Mega New Elk WIP

[RogerCooper]

@black_elk Does TripleA support the possibility of multiple move/combat phases?

[Black_Elk]

@rogercooper I'm not sure, but that could be an interesting idea, like to have a double combat phase, with a specific blitz phase. I remember at one point thinking it would be cool if we could do something like a same time TripleA, sorta how RiskII handled it's same time, but I could never figure out how to set such a thing up.

[wc_sumpton]

@black_elk said in Mega New Elk WIP:

When testing the set ups, thus far I haven't been able to persuade the HardAI or FastAI to blitz at the m3 distance with their tanks. For example, if I blank the Eastern front of USSR units and give Germany a bunch of tanks on their starting factories, they will still only advance at M2, whereas under human player control they can advance another space

By rule, the attacking tank may move through (blitz) an unoccupied enemy controlled space, as long as that space is the first one moved into, the next space move into may be friendly or enemy controlled. And this is how the AI is programmed, if the adjacent territory is enemy controlled it will attempt to move through it and end its movement in the next territory entered. If this first territory entered is allied controlled it will continue it movement until it enters the closet enemy territory, movement allowed, Blitzing is only checked for the first adject territory, so no you will not be able to get the AI to "blitz" 2 territories, or move 1 territory and blitz the second territory.

Cheers...

[wc_sumpton]

@rogercooper said in Mega New Elk WIP:

Does TripleA support the possibility of multiple move/combat phases?

Yes TripleA will support as many Tech/Purchase/Movement/Combat etc.. They even can have the same name as the engine will move through each step as they are listed. The problem comes when you want triggers to fire off of certain steps, then each step should have a unique name.

Cheers...

[wc_sumpton]

@black_elk said in Mega New Elk WIP:

I remember at one point thinking it would be cool if we could do something like a same time TripleA, sorta how RiskII handled it's same time, but I could never figure out how to set such a thing up.

Something like "Movement/Battle check if battle won, if won reset/Movement/Battle etc.. for maybe 10/15 turns. If a battle is lost, set all movement to 0. Not that impossible.

Cheers...

[Black_Elk]

@wc_sumpton Ok cool that's sorta what I was figuring was going on. In that case I think the goal should be just to hit that first line in yellow, which we can achieve with the regular tanks at M2 blitz

For France, I think we can say that there G1 is modelling May/June 1940 till basically the end of the year, since the eastern front would reflect late summer 41 with the German invasion of USSR being compressed into that same opening 'Recap' turn.

With this date in mind for the opening fireworks display, I think we can just assign control of Calais to Germany on the first turn, put 2 tanks or mech there and then HardAI will reliably blitz to take Brittany and Poitou. I can then rearrange the forces in Southern France, so that the dominoes fall sorta how the line would look in most books. With Germany overrunning the North and West and the South being taken the follow up turn. It might make sense to see if we can get a German transport to launch units from their position in Libya to take over most of North Africa on G1, otherwise Brits tend to send their air to prop up Algeria, a bit sooner than we'd want a Torch to really crack off. This may take some tinkering since it's hard to test with edit mode. Usually the computer will fail to utilize it's new transports if I simply edit them in, I have to edit the entries to check, so it's a little hard to model on the fly. Though I think we can get it to work. Say either from sz 96 A or sz 95 A, like with a transport or two to use on G1, and then they might have a bit more flexibility to reinforce North Africa.

Transport behavior is a little different between HardAI and FastAI, but I think I'd use hardAI since it's going to default to them anyway unless the player clicks in.

Since Germany routinely uses it's pair of Strat bombers vs the British fleet, I think it might make sense to give them another in N. Africa, probably a fighter or two on the new AB we gave them as well. I think I will empty the Soviet front line, move some of the German forces more forward there, since we don't have to model the one round delay anymore if I do the whole recap concept for the first turn.

I'll play around with some stuff when I get home in an hour

[wc_sumpton]

@black_elk said in Mega New Elk WIP:

Though I think we can get it to work. Say either from sz 96 A or sz 95 A, like with a transport or two to use on G1, and then they might have a bit more flexibility to reinforce North Africa.

Tried this in the XML, HardAI Germany retreats them to zs 97A. Fast AI goes after Gibraltar with one, the other is used to reinforce the Tunis battle.

@black_elk said in Mega New Elk WIP:

I think we can just assign control of Calais to Germany on the first turn, put 2 tanks or mech there and then HardAI will reliably blitz to take Brittany and Poitou.

