Funny AI Tech Investment (Full Version)

All Forums >> [New Releases from Matrix Games] >> Strategic Command Series >> Strategic Command WWII War in Europe >> Tech Support



Message


Taxman66 -> Funny AI Tech Investment (8/30/2019 2:42:21 PM)

v16.03
1939 Storm.
English
Allied vs. Axis AI on Veteran.

Germany made a 73 MPP Tech investment.




[image]local://upfiles/28692/88D27E38FFED464C978C4473CA4BF2AB.jpg[/image]




Taxman66 -> RE: Funny AI Tech Investment (8/30/2019 6:04:45 PM)

Another example, same game

[image]local://upfiles/28692/2B87A676B1AB414EB373999BBA3D4872.jpg[/image]




Taxman66 -> RE: Funny AI Tech Investment (8/30/2019 11:54:50 PM)

Same game
v16.03 Base 1939 Storm, English, no Mods.

More Screwy-ness:

Now it appears AI investment isn't showing up on the report.

Note the German Tank at level 2, and the only investment is some incorrect value on a prior turn.

[image]local://upfiles/28692/A93366666CD24A089F0EB1DC68F14FB9.jpg[/image]




Hubert Cater -> RE: Funny AI Tech Investment (9/11/2019 4:28:01 PM)

Thanks, I've tracked this down as well, similarly related to the issues with the AI Diplomacy investment reporting, and will also correct it for the next update.




Taxman66 -> RE: Funny AI Tech Investment (9/11/2019 4:30:02 PM)

Is this a Reports Display issue, or is the AI getting these investments on the cheap?




Hubert Cater -> RE: Funny AI Tech Investment (9/12/2019 1:29:22 PM)

A bit of both. In order to help the AI and for it remain focused, there are times in the scripts where we have the AI run a single research event for investments where it will skip the check for sufficient funds. It may have the funds, but if it doesn't it will simply spend what it has in order to purchase an investment chit.

Generally we have the AI randomly select from a category of research items as the war progresses and for each country and so on. It has priorities of course, but this also then ensures it is a bit random from game to game. However, there are some key items it needs to focus on at certain times. For example, Heavy Tanks for the Germans, Industrial for the US, and so on.

It's not the most ideal, but if we want to ensure a good gaming experience where the AI is performing the right research investments, has a bit of randomness as well, and does everything else it needs to do such as purchases, upgrades, reinforcements etc., this was the best way (as of right now) to get what we wanted from the AI as typically by the time it comes to research or diplomatic investments at the end of its turn there is no MPPs left for it to do so.

The fix on my end is two fold, one is to actually improve some of the code to have it do a better job overall and rely less on this above described mechanism, and for it to report the correct spending amounts so you will no longer see this anomaly in the graphs.




Taxman66 -> RE: Funny AI Tech Investment (9/12/2019 3:56:40 PM)

Instead of having (some) random Tech choices how about a random choice of path at the beginning of the game. Obviously choices could be weighted too.

For example:
At Game Start Italy chooses one of the following Tech groups and prioritizes all of the associated techs (instead of random selection):
1. Naval (elects to go down improving naval Techs, also tie in or increase chances of BB to CV conversion DE).
2. Air (elects to go improving Air Techs)
3. Tanks (includes Armored Warfare and then builds Tanks and Mech units)
4. Infantry and more General purpose (ART, AA, Industry etc).

The concept here is to improve Tech choice Synergy.
Sometimes at the end of AI games I'm looking at opponent Tech choices that don't make much sense (like Italy improving Advanced Tanks but didn't take Armored Warfare).

----
The same concept could be applied to AI grand strategic decisions if someone ever had the time to develop additional AI choices (like AI Germany going for a Sea Lion).




Hubert Cater -> RE: Funny AI Tech Investment (9/12/2019 8:36:04 PM)

I see what you are after and you are not wrong in that it would be an improvement if worked out accordingly, it is just always a matter of priorities and finding the time and so on. In fact what you've suggested has been on my list for quite some time, i.e. link research, purchases and offensive strategies more closely together.

While I mention it being a bit random, currently it is more like a singular tech strategy and grouping such as let's say Germany should prioritize and invest in Infantry Warfare, Advanced Tanks, Advanced Air and Advanced Subs. The randomness is that in some games the distribution of chits will vary and sometimes it will be heavier in Advanced Tanks and Advanced Air, or in another game heavier in Infantry Warfare and Advanced Subs and so on. Or it might spread them out more evenly between all desired categories. It could even as you've noticed invest in Advanced Tanks and miss the investment in Armored Warfare even though it is scripted to at least consider it.

Even with a more precise design in the AI research scripts it won't always be exact and you might get a bit of randomness within the grouping and priorities just the same, unless of course I force the AI into exact choices which is something I could consider as well.





Page: [1]

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
2.859375