What is M28AI
M28AI is an adaptive AI for Supreme Commander: Forged Alliance, intended to provide a challenging experience for players, making extensive use of microing skills coupled with strong macro capabilities to defeat opponents.

Competent at both 1v1 and team based games, M28AI adopts an aggressive approach to the game, utilising a wide variety of tactics that will adapt based on what the player does.

A variant of M28AI, M28Easy, disables M28AI's micro capabilities, while there are various configuration options to allow players to further refine how it plays:

• Hybrid AI option - if enabled, you can enable M28AI logic on selected units you control (note that this option should only be enabled if you intend to use the feature, as it will result in all normal M28AI calculations being applied to your own units which may impact on the sim speed with high unit counts)
• AIx overwhelm option - allows you to increase or decrease the AIx modifier over time to make it harder/weaker
• Micro level - In addition to the option to disable all microing with the M28Easy option, a 'M28 time between orders' option allows you to adjust M28's general responsiveness level for most units. By default units consider new orders every second, but this allows you to increase the delay up to 10 seconds (which, for example, will make M28's kiting behaviour much weaker, as well as making M28 run faster)

M28AI is highly compatible with mods and unusual map settings, including (but not limited to) popular mods such as BrewLAN, BlackOPs and Nomads. It can also handle a wide array of unit restrictions in FAF (e.g. it is capable of completing the original supcom campaign which in early missions can involve restrictions such as only building t1 air units).

It also supports campaign missions, both as an AI teammate; or as a replacement for the campaign AI (whether for AI allies such as the rebels in the FA Mission 2 Dawn campaign or for the enemy AI), with further configuration options relating to this to adjust your experience set out in the FAF wiki for co-op missions.

M28AI was originally developed for FAF, and this is recommended for the most competitive experience. However, it also now supports both LOUD and steam (see the installation.txt file for details of how to setup M28AI on LOUD and/or Steam; for FAF setup you can either refer to this same file or the FAF wiki which details how to play with custom AIs).

Why should you use M28AI

• Arguably the most competitive (non-cheating) AI available
• Runs significantly faster than M27AI in larger teamgame scenarios
• Chance to see the AI improve and influence it's development via feedback
• Different playstyle to M27AI and other AIs
• Greater compatibility with mods and unusual game settings
• Compatible with campaign missions (FA, SC and custom campaigns) as an AI teammate (in co-op), and/or as a hostile campaign AI.

Background
M28AI was intended as a long term potential successor to M27AI, involving a complete rewriting of the AI framework to run significantly faster in large teamgame scenarios and provide greater tools for the AI to make use of. As of February 2024 it was able to outperform M27 on most maps and settings, and has improved further since then.

As with M27 it makes extensive use of microing and its focus is in trying to win games, with an estimated global rating of around 900-1200. An M28Easy option is available that disables most micro (along with a setting to effectively set the level of micro used for the normal M28AI), and an AIx overwhelm option is available for people wanting to customise the difficulty more (by increasing / decreasing it over time).

Find out more information

• This forum thread contains summary information of M28AI updates
• A detailed devlog is available where I've noted my progress with M28AI and include far more details on updates made (e.g. the devlog relating to the v1 release of M28AI is more than 100 pages!). Due to file size issues, this is split between
  a devlog covering pre-release to v103 and a devlog covering v104 onwards and a list of possible future changes.

Provide feedback

• Send me a message on discord (maudlin27), either via direct message, the AI development discord (https://discord.gg/jTvFqGKC) or alternatively there's a general AI feedback channel in FAF.
• You can also post in this thread

Where providing suggestions or highlighting bugs it'd be helpful if you would describe the issue, include the replay ID, and the time in the replay where the issue occurs.

Playing offline
If you want to play against M28 offline and can't access the FAF client (e.g. due to DDoS attacks), you can download the mod from Github and ensure the M28AI folder is saved to:
C:\ProgramData\FAForever\user\My Games\Gas Powered Games\Supreme Commander Forged Alliance\mods\M28AI

The mod files can be obtained from Github by downloading a ZIP of the code (and then extracting these):
https://github.com/maudlin27/M28AI

While the game can be run offline either by clicking the offline mode from the FAF client, or running the shortcut C:\ProgramData\FAForever\bin\ForgedAlliance.exe /init init_faf.lua

Trophy holders
Introduced from v54, this recognises the person to complete a challenge with the latest version of M28, going to the first person to beat the highest M28 cheat modifier above a minimum level.

