AI - strategic layer
Sept 18, 2019 12:55:28 GMT -6
tortugapower, garrisonchisholm, and 2 more like this
Post by dorn on Sept 18, 2019 12:55:28 GMT -6
I was thinkig if I should include this in suggestions or bugs as it is not a bug in classical definiton but not WAD too.
There are severals issue with AI on strategic layer which where not present in RTW1 (or they were there but without airpower that was not an issue). I will show suggestions and than at the end why there are needed, what is wrong and example from game with real datas.
These issues are bigger as fleet size decreases and makes game on small fleet size from 30s playable with almost no fun. I would suggest testplayers jwsmith26 , garrisonchisholm playing game with "small" fleet size. A lot of issues start to be much easier visible. There are here even with very large fleet setting as some players have already written about some small things. But there are not easily visible and especially in big picture. But neverthless they decrease competitivness of AI from 30s quite a lot.
ISSUES AND SUGGESTIONS
1. Funds issue - AI investment into "coastal fortifications"
AI invests too many funds into coastal guns and airbases.
Possible solution:
I would suggest adapt AI thinking to limit it % of budget. Or if there is not enough funds AI should think about scrapping them instead of ships.
2. Funds issue - AI investment into land airpower
This means all planes that are in airbases. The issue is that all of aircrafts are only useful if battle occur. But if AI has no ships, not battle can occur so spending 90 % of budget on aicrafts is lost funds. I suggest several things to adress this big issue.
EDIT: This is based on implementation airpower within game and limitation of 1 battle per month to game. But it is total war so airpower is not limited to having own ships. This is something which game cannot show making airpower less important if in that location are not ships of both side. In history there was quite often situation where fleets were attacked by airpower only without any enemy ships near by. The game cannot simulate this.
Possible solution:
a) Make strategic layer to airpower similar to submarines when there will be either airwar with losses on airbases or some damaging and sinking of ships between turns similar to submarines.
b) Make AI invest into land airpower much less outside home areas and limit it to home area in a way to be competitive in ships.
c) Make airwpower vs. naval power scenarios (usually during invasions or protecting convoys)
3. AI scrapping
AI scraps by 2 reasons. But as technological progress from beginnig of 20s are slowig down, ship can be useful even after 20 years without major reift, which was impossible for ships built in first 20 years. So AI should evaluate scrapping more carefully later.
AI usually scraps ships because:
a) ships are too old (usually about 20 years)
b) AI runs out of funds (reported here)
Possible solutions:
ad a) Ideal solution would be that AI analyse players fleet and decide by comparison but it is quite difficult. Easy solution is to make time after AI decided to scrap ship larger ships built after 1920 so later ships service time is longer.
ad b) This is linked to funds so solving issue with funds in points 1. and 2. would probably solve this
4. AI builds wrong ships
a) AI starts building CAs instead of BBs/BCs later in game. In real history it proved right thing however in RTW2 the aicrafts has smaller value than in history. In history aicrafts were able to operate almost in all weather conditions (starting with UK pre WW2, USN during war). But in RTW2 in bad weather or night aicrafts are completely useless. So at this conditions battleships are still queens of the sea in RTW2 more than in history.
b) AI builds quite low number of DDs
c) AI builds too many capital ships
Possible solution:
a) Making possible some limited all weather aicrafts possibility. Or AI should build capital ships even later in game.
b) Make more priority for DDs especially if player has much more DDs
c) Change priority of ship building. In RTW1 it was quite right but it is not in RTW2. In RTW1 AI reacts on player construction program, it does not in RTW2 or does it quite limited.
5. AI builds more than have funds for
It seems there is no check for AI builds. So even if AI is spending more than 100 % in maintenance or close to, AI builds another ships. At the ends I saw AI building capital ships more than 7 years (could be seen in save - German battleship)
Possible solution:
Some easy function of prediction of budget and calculating available funds for construction.
ACTUAL STATUS
To prove all the issues are really there, I will show my save and table from 1955. From save you can easily see than AI has not large losses of ships than me.
From the table you can easily see:
1. UK is spending 120% of income for maintenance of aircrafts and infrastructure. - Wrong and not competitive unable even have enough ships for foreign stations
2. USA is spending 70 % of income for maintenance of aicrafts and infrastructure - Still too much.
3. Japan is doing well spending reasonable amout in aicrafts. - Good
4. Germany is spending 79 % of income for aicrafts maintanance. Additional 49 % for infrastructure. - Terrible and wrong.
5. Italy is spending 94 % of income for aicrafts maintenance. - This means AI has only 2 DD so no battle, all investment in air power is wasted.
6. A-H is spending 41 % of income for infrastructure. - Wrong
7. Comparison to player - I as France at the end. Next row is my standard strategy without experiments I did in last year of 1955. As it is easily seen I spent only less than 20 % for aicrafts and infrastructure. As land aicrafts are linked to position (base) they are quite costly opposite to aicraft carriers. And as aircrafts are quite costly and I cannot see so much advantage having good vs. poor(fair) crews, saving funds is much better. This allow me having much larger fleet making AI not competitive even with much larger budget.
Game2_suggestions.7z (390.7 KB)
EDIT: spelling corrections and some small additions in italic