These units reinforce the Paris battle. Brittany and Poitou are the only 2 territories left in France not captured/controlled by Germany.

UHD WIP 1940-45 1.38.2.zip
Calais changed to German control w/mech and tank
2 German transports in sz 96A

Cheers...

[Black_Elk]

@wc_sumpton Looks interesting! I will use that as the basis. With Calais under Axis control, this draws an an Axis Air attack pretty reliably vs sz 110 A. That is a pretty significant one since Allies will scramble to defend. If Axis prevail it's harder for UK to tip the scale on Algeria defenses vs G2 press against N. Africa.

If I remove the French starting factory at Normandy and add another German tank/mech pairing in Calais then computer G will go the distance to Brittany.

I also saw that HardAI was more cautious with those new Med transports, although I think this should still work, since it gives them more options in subsequent rounds. I also like that it helps to indicate the presence of Germans in N. Africa here I added a cruiser just so they'd have a little coverage.

Then I started making a number of edits to add in more Axis TUV, and push the Eastern Front a bit so it will sync up more on G1/G2. Mostly I just repositioned Germany into a more forward position.

Since Italy is pretty light on income, I gave them a battleship and factory at Istria so they'd have a springboard facing more on the Balkans, also a few more aircraft, so they're fleet doesn't get dropped too hard.

For Japan I filled out their empty tiles with some light forces and a hub on Bangkok.

Going to test from here and see how it shakes out, since I changed quite a bit there. I think the swing towards Axis should be pretty pronounced, so hopefully evens things out a bit. They're into USSR much earlier now and have more to work with in forward positions, so hopefully an alright pick up.

Here's the edit save I just put together to trial

2025-3-25-UHD-WIP-1940-45.tsvg

[wc_sumpton]

@black_elk

Here you go:
UHD WIP 1940-45 1.38.3.zip
New starting setup by @Black_Elk
All Allied starting income adjusted by 50%

I was thinking about making Gibraltar as 2PUs VC because it controls the entry to the Med.

Cheers...

[Black_Elk]

Sounds good. Makes sense to me

They may need another inf dude or two there as well for fodder, just so it's not too much of a sure shot for the Germans to take it with those minty new transports. Looks pretty cool so far.

I suspect I may have juiced the Axis just a little overboard there, since my last trial they had a pretty good drive going and were in Moscow by round 8 hehe, but we can always dial it back again from there I figure. Still sorta broad strokes painting on this, but it gives a pretty nice push/pull by sides I think.

Since computer won't purchase any intermediate factories I thought to add a couple more. It may also make sense for another +2 spot on the other side of the med like in Damascus or Aleppo maybe in to give Axis a reason to push there.

ps. By adding a French bomber to Algiers, I was able to persuade hardAI Germany to attack it on G1 with 1 of the transports. I think that might be the most reliable way to get the computer to attack a spot. I'm making another round of minor adjustments hehe. Ok so in this edit I used the French bomber to make the take on Algeria more consistent, pretty sure hardAI will go after it every time.

I added another bomber to Corsica, computer G will then attack it with the second transport.

I did the same sort of thing for Japan by adding a target bomber to Philippines, Borneo and Wake etc so Computer Japan would hit they'd hit them more reliably.

Then to get Britain to mop up in East Africa an Italian bomber at Mogadishu. Seemed to work pretty well

Here's the edit

2025-3-26-UHD-WIP-1940-45.tsvg

[wc_sumpton]

@black_elk said in Mega New Elk WIP:

By adding a French bomber to Algiers, I was able to persuade hardAI Germany to attack it on G1 with 1 of the transports.

This seems like a good idea. But what happens when these high value targets are not attached? Should they remain or should they be removed?

mega_new_elk_1940.xml
Damascus and Gibraltar changed to 2PUs and cc
New starting setup @Black_Elk

Cheers...

[Black_Elk]

@wc_sumpton Looks cool!

Yeah, so the bomber idea would be an example of me trying to script the opening attacks via target TUV, and cajoling the computer/human player into making a particular attack consistently. It was actually the opposite of my initially approach which was to add Axis TUV, here I was adding Allied TUV, but doing so in a way that advantages Axis since it encourages the computer to make a better opening move on G1. Better positioning for their combat move at any rate.