• Radde trophy - awarded for beating a 1.4+ resource (1.0 resource) M28 on a 15km+ mapgen
  
  Current holder - DillyDally (for beating v101 at 1.4 resource on a 15km mapgen)

• Sladow trophy - awarded for beating a 1.5+ AIx (i.e. resource and build rate) M28 on any 5km or larger map (i.e. 256 x 256 or larger)
  
  Current holder - Radde (v81, 1.5 AIx on a 10km map)

Installation notes
To play with M27AI, go to the mods section in the FAF client and search for M27AI. Download the mod, then when hosting a game make sure you have enabled the M27AI mod. This should allow you to then select M27AI as one of the players in the game.

Further details including screenshots on how to setup custom AI are included on the FAF Wiki:
https://wiki.faforever.com/en/Custom-AIs

Summary of how M27 plays compared with default/sorian AI
M27 is intended for players in search of a greater challenge than the default/Sorian AI. Some of its features include:

• Adaptive playstyle, responding to player actions, and actively seeking out vulnerabilities in player forces and building units to exploit these
• Making use of player tactics/unit microing and certain 'try-hard' type approaches, for example kiting with units, hover-bombing, synchronising MMLs to overwhelm TMD, manual reclaiming, ctrl-King T2 mexes to build T3
• Being built with FAF balance/meta in mind

To give a couple of examples of the challenge it provides compared to the default and Sorian, 1 (normal) M27 AI has managed to beat an Adaptive and/or Sorian adaptive AiX with a 2.0 modifier, and also beaten 7 Sorian adaptive (on the same team) on Setons, and is consistently the winner of AI vs AI tournaments (although RNGAI is a close second and worth checking out as well since it offers an alternative play style).

M27AI background and status
Following the 2021 Supreme Computer Cup I was inspired to have a go at developing a basic AI for FAF. As I had zero experience of coding in Lua, I thought at the same time I would write down things as I learnt them to have both a guide for other people interested in getting started in developing an AI, and a dev log for the creation of my AI.

See below for a link to the current AI development guide and devlog (2022-12 update - as people were having issues viewing the original file I've split it out and have separated the main 'devlog' part of M27 development from the more general notes on getting things setup to develop the AI and useful functions)
Devlog extracts: https://docs.google.com/document/d/1uQlEoN-kti7G2MnhwD60uaaHwVevPNYH/edit?usp=sharing&ouid=100973959280546778272&rtpof=true&sd=true
Development guide extracts: https://docs.google.com/document/d/1puzW5hKcOBNE7wu7bXj6qXfTJ60r2f93/edit?usp=sharing&ouid=100973959280546778272&rtpof=true&sd=true

You can also access the mod on Github:
https://github.com/maudlin27/M27AI
While in theory it should be possible to install the mod by copying the M27AI folder (from Github) into the FAF mods directory, some people have reported issues with that. It's recommended that the mod is installed through the FAF client (via the mod vault).

I've copied the current overview from the guide below:

This document is intended to do 3 things at once (meaning it does none of them very well!):

• Act as a guide to people new to FAF AI development, including collating any other useful resources out there on AI development
• Be a sort of dev log of my thought process and development steps for creating my first FAF AI, with extracts of code thrown in
• Form my main ‘todo’ list for further developments of my AI
  If you see any errors, or have any comments or suggestions (in particular any AI developers who read this since you’ll probably know better/easier ways of doing things than the methods I’ve chosen) let me know and I can update the guide where relevant.

If you’re here for useful resources and/or details of how to write AI and don’t care about the dev log aspect then just skip to section 2) Setup and background and section 5) Index of useful functions, and/or browse the index for points that may be of interest.

Latest changes (v70-74 summary)

V70 summary of changes:
Update based primarily on issues spotted from the first weekend of the Rainbow Cup tournament:

• Fixed a bug that caused MAA to stay way too far back from the ACU when there were nearby enemy tanks (i.e. the MAA would run away from the tanks even if they were needed to protect the ACU from gunships/bombers)
• Significantly improved the speed at which the AI will try and get T2 PD if the enemy gets a guncom and attacks its base
• Added new capture logic to capture certain civilian units on the map (e.g. titans on big bang dry)
• 6 other bugfixes, and 12 misc changes, mostly relating to building ground based anti-air, and being a bit more cautious with bombers (including not overbuilding them once a very large bomber force has been assembled).
  v71 - 8 Minor bugfixes and tweaks
  V72 - 4 fixes, including AIX fix
  v73 - M28 compatibiltiy fix and energy storage adjust for FAF balance change
  v74 - Minor bug fixes, primarily relating to the mobile factory and Aeon gun upgrade FAF changes

