As i promised in patch topic, after finally getting a hands over working test version of Game, working under modified rules, i can proceed to writing some stuff bout it. Version itself is still require some more tests to ensure what everything work just as should, but its fairly safe to confirm that intended result is well achieved there. So, now into some explanations why Retreats in MOO2 are important enough to bother with writing about them, why they are bad the way they been done, and what is possible to do (and whats not) with them.

MOO2 is the game plagued with many gameplay issues. While patches did a good job in fixing "technical" (including interface ones) issues of it recently, "balance" ones are normally out of scope for them; mods do instead. Mods could reach a desire for better gameplay by adjusting many values in their relation: racial picks, weapon stats etc, they could make more strategies viable, even could do good for a problem of production / research imbalance, whare research cost is horribly overpriced in original game, while cost of ships-of-war is horribly underpriced (fun fact is that weak players usually somewhy tend to get it totally reversed way; they want research to cost even more, and seems to be scared to soil enrichment if they seen Command Points indicator goes red even slightly).

There is also another thing for possible change: game rules itself, and one of such change is to be discussed mainly further. Great examples of it is strict_combat_sequencing & interactive_combat_stats of Patch 1.50 — unlike the balance issues of values, ones that affect decisions mainly, they improve game process, namely interactivity. Of course its mostly takes two to know, as AI is very bad in this game. Most of issues fixed discussed here not affect AI, improving AI is a very promising thing of future. But even in game with AI MOO2 show one of its huge problem: low interactivity.

In this stuff its about strategic layer of game. Interactions between Attacking and Defending players lack much actual interaction; Attacking player have yuge proactivity, Defending player lack reactivity to apply. AI incapabilities is also very noticeable on this layer, but we're not about it here. Attacker have all the choice where and when to strike with a packed strike fleet. Defender could only occasionally meet it with own combined fleet, but usually have to take fight with only parts of it. And while those battles itself currently could be improved on tactical level by modding of weapon's values, or applying better battle rules, lack of interaction on strategic layer still remains.
-
One of important factors, contributing to it is (finally) ]Retreats. Currently, in original game, the are basically free (for simplicity i will use more strictly-worded "free" later, tho, technically there is at least one case where they still do affect something). Player can retreat own ship if that ship is under threat, that will save it, just that ship wouldnt further contribute to battle, as if its dead. Deal is that its not. Small AI-related note: ofc AI is horrible with own retreats, it doesnt use ship-level retreats, only fleet-level ones, one or nothing, so its not really exploit main feature of retreat: to save selected ships currently under threat. So, strictly speaking, retreats are also another anti-AI exploit. But mainly its about player-to-player interaction here, as a whole, regardless of artificialness. Retreats are free in a sense they not bear much decision cost. Normally player retreat a ship, that would be dead anyway instead, so wouldnt be able to contribute anymore, so no loss caused. Possible cost there is in players's need to calculate risk; ofc retreating ship, that wouldnt be dead otherwise (say, due to player's miscalculation) its not a free action. Retreats are also free in sense of not affecting a retreated ship in question on a strategic layer: this ship is either will retreat with rest of defeated fleet, or will stay to participate in possible further actions of winner fleet.

Retreat on a ship level effectively put a ship into odd "pocket dimension" at will, where it is essentially safe and sound, with at worst it could have is to proceed to calm further retreat to own planet on a fleet level, even if fleet of that retreated ship is totally routed and decimated by superior enemy, and the ship in question is manage to escape to safety of retreat being barely mobile. Yet if fleet of retreated ship is victorious, that ship will rejoin it as if retreat never happens. Ship-level retreat affect fleet-level retreat only by lack of possible contribution in battle, but if overall fleet is strong enough, it means no consequences to fleet in question at all. Moreother, fleet-level retreats are something only losing side suffer anyway, so in short: Retreats are totally free for a Winner.