In general if the player has a chance to destroy naked or weakly defended aircraft (especially strategic bombers) before those air units have a chance to fly away, they will just always make that attack. It becomes part of the standard opening script. A good recent example of this would be the 1942.2 Tourney Rules (world war II v5 TE used by A&AO as the standard set up now) where the German bomber was moved to Ukraine. That Ukraine attack by USSR vs the German bomber in the first turn is now scripted into the 1942 TE opener, because the bomber is simply too powerful to ignore. The chance to kill a bomber on the ground in the opener is just a solid tactic and pretty much impossible to resist. There are other moves one could make, but that's the best move, and so then it just becomes the default play pattern.

So basically, with that attack vs the naked Algiers bomber, the computer is doing exactly what I would do as a human player, prioritizing killing the enemy aircraft. As a player I might simply attack the bomber on the ground with other aircraft, rather than doing the whole amphib thing. Or say for the bomber in Corsica (since there's no factory on that tile) I might approach the attack in a different way, but the priority would still be to kill the lone French bomber before it can get in a move, or fly off and cause havoc elsewhere.

Here I was simply trying to see how far I could push the computer's behavior, which turns out can be pushed a fair bit hehe.

I don't think I would auto add/remove units from play based on what the computer is doing, as I think that would be somewhat confusing. Though of course we'd have to fine tune if it's an either/or sort of attack. It's still possible for example (since the computer edges it's attack) for a defending bomber to roll a 1 and throw a kink in the plan. Although here I think in the Pacific that could provide some interest in the variability game to game.

TheDog in the other map used high value non combat targets (oil derricks and the like, to achieve something similar) but here I wanted to try the simpler method using standard air units. Your mention of what the FastAI was doing vs Gibraltar game me the idea, since I remembered how much the computer loves to try and kill aircraft on the ground, especially bombers (but also tactical bombers and regular fighters), any time they don't have a lot of fodder cover, computer tries to pick them off before they can fly away. Another relatively high value unit that could be used is the naked AAgun, but I saw that computer was not quite as consistent there. Using bombers they will lock in on those targets like a magnet, so it just seemed like an easy way to get things moving.

Once the computer is making it's standard scripted moves, then the challenge is to find the breakers in a more PvP oriented framework. Because we know the human player will have a much better grasp on how to defeat the opponent than the computer does. For example, human will prioritize capital capture above all other considerations, since it's the most impactful game resolution mechanic. They may ignore everything else, send all tanks to nail Moscow and know that they'll have won, because the opponent loses their purchase, purse gets stolen etc, whereas the computer will fan out and sorta cautiously trade territories instead of going for the killshot. For the most part this is fine, since we don't need to create a situation where the computer is always mopping the floor with the human player, but just to put up a reasonable challenge. We can always use stuff like income modifiers to scale the difficulty similar to bids in PvP, which would probably be my approach here since it's pretty easy. Example might be HardAI 125% resource modifier for Very Hard, or 150% for Iron Man mode or whatever hehe.

ps. here is the pattern for the first round. I'd say the script there is pretty consistent, meaning the computer will at least run it's standard attacks reliably. The results of the individual battles will have some variability sure, but the computer will tend to do the same thing on G1/J1 for it's opening attacks, which I think is desirable. They may non com or edge the attack in different ways, especially HardAI vs FastAI, but even there, it's pretty consistent using this method. Here's an image of the first round HardAI, on Frances first turn...

You can see the modelling for the Fall of France/N. Africa has most of the production being overtaken by Germany on G1, and this is fairly reliable game to game. There is a bit of variation on whether G will attack into Toulouse, or how they will activate the Pro Side neutrals in Finland, but for the most part I think we can control it this way.

Quickie save/screen to show the G1/J1 attack pattern.

2025-3-26-UHD-WIP-1940-45_France_1.tsvg

Goal for Japan was just to get them to use those transports in attacks rather than non coms, to make the standard pickups and to press a bit harder vs the Dutch islands and such. Here they will take Wake and Guam etc. If they capture Wake with a tank, they will often advance vs Hawaii on J2, so it might recommend a second US infantry unit there, but for the most part felt pretty good. In this game Japan faired rather poorly at the Pearl Attack. I added a US battleship, to try and make that fight somewhat more interesting. The Axis control of the Med is a bit more reliable than before, so I think it's just a matter of making sure USA/UK can still get a Torch going.