Note: Most of my focus is currently on M28 - the below links to a separate M28 devlog:
https://forum.faforever.com/topic/5331/m28ai-devlog-pre-release

Areas of focus by version (history):
v1-3: Release
v4 - Intel and threat detection
v5 - Use of ACU and hotfix
v7 - Engineers and building upgrades
v9 - Air
v10 - Anti-Turtle
v11-v12 - Pathfinding and misc changes
v13 - Anti-navy
v14 - ACU and mobile shields
v15-v18 - CPU optimisation and hotfixes
v19 - Rally points and air domination strategy
v20 - Late game
v21-23 - Eco management
v24 - Bombers
v25 - Misc
v26 - Nukes
v27 - T3 Arti
v28 - Game options/compatibility
v29 - Optimisation and TMD
v30 - Czar, Soulripper and Defences
v31 - Bombers (again) and TML
v32-33 - Plateaus
v34-35 - Turtling
v36 - Skirmishers
v37-38 - Game enders
v39 - Quantum optics
v40 - Deceiver
v41 - Hives and shield disruptor
v42 - Misc and multi-TML defence
v43 - Misc and UEF experimentals
v44-46 - Teamwork
v47 - Optimisation
v48 - Personality
v49 - Hotfix
v50-51 - Navy
v52 - Desync hotfix
v53 - Shield boats
v54 - Temporary eco strategy
v55 - Experimental defence
v56 - Stealth boat
v57 - Pathfinding (again!)
v58 - Misc (RC tweaks)
v59 - Gunships
v60 - Naval enhancements
v61 - Stats
v62 - Land spam
v63 - T3 arti prioritisation
v64 - Mass fabs
v65-v66 - Skirmisher and T3 arti tweaks
v67 - Misc
V68-tml enhancements
v69 - Misc
v70 - Guncom defence and capture
v71 - 74 Mini updates - bugfixes

M28AI's release day has arrived!

The main changes made from before were focused on compatibilty, (cpu) optimisation, and increasing its difficulty.

Compatibility
Tested M28AI with the following mods and map settings (usually this just consisted of a single game, so please let me know if you come across any issues):

• Maps with water starts (including all-water maps such as Dreadnought Chasm)
• Survival maps
• AI Wave survival mod
• Flying engineers mod
• Norush game setting
• Low unit caps
• Crazy rush
• Nomads
• BlackOps
• BrewLAN
• Max AiX modifier (IIRC 5.9)

Optimisation
Big thanks to @Jip who helped me achieve a significant reduction in the time taken by M28AI to analyse the map at the start of the game. Previously M28AI was taking more than 2 minutes on 10km maps such as polar depression, with one of the functions being run taking c.102 seconds. With Jip's help this function's time was reduced to roughly 0.2 seconds, with M28AI completing its analysis in most cases in 30s or less (slower CPUs will take longer). The devlog contains more information and profiling for this.

This also allowed me to make use of more accurate water zone generation (without increasing the analysis time).

Jip also helpfully provided code to allow M28AI to get the navmesh label (broadly, whether a particular part of the map is pathable by particular units) for a particular point on the map (which was needed for the optimised code mentioned above to work properly).

Competitveness
Various tweaks and improvements were made to make M28AI provide a harder challenge. As part of this I made sure that M28 was capable of beating RNGAI and/or DilliDalli AI on a number of maps (including those maps I'm aware of where they're capable of beating M27AI some of the time).

Rather than try and achieve >50% win rate on all these maps, I instead just aimed for a single win, and did the AI version of 'save scumming' with replays until M28AI either won the replay, or creating a new game on the map resulted in it winning. Outside of Astro Crater (where M28AI is able to beat M27AI a majority of the time) I didn't test M28AI against M27AI. This is because I realised early in the development of M28AI it was too demotivating to watch it get crushed by M27AI (wheras against RNGAI and DilliDalli AI even if it loses often it's a close fought game, providing me with motivation to try and improve M28AI as opposed to giving up).

As a rough count this process involved 109 tweaks and improvements to M28AI (examples of more impactful changes within this include T2 artillery construction, destroying obsolete radar, and basic logic for building and managing bombers).

Since I've played a grand total of 1 game against M28AI (just to make sure M28 didnt take control of my units when it was my ally against 2 enemy M28AI), I expect it will need a lot more changes to improve its performance against players and become 'cheese resistant'. Almost all games were 1v1s (since they're faster to test) so although the code was all written with the aim of working in team game scenarios, I expect these aspects of it logic will be a bit rougher around the edges.

