@ptt196 The Triple AI has has some tweaks over the past couple of years, but it is really up to the map maker to squeeze the most out of the AI. See here if you fancy improving your favoured map. https://forums.triplea-game.org/topic/3743/how-to-make-a-map-ai-friendly WW2 wise 1941 Global Command Descision will give a good game, as its AI has been tweeked. https://forums.triplea-game.org/topic/3326/1941-global-command-decision-official-thread

Also, you get a NoSuchElementException when doing a battle with the UK-aaGun unit: [image: c80a9ad1-4141-4dac-a091-41bd276796e3-image.jpeg] In this battle, I gave the British infantry, tank and aaGun the ability to retreat on the first round of combat. When choosing to retreat and selecting a territory, it immediately throws NoSuchElementException and the game crashes. I think that is caused because there is a phase where the aaGun shoots at the air units and that comes after the phase where the British decide to retreat. This results in a phase where the aaGun cannot shoot at the air units. I think it's better to move the decision to retreat for the defender after AA fire, according to the North Africa rules. [image: 4d5afbaa-b922-4031-9a53-9961e87b81b3-image.jpeg] In other words, "British withdraw?" should come below "Germans notify AA casualties".

@Kindwind A neural net certainly makes more sense than a LLM. I didn't see anything obviously illegal in the play, but you didn't document the non-combat moves which caused a problem with your previous model As you noted, your neural net is inferior to your algorithmic WeakAI, which is itself comparable to TripleA's Easy AI. However it does attack and would beat TripleA's Do Nothing AI, which is something. (I have seen a GGP (Ludii) that plays Chess so poorly that it has trouble against random play). However, that it can't beat the training AI seems to point to a fundamental weakness with this sort of neural net. Note that AlphaZero played 44,000,000 games of Chess in its training process (600 years on your computer), and Chess is much simpler game computationally.

Notable mention, bot hosts are an exception. They need a manual docker pull & restart to pull in changes. This is largely intentional as I would consider them already as legacy. edit: plus, restarting a live game is incredibly disruptive. I wasn't comfortable having the lobby restart without warning until there was the auto-reconnect capability implemented - let alone interrupting a full blown in-progress game.

@Cernel wasn’t even aware of all of these existing commands i do use the control click for getting an entire unit type. didnt realize at all that right clicking current deselects units selectively. i dont think it’s a great idea to flip the click tho not sure the current thing is broken just a bit convoluted and not well understood probably

@jayc707 Hi JayC The images will be much sharper in game then they appear when posted here as well. The UHD maps use mostly Frostion units also , as well as his Iron War and Age of Tribes. You can add units to your preferred map and customize it. The map making threads have the information on how to do it. Ask there if you are interested in learning how. Donations are always welcome

I like Darcula - except the politics panel - those colours are a pain, indeed.

I suppose it is worth pointing out that a 'revert' in git is just an undo update. The code history in Git is write-once. Git never deletes anything. So, in a revert update, based on the reverted change: any lines that were added are removed, and any removed lines are re-added. So, a revert is a new update, it just an inverted update to cancel out a previous one. Revert PRs are not at all special to Git, it's just another update. Once a revert is merged, a person can revert the revert update, add more changes to it, and then merge all of that as one bundle. So, a flow might be like this (the letters on the left hand side represent the commit ID, AKA the commit SHA): a43ae Some update that introduced a bug 3ads2 Revert: Some update that introduced a bug fae54 Revert: Revert: Some update that introduced a bug + Bug Fix Changes to the code base are "squashed" together. When a person clicks the revert for the revert, that creates a new proposed change. Bug fixes can be added to that proposed change, then it'll all be squashed together as a single commit that is then released. For example, when reverting the revert, this change would be proposed: abc12 Revert: Revert: Some update that introduced a bug In that proposed change, we can add the bug fix, eg: abc12 Revert: Revert: Some update that introduced a bug fea23 Bug fix Then, when merged, that would land on the main branch and in the history as one commit, eg: fae54 Revert: Revert: Some update that introduced a bug + Bug Fix

@Nothingtoseehere You will need to use workarounds here or play at https://www.axisandallies.org/forums/ PBF works fine there, as the forum-software is on a lower (but compatible) version than here.

Everybody feel free to join the discussion here: https://github.com/triplea-game/triplea/discussions/14596

@LaFayette ok so maybe all that wud be needed at this point then since the menu is icons i’d reverse the up arrow that wud make it more obvious that thats where you go to download

@VictoryFirst How bout "Onward Christian Soldiers" https://www.hymnal.net/en/hymn/h/871 Edit Or some sort of drum beat https://audiojungle.net/search/war drums but yea, the air raid kinda harsh. Not to mention it's still used to this day. Don't wanna freak anyone out in a current warzone

@thedog downloaded the 393 and started a game. It's working great and the AI is much faster. Now I have a more challenging game and stronger allies & axis bots.

@johnnycat you raise good questions: Is the answer time and programmers? Or just programmers or just time? Or management ? Does it come down to Money? If the answers are: yes, no, maybe, and only on weekends, how could I and others help? Time and programmers is the biggest need. Every bug fix and issue probably takes 2 to 16 hours to fix. This is the biggest gap. Further, TripleA is a larger code base that is very brittle. Making that codebase be easier to work with would help. Money is useful for keeping the servers running and paying the monthly bills. As much as I would love to be payed for TripleA work, paying any type of reasonable developer wage becomes way too much very quickly. Additional management would be helpful. Getting issues sorted, cleanly stated, boiled down to a bare minimum would all be helpful.

@kindwind It would seem that training against random games would be every inefficient compared to training against the AI that TripleA comes with.TripleA is not Go, random play should be very bad.

@Darmok72 here is a new starter request, it was half done but not completed, it will affect about 1/4 maps, so please attempt this. https://forums.triplea-game.org/topic/2999/2-6-release-getting-close-need-volunteers-to-help-beta-test-2-6/73?_=1762176127175 EDIT: Also copied links from above link https://forums.triplea-game.org/topic/2752/maxbuiltperplayer-error-resets-purchase-ui-was-how-to-limit-the-number-of-units-purchased/11 https://github.com/triplea-game/triplea/issues/9231

@kindwind Efforts You seem to have in mind using a neural net style AI to play TripleA. I suspect that this we actually be quite difficult given the huge number of possible board positions.You might be better off using a very small mod like Minimap even without a large record of games to train it. TripleA's own AI can be used for training.

@ubernaut yeah, redoing MARTI would potentially save us $10/month on costs as well. I think redoing the bots with 'network relay' might be higher up on the list. That will save a ton of code, create an effective infinite number of bots, and save us maybe like $25/month. The 'network relay' will be an update where the bots will just be responsible for only forwarding network packets. Today, a bot is basically an entire player but is made somewhat invisible. We are bounded by the amount of RAM on a host for how many bots we can host. If bots are just sending network packets, they'll take up very little RAM. We'll only be bounded by network traffic. We could theoretically run dozens of bots per host instead just 3 to 5. This should cut our bot host count down from the current 6 or 8 today to maybe 2. I think there might be some quick wins with the server's as well. AT least for me, I might be inclined to do a quick ELO system before redoing MARTI.