Here computer UK will pretty consistently invade Italian East Africa with something, which while still a bit of an anachronism, at least reflect that whole idea of the Recap turn, like the last pockets of resistance there. They tend to nail whatever German transport off the coast of Algiers as well, so that seems to work pretty well. Axis have sprawled, so now it's just making sure computer Allies can clap back, so that rounds 2-3 have the right vibe generally.

I think we can tune from here, like trying to get USSR/Japan front to stay stable, since Japan will press harder now. Vladivostok is a bit difficult to manage, since they tend to skirmish, again a bit of an anachronism for the 1940-41, but it's not entirely off base. Khalkhin Gol was 39, so I guess any back and forth can be seen sorta like porous borders, or minor clashes. There is always an option for a NAP, but I think there is something in traditional A&A "what if" that lends itself towards letting the players brawl on that count.

[wc_sumpton]

@black_elk

Hard AI Germany will go after Algiers and Corsica, but Fast AI does not. Fast AI Italy will, as long as Britain does not reinforce Algiers. I've also notice Fast AI British not attacking Mogadishu. Japanese Hard/Fast AI does tend to clean all their 'lonely' bomber targets. This is why I was asking, because of the inconsistent Fast AI play.

But all good.

Cheers...

[wc_sumpton]

@black_elk

When "Tech Development" is not selected, the map does not center on the player's capital (only matters for human player's). This is because the first step, which is used to center the map, is being skipped. This can be corrected with a dummy endTurnNoPU at the beginning.

Thoughts

Cheers...

[Black_Elk]

@wc_sumpton Makes sense. Yeah I was noticing some more inconsistent behavior from FastAI as well, I tried to see if I could find a happy medium there, although it could be a bit swingy. Still pretty fun to see the computer doing a nice sprawl now.

Here was my first trial just HardAI vs itself for 5 rounds. Not too shabby
Axis were definitely on the march for that one so far

2025-3-26-UHD-WIP-1940-45_France_5.tsvg

ps. Here's a Fast AI game. In that one Germans went after Malta rather than the French bombers. Italy took Algiers, and the French bomber at Corsica escaped hehe. They ended up bombing Tripoli too, so some flash there I guess. In that one Brit's didn't handle E. Africa so that one may be an open question too, they may regret not making as many calcs as HardAI lol, though in the Pacific computer Japan still pressed pretty well. They definitely like to keep it moving now which is nice to see.

2025-3-26-UHD-WIP-1940-45_Fast_France_3.tsvg

[wc_sumpton]

@black_elk

Hard AI tends to follow the 'scrip'. While Fast AI tends to be a little more unpredictable. What Germany misses, Italy tries to clear up.

Cheers...

P.S Putting a dummy first step and attaching territory setup removes the "star" problem.

Cheers...

[Black_Elk]

Right on

I was just spending the afternoon watching the HardAI make it's moves hehe.

Here was another interesting game to round 8, where Italy captured the syria VC and Japan grabbed Honolulu. Took Japan about that long to overrun China. USA eventually moved on North Africa but the delay was pretty pronounced since Axis were fierce in the med.

2025-3-26-UHD-WIP-1940-45_France_8.tsvg

I was thinking perhaps USA starting minor factory on Rio de Janeiro might give them a better angle on Africa.

Also, watching Japan go all ham vs Hawaii made me think it might be interesting to see if we can get HardAI Japan to sometimes flex vs Alaska or North America in the midgame rather than just always slogging it out vs the USSR up there.

Ideas for territories we might raise to 2 PU value, I did a pair in the case of Allies so the starting income thing would still divide cleaning, with a mind towards giving the Axis some more targets, or making the payoff somewhat better if doing particular moves, like say a Sea Lion invasion or maybe an invasion of Alaska or South America etc. I also thought to do a 2 PU for a couple of the neutrals since they don't have starting income easier to add there.

Anchorage (USA)
Brittany (France)
Calais (Germany)
Bulgaria (Pro Axis)
Dakkar (France)
Denmark (Germany)
Dovao (USA)
Dt. Guiana (Dutch)
Hungary (Germany)
Morocco (France)
N.Ireland (Britain)
Narvik (Germany)
Toulouse (France)
Wales (Britain)
Yugoslavia (Pro Allies)

Then making Oslo a VC, would give what, like 42 VCs across the board? That's probably pretty solid. Puts a couple in a each major theater of ops for the back and forth, push and pull. Idea being to make North Africa, Balkans, Scandinavia, France etc a somewhat bigger get, and to make an Axis push vs the Americas somewhat more viable should they attempt it, even if that's less likely outside of endgame solos and such.

