I discovered the mathematically proven best fleet composition

    The result seems interesting and finally i knew there would be some guy doing code stuffs.


    Just a small reminder, while look into your code, the shield in attack should be calculated by int%(i believe?) for example when gs shot ds ds would only lose 1000 shield(rather than 1100 as I think). It doesn't that matter but still need to point that out. ( also I'm not sure but might be the edge case the explode happens only when <70% rather than <= 70%? It might matter when LL shoting heavy ships and one more hit on 70% border might matter the ratios a lot).


    Anyway, why gradient Ascent lul( I personally dont think the problem is near-convex but anyway it might be a good way to approx) (but since you have already use the GA on attackers, why not on defenders as well 0 .0)


    Anyway GJ!


    So much time spent on this, while the answer is so simple. The best fleet composition is the one that can optimally destroy your NEXT target. You adjust your own fleet depending on the target's composition.

    .

    I don't think the problem is very convex either, but with multistart, it seems to work okay. As you can see in the result, the ascent isn't very big in many cases. As for not doing it on the defenders, that's simply because I wanted to stabilize the output. In particular, if the defenders are co-ascending, the two populations can get stuck in a rotation where they could potentially spiral into a suboptimal pit. (Also -- I was lazy to think about how to actually co-evolve them, as there are many different strategies).

    Quote from Kelenken

    So much time spent on this, while the answer is so simple. The best fleet composition is the one that can optimally destroy your NEXT target. You adjust your own fleet depending on the target's composition.

    I don't know about you, but I can't adjust my fleet composition on a whim -- doing so in any significant way would require growing the total fleet by at least 10% to be significant, which would be an enormous amount of resources, hardly what you could do on a daily basis. The only thing you can do is select a subset of your fleet, which DonutZ and I discussed on the last page. Plus, hunting a single target is not always profitable, especially when they're on to you. Hence, playing with limited knowledge requires making a choice on how to grow your fleet in the very long run, for which this data can be valuable.

    Very thoughtful analysis, even if as others point out, it strays somewhat from the established meta of ogame. But that's not necessarily a bad thing. It does affect what you are optimizing for, however.


    My thought from fleeting is that in reality, as you pointed out, ogame is a big-fish-eat-small-fish game, but also that attackers have a strong advantage: they can decide when and what to attack with, whereas defenders are subject to constraints in having a fixed fleet composition (more on this later), ACS defenders may be on fleetsave, etc. - and attackers can also decide not to attack in situations where fleets may have equal value, but unprofitable for the attacker's given fleet composition (or cost to launch). There's many targets to pick, and it is generally inadvisable to go against, say, that outsized ripfleet.


    Then there is the problem that ogame is, ultimately, a pvp game. This means that fleet compositions of your defenders generally fall into two categories: (1) the newbies who build fleet with total disregard to fleet composition, in this case they're likely to be crashed while small and their presence is negligible (almost guaranteed to be crashable with pure BC fleet), and (2) the fleets that follow meta and have combinations of the classic BC:BS 1:1, good fooder, dessies to counter pure BC fleets, etc. Then your problem subspace becomes just that - for bigger opponents you just have to deal with classic comps, and as I was once mentored, BC are the pinnacle of fleet destruction (on equal footing) with few exceptions late game where dessies and rips start to emerge on a large scale.


    Other posters above have harped on other general ogame concerns as well. The cost to fleetsave a fleet is not negligible (which is partly why a destroyer-heavy fleet is not as popular even if your opponents can be pure BC fleets), the cost to launch **and speed** is not negligible, because if your target comes online you'll waste deut, time, and subject to ninja, and there's no 'spare resources' concept - there are many ways to spend resources than fleet, and a pure-fleeter player with a bad economy will be impossible to recover from a crash.


    I really appreciate your code however and will play with it to see, and I'm happy you're getting good results in your uni.

    Im sorry for not reading everything, so if my points have been brought up already, just disregard my message.


    First off - light fighters is the most overrated fleet unit in the game currently. Everyone feels like they need to have big numbers of light fighters - and they sure are good when you manage to get fleet on fleet hits (assuming your fleet isnt way bigger than theirs), but reality is that 99/100 hits you do (assuming youre looking for the "correct" targets), you wont need light fighters.
    Majority of your hits (probably like 7/10 hits) should be battlecruiser/battleship only. Then you have 3/10 of your hits including cruisers and pathfinders.
    If you bash def then obviously having the destroyer/bomber/reaper combo is a valid option, but they are expensive to fly with, expensive to fleetsave with, not to mention slow.


    So when it comes to raiding you first and foremost need to figure out your target more so than the optimal fleet composition.
    Say your target has 100k cruisers and almost nothing else - then it doesnt matter much that you have 25 lf to 1 bc. Would be much better to have 0 lf and double the bc numbers.

    So for anyone reading up on all of these suggestions in this thread - take it all with a grain of salt. You need to figure out what players you can "hunt" and figure out what the optimal fleet composition is to take those down.
    In majority of the situations it require a fast fleet and as such light fighter/bomber/destroyer/reaper isn't a reliable option.

    Hope I didnt repeat what too many others have already said.


    EDIT: and as someone else already said - if you are discoverer, then building a solid amount of pathfinders is rarely a bad idea. They dont do terribly in fight and they have amazing capacity to grab the ressources your targets will have.

