this, but it's good practice anyway, and I'm sure he doesn't have time to add in all the optional stuff.

[Edwin P.]

Timskorn, could you post the code you used for D-Day. I wrote a transport code to the UK and for some reason my units - including HQs where being off loaded at St. Johns. My guess is that a Length of 1 might have had something to do with it, but I though that once a force was assembled it would turn to a 10 turn transport order.

PS: I used a size of one to insure a constant stream of units to the UK.

[Timskorn]

All I did with this is I delayed the minimum start date by 6 months and changed the min/max units required to start it (#Activate). It was initially 20,20 so the AI would need precisely this amount within a 7 tile radius in the UK for this to be true, so I changed it to 12,20. I'd like to be able to say EXACTLY 4 US troops + HQ though, for example, before D-Day is attempted.

Before I had changed the date by 6 months, the US dropped two Armies and 1 Tank division to take Brest in mid-1943. All the US HQ's were dancing in Boston and New York still. Grr...

I don't think I changed the Length though, which is currently 1.

{

#NAME= USA Build Up Amphibious - Brest (D-Day)

#POPUP=

#FLAG= 1

#TYPE= 1

#COUNTRY_ID= 3

#TRIGGER= 100

#LEVEL= 0

#PLAN_ID= 2

#SIZE= 3

#LENGTH= 1

; Brest

#GOAL_POSITION= 59,18

#DATE= 1943/07/01

#STEAL= 0

; Set friendly positions:

; 1st Line - Manchester

#FRIENDLY_POSITION= 61,13

; Set variable conditions:

; 1st Line - USA politically aligned with Allies and not surrendered

#VARIABLE_CONDITION= 3 [2] [100] [0]

; Set tactical conditions:

; 1st Line - Washington D.C. not tactically threatened (dummy condition)

#TACTICAL_CONDITION= 2,23 [3]

; Set activate position:

; 1st Line - 20 Allied units in England

#ACTIVATE_POSITION= 64,15 [7,7] [12,20] [2]

; Set dummy cancel position (single neutral unit at position 0,0). This is not possible as no

; unit can occupy tile 0,0 so event will not be cancelled due to #CONDITION_POSITION

#CANCEL_POSITION= 0,0 [0,0] [1,1] [0]

}

[Edwin P.]

Many thanks. I may have a way to trick the AI to send a HQ unit. I will let you know if it works as I want to test it first (its a cheat and involves giving the AI a free Str 1 HQ but it should work - i.e. Give AI a Unit HQ Str 1 on Position XX,YY. It reinforces it to 10 and then you send it to Brest on Turn X.

HC - Adding a parameter to limit a plan to include specific unit types would be most useful.

[Timskorn]

You mean give the US, for example, an HQ in the UK a few turns before D-Day?