Please let me know if you have any feedback, suggestions, or encounter bugs.

Progress update: The end of the beginning? or the beginning of the end?

A lot has happened with M28 in just 2 weeks, such that I'm hopeful that it's release date could be in a matter of weeks rather than months. To summarise what's been implemented:

• Naval logic, including building and upgrading naval factories, building naval units, and using naval units
• Novax
• Nukes
• TML
• T2 PD (although in a very limited/niche way comapred to M27)
• Ahwassa
• T3 arti
• Various small changes to improve its performance slightly (e.g. less likely to overflow mass, doing better on maps where it can only reach the enemy with air, logic for engineer order reprioritisation)

M28 now has all the core features that I'd set out to implement ahead of the release (some original points such as T3 bombers were pushed to be post-release but others like T3 arti and T2 PD that were planned for post-release have already been implemented). The main areas it needs work on are compatibility (making sure various niche scenarios work), and a new set of 'competitiveness' related goals I recently added - in short, making sure that while it may not be as good as M27 it should perform similar or better to the other custom AI when released.

The most exciting point for me personally though is the performance gains M28 is still suggesting it will provide compared with M27. When I'd previously ran it on the scenario I used for M27 for benchmarking (4v4 UEF on Africa) it was significantly faster, but that was before air and naval logic were added, and there was always the danger that I could spend half a year building an AI hoping it would run faster, only to find out it was slower. Now though it has almost all core features, and is still much faster (more than double the speed).

• For example, taking M27's 'benchmark' map of a 4v4 UEF mirror on Africa, M27 v67 took 167s to run the first 10m, and encountered stuttering for 20.8 seconds cumulatively.
• By contrast, M28 took 67s, and the cumulative time of freezes was just 0.4 seconds. The only main caveat to this is it freezes the game for a while (probably at least 1 minute) at the start of the game which is much longer than M27, but overall it should still be faster for games running 10+ minutes even after factoring this into account.
• Although not tested (as it's hard to produce an easily comparable scenario), I'm also confident that logic for T3 arti should be far lighter on the CPU than M27's approach which should help late game (as I suspect this was a likely culprit of some of the slowdown, along with M27's preference for swarms of T1 bombers).

Having only started following the FAF forums and news the start of this year I'd not come across the PC election before, but one thing that's surprised me is just how long the process has gone on for. To summarise the timings for the PC election that's almost now finished:

• 6 May 21 - election announced, candidates requested
• 30 May - Signups closed (includes 1 week extension to accommodate further debate following disqualification of certain candidates/clarification of conditions to apply)
• 6 June - voting begins
• 6 July - voting ends

So essentially, 3 months taken up with the PC election.

Regardless of whoever wins the current election, I'd suggest shortening this period significantly, as 1/4 of the PC's time will potentially be in 'election mode'.

I'd therefore suggest allowing 3 weeks for candidates, and then 3 weeks for voting, to halve the time involved overall, for next year's election, while still giving enough of a period to allow most people who might be on holiday at some point in that time to be able to stand and/or vote, without having the election go on for so long.

I'd also be curious if anyone has stats on the timeframe of when votes were cast, as I suspect they will be weighted significantly towards the start of the voting period (which would give further support for there not being a need for a whole month to vote). If this isn't the case and votes tend to be evenly spread over the time period (or even weighted more towards the end of the voting period) then it may still be worth retaining 4 weeks for voting.

Progress update:
Most of my work (when I wasn't updating M27 for the rainbow cup) has been focused on having M28 function on maps with water - for now, without building a single naval or air unit. This took far longer than I expected despite using a very similar approach to how M28 handles land units, and I expect there will still be a number of issues to work through once I have M28 able to build naval units (although most of the hard work should be done now).