    Old school gamer

    Quote from Arcticwarfare

    what's the real world $ value to build a million Dessies. I wanna know how thicc these whales are

    That depends on the pack size of that player. the pack size has many things that can affect it.

    In a perfect world (3:2:1 merch rates) 1million dessies costs 180billion metal (once you merch correct amount into crystal and deut).

    someone with 17bn packs needs to buy 11packs

    someone with 2bn packs needs to buy 90packs


    now ill leave the small task of working out how much a pack is to you/or to the whales in question.

    you calculation is wrong,best fleet composition is 10 crawlers,10 sats,10 death stars and 5 billion death stars. 4:5:1 m c d ratio.Square root of that is basicly 2 LFs.I simmed it,and trust me.

    I have 15 planets, with an average of 41,4 metal mines with slightly below optimal lifeform buildings and techs - my packs are 528m metal (330m points atm. and Im not collector)

    So Id need in the ballpark of 350 metal packs assuming your calculations are correct.

    (just writing this here so people can kinda relate to what a "normal" player with a "normal" sized account can get from their packs)

    Old school gamer

    So after going over the code, it doesn't appear that it takes debris field percentages into account. Wouldn't that have the potential to significantly alter the profit calculations? Not to mention that it doesn't consider that destroyed defenses (generally) aren't part of debris fields. I'm aware these are server-specific settings, but I'd expect taking DF % into account would make a significant difference, since it would make it more important to conserve your own ships. Unfortunately, I don't actually know C++, so I'm not sure I can make the changes myself.

    Thanks for the kind comments and for the thoughtful analysis! I think you're spot on with your observations about the meta and how things work in practice. Something that would be interesting is manually sticking all the "classic" comps for the big fish in (2) into the canonical defenders, and optimizing against that. Depending on your goals, changing the equality of values of the two fleets may be warranted. I have no idea what the output is and it's possible that the top fleet will have mediocre performance (that is, a profit hovering around 0).

    Quote from mh227

    So after going over the code, it doesn't appear that it takes debris field percentages into account. Wouldn't that have the potential to significantly alter the profit calculations? Not to mention that it doesn't consider that destroyed defenses (generally) aren't part of debris fields. I'm aware these are server-specific settings, but I'd expect taking DF % into account would make a significant difference, since it would make it more important to conserve your own ships. Unfortunately, I don't actually know C++, so I'm not sure I can make the changes myself.

    The debris field percentages actually don't matter for purposes of simulation, because all the comparisons of fleet values are relative. For example, if one battle results in 100,000 profit, and one results in 200,000 profit, then the amount of profit in the second case is 2x, no matter the DF percentage, since both battles are subject to the same percentage. (If the DF percentage is 50%, then it's still 2x, if it's 30%, it's still 2x). Only the relative profits matter for purposes of ranking. Also remember that the fleet sizes are subject to the same invariant. If we double both fleet sizes, they will roughly still be 2x apart. Since everything is measured in the same "unit", only relative comparisons matter, but of course, you can multiply the profit with whatever percentage you want if you want a representative profit, whichever value function you picked.

    foxove5154 but deut to debris does indeed matter.

    and to say % dont matter, isnt really something I can comprehend. Sure in a perfect world with the perfect setup, but OGame isnt perfect. Most hits wont be perfect - and if you told me I had 1 bn metal to convert to whatever I wanted and make any composition of ships - then Id likely build 3 bc for every rc and thats the only thing id build.
    Depending on the uni size - but i wouldnt stop building bc until I could make small hits only with them.

    Once i had done that I would start building pathfinders, cruisers and maybe some bs.

    Once I had a decent size fleet of those, then (and only then!) id ‘waste’ ressources on lf and MAYBE heavy ships.

    Id likely never ever build heavy fighters.


    oh and id likely build rips before building lf/heavies


    again - it depends on the speed of the uni. The faster the uni - the more likely I am to build light fighters and heavies. In a 1x uni, i might even prioritise cruisers over bc. Unsure, but its close atleast

    Old school gamer

    Quote from Krumme

    but deut to debris does indeed matter.

    Yes, apologies, that's true. I don't think it makes a huge difference in simulations, but you can modify the code and try. Excluding deuterium from profit calculations might favor heavy ship compositions slightly more.

    Quote

    and to say % dont matter, isnt really something I can comprehend. Sure in a perfect world with the perfect setup, but OGame isnt perfect.

    It doesn't have to do with perfection, but with math and linearity. The unit you see in the "Profit" column is arbitrary -- it's technically kilo-max-max-equivalent-resources (call it kilowobbs, or something), but the point is, the ranking doesn't care what unit it is in, as long as they can be compared to numbers in the same unit (kilowobbs). If you want to multiply the "Profit" column with 0.5 for a universe with 50% debris field ratio and get kilohalfwobbs, you can do that just fine, but the ranking stays exactly the same. The question of which fleet composition results in the max profit is independent of the amount of resources that go to the DF. If you're still confused, read lines 126 - 146 in gradient.h.

    Quote


    light fighters is the most overrated fleet unit in the game currently

    Temper your judgement with the fact that light fighters spam the infamous 70% explosion rule. From the wiki:

    Code
    1. It is worth to mention that the probability to explode is calculated at each shot (as long as the ship is under 70% hull), and not only once at the end of the round.
    3. This is the reason why a lot of small shots received under 70% create a very high chance of explosion even if the hull remains at a high percentage.