Ps. Was going to try something like this, 2 inf and the factory minor at Rio, 1 transport 1 destroyer off Brazil. Figured it gives the German sub down there something to do with a bit of a coin toss on whether USA can jump out early or has to rebuild. I expect they'll probably drop a few ships down there more maybe fighters to push up on West Africa. Not sure haven't run it yet to see, but maybe it'll give Allies a bit of a leg up on simulating torch. Last game Allies were a bit on the back foot, but I think it's looking pretty good thus far. I'm not sure what sort of trick we might use to entice Japan more towards dutch harbor, aleutians, alaska etc. since it's a bit further afield. I think a production foothold or VC up there would probably make USA just sorta camp up there and drop into USSR the whole time, as they already seem to want to do that. Right now the balance Japan vs USSR felt pretty stable, so I don't know if it's better to weaken USA position to reinforce with transports, or to give USSR a secondary hub at like Yakutia or Yakutsk to cover the Soviet Far East backfield. If USA covering that area they tend to march south and eventually end up in Manchuria, or at least last couple games. Japan was on China a bit harder, and of course Germany had USSR more on the backfoot since the last iteration. I do kinda like that Axis charge and sprawl though, since I think that would make playing as one of the Allies a bit more entertaining, which would be a like to have the Allies just naturally a bit more behind the 8 ball there. Anyhow this was the one I was looking at

2025-3-26-UHD-WIP-1940-45.tsvg

[wc_sumpton]

@black_elk

Wow, what a list of changes.
UHD WIP 1940-45 1.38.5.zip
EndTurnNoPU step added to the beginning of every player to allow the map to center on players capital even when Tech is not checked
New starting setup by @Black_Elk
The following territories changed to 2PUs
Anchorage (USA)
Brittany (France)
Calais (Germany)
Bulgaria (Pro Axis)
Dakkar Dakar (France)
Denmark (Germany)
Dovao (USA)
Dt. Guiana (Dutch)
Hungary (Germany)
Morocco (France)
N.Ireland (Britain)
Narvik (Germany)
Toulouse (France)
Wales (Britain)
Yugoslavia (Pro Allies)

Oslo changed to vc

(I think that's it)

Cheers...

P.S Start PUs adjusted (I did miss something)

Cheers...

[beelee]

@black_elk

as wc said, it doesn't matter for the computer but It Totally Kicks Ass for the Player having it center on the Capital !!!

with the tech off

[Black_Elk]

Hell yeah! Nice

Looks cool! I'll be playing this one all night to see how it pans out
Great work

Here was my first trial with that new set up HardAI vs itself. Brazil did seem to help the Allies get positioned somewhat better on N. Africa. Took about 4 rounds for them to get enough transports to start dropping. I might give them a few more tanks maybe, just so their initial push is slightly more credible, but otherwise felt pretty good.

In this one Japan came crashing hard vs Anzac and overran them I think in round 7, which was around the time that China collapsed. It almost looked like they were going to take a crack at Alaska, they had some floaters hanging out, but then doubled back at the last minute lol

HardAI Axis just met up at the middle of the board, when they went godzilla mode and cut the USSR in half hehe. Least it seems Axis are back in the running now. If anything might be Allies need the slight boost. I did rather neglect Anzac thus far, since they've been pretty stable. Probably they just need another bomber, or again a tank or two prepositioned somewhere. Overall all though, pretty entertaining for the popcorn into round 10 hehe

2025-3-26-UHD-WIP-1940-45_France_9.tsvg

Here's a second game around the same point HardAI. In that one computer Japan went after India instead of Anzac. Computer G took Gibraltar and Axis were dominant in the North Atlantic after that. Computer Italy held suez and then made a play for the Mid East and Africa. Axis made a breakout to take N. Ireland and started chopping it up. Seems the boost up there was an enticement. Computer Anzac had sorta the reverse turn of fortune compared to the previous game, and made some alright gains while the other Allies were taking the heat. I think Axis were clearly in the lead. Just based on those two outs, probably it'd be team Allies from here, adding TUV for tinkering around the margins for the balance by sides. Probably mostly for Torch and to keep China from getting swept too quickly.

2025-3-27-UHD-WIP-1940-45_G9.tsvg