1. We have added a Gift Upgrades feature that allows you to gift an account upgrade to another member, just in time for the holiday season. You can see the gift option when going to the Account Upgrades screen, or on any user profile screen.
    Dismiss Notice

The Mechanics of Overflow Inflation

Discussion in 'Civ5 - Strategy & Tips' started by Fallen_Kings, Dec 31, 2013.

Thread Status:
Not open for further replies.
  1. Fallen_Kings

    Fallen_Kings Chieftain

    Joined:
    Oct 21, 2013
    Messages:
    97
    Not from Japan.:crazyeye:

    In China. we simply name it 溢出瓶(read:YiChuPing,Overflow beakers)
     
  2. Fallen_Kings

    Fallen_Kings Chieftain

    Joined:
    Oct 21, 2013
    Messages:
    97
    32-bit sighed Integer.
    2^31=2147483648
     
  3. Cromagnus

    Cromagnus Deity

    Joined:
    Sep 11, 2012
    Messages:
    2,272
    It's probably a fixed-point. (IE 214748.3648)

    Just a guess.
     
  4. Cromagnus

    Cromagnus Deity

    Joined:
    Sep 11, 2012
    Messages:
    2,272
    Side note: I was trying someone's salt-heavy start with the Maya, and hit T137 Dynamite *without this exploit*... DAAAMN the Maya are strong. With an insane salt start. ;)

    Anyway, that tells me that t130 Dynamite should be achievable with the exploit. Just food for thought. :)
     
  5. neoaspilet11

    neoaspilet11 Chieftain

    Joined:
    Sep 23, 2013
    Messages:
    82
    Location:
    Hong Kong
    it seems working. played china. immortal, quick pangea. i just bulbed one scientist only for the sailing and my current overflow now is 160K at turn 180. however what to do if i get into the negative overflow? my reseach with rocketry seems to be affected with the negative overflow causing it to be completed by 380+ turns.
     
  6. Browd

    Browd Dilettante Administrator

    Joined:
    Aug 5, 2012
    Messages:
    11,886
    Gender:
    Male
    Location:
    Rural Vermont
    Burn some overflow on an expensive tech and then resume with lower techs.
     
  7. Hammer Rabbi

    Hammer Rabbi Deity GOTM Staff

    Joined:
    Jan 3, 2012
    Messages:
    4,460
    Location:
    USA
    i tried twice with babylon to see how early i could pop the free GS with a minimum of 100bpt like it suggests and both times it ran out of steam after about 10 techs, somewhere in the middle of Ren era techs. this was starting with the bulb about t110-120 with about 110bpt. both times also had a 2nd GS spawn during those turns (had to work Unis to get the earliest 100bpt) so i tried to pop it too to keep it going but not to much success since all the low level techs were already gone. im going to keep trying with different mixes and timings, but even with micro-surges through the techs it feels really imbalancing.

    not paying this much attention to every number in my games i realize ive probably gotten these in the past for up to 2-3 techs at a time and just never realized it (not hovering over the tech to know that it had built up combined with them all being lower techs). those were without bulbing anything, i think just due to the modifier for every other civ already knowing those techs.
     
  8. Cromagnus

    Cromagnus Deity

    Joined:
    Sep 11, 2012
    Messages:
    2,272
    You can maximize the overflow by almost completely researching the early techs and not finishing them, so that when you do research sailing, it's more overflow, which helps with the snowball.

    Also, Liberty would get you 2 GS at the same time... makes a big difference to double your bulb strength. Maya should be able to take it even further. They can achieve 3 GS by beelining Printing Press and using the GE on Pisa. The problem is that you also have to beeline Theology AND Printing Press, so you're researching all of the cheap techs except for the Sailing line. Whereas, if you beeline Astronomy instead, you can blow off tons of cheap techs. So 2 GS with more cheap techs might be better than 3 GS with less cheap techs. Of course if you beeline astronomy and blow everything else off, good luck not getting pwned along the way. Tough not to research Construction on Deity. So, I think ultimately, the best options are Beeline Theology->Construction->Education with the Maya, or Beeline Physics with Babylon, and bulb when you get there. I don't think you save enough turns with 1 GS to justify bulbing Chivalry with Arabia this way. (Because skipping things like Mining/Calendar until t80 is counterproductive at best: slows your growth/economy and probably gets you killed)

    I think Babylon has the potential to get more cheap techs by only taking Writing and otherwise hugging the bottom of the tech tree to get Physics. Especially since you can take Mining/Masonry/Construction along the way. You can skip the horseback riding line, even calendar in the right conditions, not just sailing. On the right map, the rest of the AI will all research Education, giving you tons of free techs when you bulb Physics, probably enough to get super-early artillery. But it's not NEARLY as broken as a later-game exploit.