• Shields should be built to defend against enemy T3 arti/novax
• SAMs should be built across the map if lacking in air control (which M28 will since it doesnt built any air units!)
• Added logic to consider locations on an island by island basis, so engineers can move from one island to another, and establish a beachhead on an island to try and take control of it
• Put in place framework for 'water zones', which divide the map up into areas similar to what M28 does for land zones, so that naval forces can be handled on a localised manner. These zones are much larger than land zones in size, and for now use simplistic pathing which I expect I'll need to revisit at some point if I can speed up the logic for this.
• Combat units in water are now managed, and will consider fighting other water based threats or to move onshore (if amphibious) to fight for an island
• Engineers can now reclaim and build underwater mexes

I'm also going to start using branches for future updates (so if anyone's running the AI off github, the main release version is likely to be more stable as sometimes when I'm in the middle of a major update Ill post the changes to backup them, which could lead to the logic breaking if trying out the latest branch)

Another thought (sorry for the double post) - could you record your ‘client briefing’, a bit like what Jip did a while ago with some of his tutorials for doing work on the mainFAF code base? That way anyone interested in contributing who misses what you talk through would still have a basic starting point to refer to

v9 Release
Main changes were reworking how Ahwassas target units (including a bugfix that stopped them targeting buildings) and replay based changes (12 tweaks and bugfixes in addition to the Ahwassa changes), including:

• Ahwassas should be more aggressive, particularly in larger numbers, while also being slightly better at avoiding ground AA fire
• MMLs should be slightly more likely to try and overwhelm a TMD defended positon
• ACU should be more aggressive (by factoring in the threat of allied units approaching its current position)
• Fixed some bugs with gunship targeting logic
• Adjusted experimental behaviour slightly so they're less likely to attack-move if an enemy experimental is attacking them just outside their range.

Also implemented logic provided by Jip to make M28 compatible with the upcoming FAF develop release.

Acknowledgements:

• Jip - Providing voice chat and code for the upcoming FAF develop changes (that introduce a new brain class for each AI)
• Azraeelian Angel - Human vs AI replay
• Radde - Human vs AI replays

UEF
Order: The United Earth Federation is the successor to the empire that once controlled the galaxy. The UEF imposes its authority with durable armies designed to control territory.
Cybran
Freedom: The Cybran Nation began as a cyborg conspiracy to liberate workers enslaved by the Earth Empire. Because it began as a resistance movement using hit-and-run tactics, its armies are fast, stealthy, and versatile.
Aeon
Purity: The Aeon Illuminate is a religious order based on devotion and sacrifice. Followers of "The Way" are zealous and single-minded missionaries. Aeon units tend to excel at only one single purpose.
Seraphim
Revenge: The Seraphim are alien invaders from another dimension. Seeking revenge for a bio-weapon attack by the Earth Empire, and fearing humanity's genocidal intentions, Seraphim weapons create overwhelming destruction.

For anyone interested in getting into AI development I've created 'Mini27AI' which is a very basic AI acting as a proof of concept of how to write an AI to carry out some simple tasks outside of the default AI framework, to help as a potential starting point for developers.

Anyone is free to copy/make use of this AI as they wish and build on it for creating their own AI.

At a high level, the AI takes the following approach (which obviously has a number of flaws, as the AI isn't intended to be competitive):

• ACU will build land factories, power and mexes, unless there are nearby enemy units
• Factories will build tanks and LABs
• Tanks and LABs will attack the nearest visible enemy that it can path to (or the enemy base if there are no visible targets)

Mini27AI can be downloaded from the FAF client as a mod, or from github:
https://github.com/maudlin27/Mini27AI

@saver Anything like that I’d need a replay to assess as there will be competing logic:
-Retreat to avoid damage from subs
-Avoid a cycle of retreating and advancing all while taking damage without doing anything
-Avoid making landfall to die to a superior force
-Avoid overreacting to a small naval threat

The logic will likely have room for improvement in specific scenarios

@miki1900 It shouldn’t be what you personally think is best, but what the majority think. Given the Olympics and sporting events more generally, most people are familiar with Bronze < Silver < Gold, and even if they aren’t they would have a decent idea based on market values or other video games (which typically have gold being harder to get/more valuable than silver), far more so than the specific hierarchy of particular US military ranks.

As another example, personally I’d prefer 1 to be weaker than 4 but if a majority thinks that 1 is stronger than 4 then by similar logic it makes sense to have rankings based on the current way around.

v136 Update
18 changes and fixes including:

• Disabled some of M28's kiting logic if it has poor intel
• Improved experimental targeting when faced with an enemy under shields (so they should focus down the shields first)
• Fixed a bug that meant submarines wouldn't run from torpedo bombers
• 6 LOUD specific changes (Aeon land experimental preferences, and various gunship adjustments)

Acknowledgements

• Vortex - several replays
• Azraeel - Replay that highlighted the bug with subs not retreating
• Kemi - Replay showing overbuilding of protector bots in LOUD
• Waffelznoob - Replays of M28 being crushed, and crushing

v134 Update
Another 13 changes, including:

• Greater focus on ecoing, with mexes being more likely to upgrade if held for a long time, while making it more likely mexes are prioritised over factories if stalling mass
• ACU should be more likely to get high value reclaim near it if it's not in immediate danger, even if it would like to get a gun upgrade
• Slightly reduced the amount of engineers to be built in cases where the zone has run out of assignments to give them relatively recently
• 4 LOUD specific changes, including fixing a bug that would result in lots of land factories upgrading at once, and reducing MML production where the enemy has TMD.

