Reclaim Tech Question for Everyone

[Edwin P.]

What rules should the AI follow for deciding whether or not to reclaim its research chits?

Russia has at least 3 chits. This could provide it with an extra 3 corps or an army and a corps if they were reclaimed. The US has at least 2 chits. Enough to finance the mobilization of 2 corps.

I would like to see what your thoughts are for the rules that would govern such an action for; Russia, UK, US, Germany, and Italy. Do you think that the AI should never consider such an action?

Example:

Russia should reclaim its chits when ?

US should reclaim its chits when ?

[ May 15, 2003, 08:25 PM: Message edited by: Edwin P. ]

[Iron Ranger]

Are you polling for SC2?

The AI should never reclaim unless it reachs L5 in the tech. It already does a poor job in distrubution of Tech and the game is too complex for its limited fuzz logic.

Example: One game the AI as Axis was lucky or put alot of money in Subs - reaching L4 by 1941 but it only purchased one sub. Mean while its L0 tanks (alot of these) are pounding themselves to death on my L3 AT Russian infantry.

[Edwin P.]

I was thinking that if Great Britain faced a Sea Lion and there were 3+ Axis units in Great Britian that it should reclaim any tech chits so that it could better defend the home islands. What good is advanced tech if there is not anyone around to use it.

Similarily, at a simple level, if Russia lost Rostov or Stalingrad or Moscow it should reclaim its tech chits to better defend the southern oil fields and cities of the Caucaus mountain region or the Urals.

Or if Russia has conquered Warsaw then the Germans should reclaim their chits to buy more units.

This is a strategy that I have seen many human players follow and I was wondering what logic rules, "furry" or otherwise, should guide the AI in making a similar decision.

[ May 15, 2003, 11:01 PM: Message edited by: Edwin P. ]

[JerseyJohn]

Fully agreed. But would the AI buy new research after the crisis is over?

[Edwin P.]

I don't know? I assume that at the point that the AI reclaims its chits that its almost a lost cause, except in the case of the UK.

[santabear]

Here are some simple thoughts for further refinement:

AI should reclaim chits if a "supply center" falls (Moscow, Stalingrad, Sverdlovsk, London, Manchester). In addition:

Russia: AI should reclaim chits 3-4 turns after Siberian army transfer, if German advance continues toward Moscow.

Britain: No additional condiions.

US: Reclaim chits if more than 3 Axis units land in US.

[Edwin P.]

SantaBear

How would you quantify "if a German advance against Moscow continues"?

[pzgndr]

My suggestions have been for increased per-factor unit costs for each tech level advance. This would make the build/reinforcement costs higher for those fancy jets and heavy tanks. If implemented, players would have to consider the cost/benefit tradeoffs for continuing to research to the higher levels. This would make things more interesting than what we have now.

If implemented, the AI would have to consider these things also. Even with the current system, the AI needs some improvement to increase its research efforts regarding both the number of points it buys and where it allocates them. It tends to be much more limited than what human players do. It also needs to learn to cash out when immediate homeland defense requirements outweigh future research goals. These are higher order decisions rather than specific triggers to do something. Hopefully SC2 will address all these concerns.

[santabear]

Bill Macon: Agreed, especially if the higher per-unit cost is implemented. If not, I think the decision to cash out research chits would be triggered by specific events.

EdwinP: I assume that AI now uses some kind of "zone" system with the Siberian army trigger. In other words as the Germans approach Moscow there is an increasing probability of the Siberians arriving (up to some very close distance at which the probability = 100%). Certainly the system can't be as simple as having one line of hexes always trigger the Siberians--we likely would have noticed that by now.

So: I would quantify "continuing the advance on Moscow" as entering the next "zone" of hexes--or coming immediately adjacent to Moscow. In Russia, AI would probably want to err on the side of cashing in too late to save Moscow, since there are the other two "safety valve" cities, and another turn or two could provide useful tech advances.

============================

There should also be a protocol for the order in which chits are cashed in. Expensive units like jets wouldn't be as useful in defending a capital as, say, IT. Here is a suggestion:

First to be cashed in:

Anything naval or air-oriented

Sonar

Gun-laying Radar

Strategic Bombers

Long Range Air

Jet Aircraft

Next would be primairily offensive weapons

Rocket Technology

Heavy Tanks

Last would be the ones most useful for defence

Anti-tank weapons

Industrial Technology

(I think I'm missing one here, but I can't remember which one...I hope you get the idea)

[SeaWolf_48]

GENTALMEN:

Right now there are 11 Technologies and yet you are allowed only ten chits? Is this to say that America, Germany, and England couldn't and didn't research all items? I think not.