As only losers do suffer fleet-level retreats, and as ship-level ones not directly contribute to them, its not so good on strategic level. Above it was mentioned that Attacker already have upper hand versus Defender on strategic layer; Retreats add more to it. As Attacker have no cost in retreating own ships as long as he/she is winning, as they will either join other, remained, part of fleet to assault a planet or merely regroup to continue moving together from this star, Defender, who is usually very limited on resources, actually often cannot interfere with Attacker's fleet at all. Unless Defender can actually kill some incoming ship, most often battle is totally meaningless for him/her, severely damaging interactivity. Even if Defender have means to effectively force to retreat (but not kill) some (but not all) incoming ships, it equals to inability to do anything. As after battle will be eventually won by Attacker, all those ships will act as if none of them actually been in danger enough to retreat. It put some additional threshold on Defender's ability to act: it will require way more resources from him/her to do anything in comparison to Attacker, yet its Defender, who is limited on resources instead. In addition, while its not really symmetrical on strategic layer, its not in tactical one too (note: in Strategic Combat rules ship-level retreats are not exist, adding some twist to all of that stuff): in case Attacker could have ability to force some parts of defending fleet to retreat, while still have no power to actually win a battle (say, Defender do have undefeatable, yet relatively harmless ship), Attacker could use weakening of opposing fleet to inflict some damage to defending Planet prior to actually losing the battle and retreating on a fleet level. Yet for Defender there is no such option. If battle is not won outright, only possible direct kills do matter, and if they prevented by timely retreats, then nothing happens. Defender, who can repel, say, 10 Battleships, but still cant deal with remained couple of Frigate runners are equally (un)effective with one, who cannot deal even with just one autodesigned slow Scout: they lose a battle, and then whole combined Attacker's fleet participate in further assault. Thus Attacker fleet, that just big/powerful enough to overload a Defender ability to kill all its entirety, could win any number of battles versus Defender's fleets just below that threshold; unstoppable and without any losses. With each loss likely reducing Defending player resources even more, as each such loss is likely a Colony.

The power of Beams and other "instahit" weapons are in partial circumvent of that problem. Often its not that player cannot deal something to enemy ship at all; its that player cannot deal it to it timely enough. Slow-reaching weapons let their target to went to safety of their retreat "pocket dimension", while important immobile targets of them stay where they are, unable to evade similar slow-reaching weapons. And again, its Defender, who have such targets, not Attacker. While Beam-likes do change its slightly, real problem is in retreat mechanic.

So what is possible to do with effect, that already hurts bad on a layer where players are unequal, and adding to it by actually providing even more versatility to Attacker in tactical combat layer? Separate ship-level retreats from fleet-level ones. Player can save own ship from danger by retreating it from battle. Thats fine, that ship will now just head on to flight to own's closest colony then (even if it in a same Star system), but not to some odd unreachable safe place located close enough, from that it could then go to bomb a felled planet, or go full fleet retreat, depending on other ship clashes outcome. This will add actual cost to retreat action, so for Defender, who had managed to force some part of enemy fleet to retreat, even despite still lost a battle, it will mean that on next turn Attacker's victorious fleet will actually be weakened, so even may be not able to continue further without reinforcements. Meaning interactivity is raised, even lost battles could matter for Defender now, providing reactivity time to regroup. Without it, Attacker's fleet could continue unopposed, till finally met a opposition above that "able to actually kill ships" threshold, but due to nature of strategic layer movement, it could be merely a guessing game.

So new rule will mean that Retreats are not totally free even for Winner, just like some ships from fleet being lost in process. They could be fully acceptable still, ofc, just not entirely free. Mostly it will make a difference for Attacker (as retreated ships are not participate further, there is, say, possibility that Attacker wouldnt be able to mind-control planet, even despite winning a combat over planet without any ship actually lost in process, because all big enough ships are retreated (instead of being killed otherwise, so its Defender's virtue in effect)), and thats good thing to achieve. Instead of Attacker enjoying whole strategic initiative in evading Defender fleet via jumping any underdefending colony to destroy it in a battle bringing no losses, such battles, while still won by Attacker, could weakening his/her fleet.

New rule also affect Defender, thought way less. Previously, if Defender been Winner, retreats being totally free aswell for him/her, as no fleet-level retreat happens. But even under new rules, individual ship-level retreats, happening during defense of planet (most common case), would proceed "old-style", all or nobody: either fleet-retreat in case of losing a battle, or staying together, as closest colony is the current one, not lost in battle. In case of no colony, or colony being destroyed by Attacker during a battle overall won by Defender, individual ship retreats would direct to next closest colony. Note, its actually already possible in classic game rules during Draw in battle: both sides could do ship-level retreats then (Draw work just the same new way: both fleets retreats). There is also no special need to treat individual retreats of Defender differently while battle is not at his/her planet. While planet indeed could act as some "gathering point" to possibly keep retreated ships together with fleet later in case of successful defense, there is no such place in case of no planet to defend, or if colony is destroyed in process.