Acknowledgements
Vortex - A number of LOUD replays

v135 Update
Re-uploaded as I forgot to turn off some debugging options.

v133 Hotfix - Summary
11 Changes and fixes, including:

• M28 should no longer try and protect a friendly teammate's core base afer the teammate has been killed and their base destroyed
• Changed how M28 prioitises which zones to request units/give them orders, to take into account enemy buildings (in theory making it slightly more likely to attack undefended enemy mexes over defending a friendly location)
• Fixed a bug introduced in the last version for managing gunships that could sometimes break their logic
• 7 LOUD specific changes, including supporting factory enhancements, and updating the ACU threat assessment to take into account the ACU health and to account for the massively higher mass cost given to it in the blueprint (10k, which is almost 200 t1 tanks)

Acknowledgements

• Vortex - several LOUD replays

Assisting costs mass and energy, if you have any. However, if you're stalling mass, unlike any other construction/assistance the rate at which you restore the shield health doesnt decrease. So for example if the only thing you are spending mass on is assisting the shield, and you spend 20 times your mass income assisting, then you are effectively assisting for 5% of the normal mass cost.

As to assistance costs, these vary by shield. Seraphim T3 are the most mass efficient for assisting, while Cybran are the worst. Assuming my figures (from 3+ years ago) aren't out of date, as examples, if considering in terms of how much damage you restore per mass assistance cost, you'd restore 5.79 health per mass for a seraphim T2 shield, or 6.40 for a seraphim T3 shield, or 3.19 from a Cybran T2 shield.

Usually assisting shields is good when defending against artillery since it's a low apm option. In theory the cheapest maintenance option would be shield cycling but that would require a lot of apm. The other option is building new shields so when an existing shield is about to fail you construct the new one.

v132 Update
21 changes, including:

• Gunships should group together better, and take into account how successful they've been in the game with engagements when deciding whether to attack
• Fixed some issues with transports that would e.g. prevent them dropping early when dying to inties
• Reduced tendancy to build more factories late-game
• 10 LOUD related changes, including reviewing/tweaking various LOUD focused PRs from Azraeel (covering building more T3 PD, bombers, ecoing more, and changes to unit prioritisations)

Acknowledgements:

• Azraeel - Various pull requests

v131 Update
28 changes and fixes, including:

• 8 adjustments for when M28 hits the unit cap, primarily focused on when ctrl-king its T2 units is not enough
• Reduced M28's tendancy to overbuild aircraft carriers
• Engineers should no longer suicide into enemies firing at them that it lacks radar coverage of
• 9 LOUD specific changes (including adding a workaround to allow prioritisation of fatboys) and 1 steam change (fixing a bug with T2 kennels not being upgradable)

Acknowledgements:

• MrPhalanx – Mentioning T2 kennels couldn’t be upgraded in steam
• RD – Bitterwind Crag, Elephant Rock and Mellow Shallows replays against v1.29
• Vortex – Replay vs 1.30

Due to the DDOS attacks FAF has suffered, some protections had to be added which had the unfortuate side effect of stopping Kazbek's site/tool from working.

I've not been able to reproduce the Czar issue you mention (even if I activate the total mayhem mod since you mentioned that in your reply), but from the description it sounds like it's broadly working as intended since it's meant to be used to attack ground units rather than air units. So while it might not be optimal (re there being some scenarios where its AA abilities would be good), the difficulty with reproducing it coupled with then the difficulty in having logic that figures out if it wants to have its air to ground attack active vs its AA attack active means it's unlikely to change.

I've fixed the bug with the kennel not upgrading though (will be in the next release).

M28 should build T1 experimental base defences, but I've no plans to have it build other T1 experimental units (e.g. tanks etc.) as M28 is setup on the assumption that engineers can only access experimentals at the T3/experimental stage (not T1-T2) so it'd be non-trivialto add and I'm not clear that it'd have a significant competitive benefit that would justify the extensive recoding required (compared with just getting T2 land or more T1 non-experimental units).