Also the AI should just put one chit toward each technology, this would simplify the process.

I agree with Bill if you put a second chit towards a technology it should cost more, say 50% more.

Finally, we need a 12th technology! Many postings about that, but I thought that it could still be mentioned here. (Military Intel/ Spy, Partisan, Underground)

[pzgndr]

I agree with Bill if you put a second chit towards a technology it should cost more, say 50% more.

Seawolf, that's not what I said. Additional costs would apply to the unit builds/reinforcements, not to additional research chits. An increasing cost to research could be considered, but not 50%. Maybe 5% for each additional chit? That 10th chit would then cost about 300 MPP's.

Right now there are 11 Technologies and yet you are allowed only ten chits? Is this to say that America, Germany, and England couldn't and didn't research all items? I think not.

Having more technologies than chits certainly forces you to decide what to include and what to leave out. This makes each game interesting. I'd rather keep that than water down the costs and effects of chits to provide more across-the-board coverage.

[Shaka of Carthage]

I think the AI should reclaim its chits for any nation, when the capitol of that nation is threatened.

Whatever the "fuzzy" conditions that trigger the Siberian reinforcements in Moscow should be used for any nations capitol.

As was pointed out, the AI does not do a good a job as a human in the tech area, but if the nation is about to fall, give it all the MPPs that it can find from the chits. I don't believe that should be expanded into disbanding other units.

JerseyJohn made an excellent point, if the AI does cash in its chits, what will happen after the crisis is over? Whatever logic it uses to determine when to invest in tech should be triggerred after the crisis.

Since we have seemed to stray off-topic:

increased per-factor unit costs for each tech level For some reason, this sounds like an engineer wrote it. But I like it. Especially if the idea about consuming each chit after each tech level is achieved is adopted as well.

(Now I know why I liked this idea so much. Its one of the things I was trying to accomplish with the "enhancment" of Remove Strength point increase from each tech level advance ... edited 05/16)

AI needs some improvement to increase its research efforts regarding both the number of points it buys and where it allocates them

I'm not sure, but I think the AI has a few different "tech research" plans, that it chooses from in each game. That way, it provides a different feel to each game. It would be interesting to see if Mr H has "tweaked" them to follow the optimun AI tech strategies that many people here have posted about.

[ May 16, 2003, 01:08 PM: Message edited by: Shaka of Carthage ]

[Edwin P.]

My thoughts on Reclaim Research Chits AI

0% per turn Novice, 5% per turn Beginner, 25% per turn Intermediate, 90% per turn Expert

Thus the AI reclaim chits sooner at Expert level than at beginner level.

1. The UK AI would reclaim all research chits if London or Manchester was captured by the Axis.

2. The UK AI would reclaim all research chits if 5 or more Axis were in the UK.

3. After 1941 the German AI would reclaim all research chits if Italy has surrendered.

4. After 1941 the German AI would reclaim all research chits if two of 6 minor countries were liberated by the allies: France, Low Countries, Bulgaria, Rumania, Yugoslavia and Hungary.

5. After December 1942 the German AI would reclaim all research chits if more than 15 Russian land units are spotted within 1 hex of the German border or that of its minor country allies or if Germany held 2 or fewer Russian cities.

6. After 1941 the Italian AI would reclaim all research chits if two mainland Italian cities were captured (Venice, Bari, Florence, Salerno) or Germany has surrendered or if 4 of the following countries have been liberated by the Allies: France, Low Countries, Bulgaria, Rumania, Hungry and Yugoslavia.

7. The Russian AI would reclaim all research chits if Russia controlled 9 or fewer Russian cities.

8. The Russian AI would reclaim all research chits if the Germans conquered one of its three capital cities.

9. The US would reclaim its chits if the UK and Russia have surrendered.

[ June 04, 2003, 05:45 PM: Message edited by: Edwin P. ]

[SeaWolf_48]

What is reclaiming a chit anyway? Do you pull down your buildings and sell the steel and bricks to the public and draft the scientist to create Armies and tanks. That's weird! Engineers, scientist, research people would move on to another research project, not become armies. I think that research points should be sold only to reinvest into other research!