To summarize what new retreat rule can do: it could improve interactivity by adding some meaning to many battles, that would eventually "wear off" Attacker fleet, instead of providing no opposition at all. Just as example, beside of Planet attack battle discussed above, even raid of Attacker fleet by smaller Defender fleet orbiting a colony could be of use (in case Defender would got opportunity to attack first in StarSystem selection), if it will return in some Attacker's ship retreats, while Defender can retreat to Planet after to continue battle there, if Attacker will proceed still. Under old rules such raid would be pointless until strategic-layer Defender (being Attacker in tactical layer there) actually had a way to directly kill incoming ships. Also it would make Attacking fleet more unwieldy and heavier to maintain, strictly hurting Blitz, the most viable strategy of the whole Game, one that make other ones redundant (ofc also its about balance issues, discussed above, research/ship cost etc). On a tactical layer it should bring players to develop more advanced maneuvers instead of "ok, ill just retreat this ship then".

To summarize what new retreat rule still have problems with: AI is horrible. Its so horrible, that it just slightly will rise a cost of needed player's fleet to operate. AI can be misguided by decoys, so larger ships still would be more safe than they should be. Strictly speaking it will make Fighters, already the most imbalanced and overpowered weapon in vanilla game, even more imbalanced, as AI tend to see more threat from missile ships instead. But as top-performance fleets of Teleblitz player already operate on a very squandered resources, new rule should push top performance way later, instead of crazy average 80 turns for all the 7 AIs together on Prewarp Huge Impossible (but still nothing close to a time AI could actually defend from such fleet, but its not a problem of rule itself). Also it will make Communications way more viable tech (possible balance issue only, as there is nothing weird to collect various retreating in hyperspace ships together and turn them back to target, if its possible to communicate with them). Also ships, no matter if they retreat individually and/or fleet-level still retreat as a single retreating fleet, with same speed and with Leader on one of those ships still affecting them all (only possible "story" issue only, as its the right way to do in game). Another problem is possible interface issue with large fleets: from now it will be impossible to just retreat some "unneeded" ships to simplify actions performed every turn. Im afraid that list of issues is actually way longer, but im too tired to type moar, and its continuation could appear in further posts if needed.

Currently i cannot announce an estimate time for final tested version to appear. Moreother, im not even really sure it will have any need in actual distribution, regarding the dire state of MOO2 playing overall. Still i decided to put this information publicly.

That mentioned dire state is being actually sad, as even in current way the Game offer great opportunities. Say, many insights on retreats described above, appeared when i played a "One Planet Challenge" on Huge Impossible 7 AI Average. Its not that sometimes called that way somewhere, its not abou just t not building more Colony Ships or Colony Bases, as its already not needed in vanilla game at all if played by proper race (its simply less effective than building warships instead). Its about having only one settled Planet in the whole game, so no way for capture other planets too; for Average it means also scrapping of the provided Colony Ship, making Prewarp easier in comparison (as AI have more economy problems there due to lack of free CS of their). It heavily different in difficulty from the fairly easy "One Colony Challenge", where Outposts are permitted (but still no capture of other colonies). But even latter one is fine enough to show that its not quite right to have a relatively small fleet, that redundantly retreat almost as a whole in every battle, but still move forward fast, covering a great distance. Both of those challenges offer some good learning for regular players, and if "One Planet" could indeed seems too harsh, "One Colony" should be won 100% of times still (assuming proper race), due to immensely unbalanced sheer power of Outpost Ship. After getting enough immersion in them, you'll see Retreat more and more just as merely cheap anti-AI trick, that it currently is, under old vanilla rules. Probly seeing it as too obvious than its maybe is, i forgot to mention that those challenges are ofc intended to be played without any Diplomacy, as it will otherwise change focus from actual playing into trading/begging tech/peace from AI insteаd.

Thank you for this information. I am looking forward to it!