    Bottom line, the overflow just isn't that dramatic if you do it before the late Renaissance. It might shave a few turns off an early domination run, but you still have to be awesome at domination. Where it really shines is getting you a sub-t200 victory with *virtually no skilled required*. XCOM on t190? Game over in 3 turns even if you suck. t200 science victory even if you suck. :lol:

    For that, you're better off IMHO going with a civ that doesn't bulb a free GS early. You'll get 3 GS naturally beelining science. So the Maya and Babylon don't really help you there. Actually they kind of hurt you by saving a GS that long.
     
  9. erisiamk

    erisiamk Chieftain

    Joined:
    Aug 23, 2013
    Messages:
    83
    I just found out about this accidentally in a multiplayer game... coming out of a disastrous early game war barely alive, i proceeded to expand a little bit after a friendly player dealt with my aggressor. My empire was still Medieval whilst others were Industrial. Nevertheless I managed to pass Scholars in Residence at the World Congress because people were too busy voting over the World Religion, and started researching Sailing as I was landlocked the entire time and had only just acquired a luxury resource across the ocean in my territory.

    Needless to say, everyone was surprised about 30 turns later when the tiny Brazil had managed to overtake the highest Literacy player with a meagre 180 bpt... This bug is insanely powerful and could have won me the game if I didn't concede upon discovering the glitch.
     
  10. Hammer Rabbi

    Hammer Rabbi Deity GOTM Staff

    Joined:
    Jan 3, 2012
    Messages:
    4,460
    Location:
    USA
    @Cromagnus: forgot to say i was doing this on Immortal. my goal was trying to accomplish this with just 1 GS as early as i can. since this is suited for the non-science civs the most i wondered if you could still use Bab/Maya to push an earlier start for the snowball effect. the free GS let me delay all the science techs/buildings until i realized i still need to get to 100bpt early. the problem was the initial amount of beakers didnt multiply enough pre-Ren era. 3 would definitely do it but now im going to mess with what i can get with 2 simultaneous bulbs. with Babylon an early focus on culture should get you a fast enough liberty finish to time with your first naturally generated GS (from oracle, bonus% and unis) to get 3 very early.

    and yeah, maya will be up next in the comparison if i feel like continuing this. it already feels like the time i learned i could type in "godmode" or "clipwalls" in Doom in 1993 so we'll see how much i care to actually continue this experiment.
     
  11. dinanipedro

    dinanipedro highlander

    Joined:
    Jan 8, 2013
    Messages:
    203
    Location:
    an island
    civ 2 had god mode ;)
     
  12. qemist

    qemist Prince

    Joined:
    Jul 16, 2011
    Messages:
    562
    Oh, what happened to the other 4 orders of magnitude? Fixed point I guess. That sucks. Just use double.
     
  13. phillipwyllie

    phillipwyllie Wannabe Deity

    Joined:
    Dec 11, 2003
    Messages:
    1,458
    Location:
    4 Hexes from FoY
    I've found it's best to bulb your GS when you have no tech selected as it seems the increase in overflow is applied the next turn, and you don't want the GS to research the tech instantly. I've also found the the best techs to research are the one at the top of the list the game gives you as that is the one that has the least cost(due to more civs knowing that tech). If you right click the tech you can see the base cost and judge how many civs know that tech. RA's, stealing techs, and finding them in ruins can bugger things up as you wont get your overflow multiplied had you researched those techs normally.
     
  14. civjou

    civjou Chieftain

    Joined:
    Nov 12, 2013
    Messages:
    43
    I observed this in Deity as Babylon, although I started bulbling in Atomic. I had i think 8 scientists? (Babylon...) and kept researching techs 1 at a time while observing the beaker count grow...

    Then it suddenly reverted to normal which is I guess when the integer overflowed. I didn't think much of it and was too tired to investigate and did not suspect a bad bug like this, but now reloading the game I was definitely seeing this!


    Makes me more inquisitive now.


    Abuse of this aside, these various civ variable overflow bugs are getting really tiring (the Ghandi nukes thing is also a manifestation of this bug).

    Someone at Quality Control at Fireaxis should really look into testing this properly.... come on, it's not hard!
     
  15. qemist

    qemist Prince

    Joined:
    Jul 16, 2011
    Messages:
    562
    Wait, Firaxis has quality control??? :confused:
     
  16. ravenmagus

    ravenmagus Warlord

    Joined:
    Dec 28, 2013
    Messages:
    251
    Gender:
    Female
    Location:
    Washington, DC
    You'll notice if the beakers overflow. It'll suddenly apply a very large amount of negative beakers to your next tech, making it take forever to research (I once had Industrialization at 1170 turns to finish). If you run out of techs known by other civilizations, or you're researching high enough techs that the cost is greater than the amount you'd gain, then you'll start actually spending science again.
     
  17. Cromagnus

    Cromagnus Deity

    Joined:
    Sep 11, 2012
    Messages:
    2,272
    I feel the same way. Not sure how much I care to continue the experiment. Unfortunately, the whole thing is tarnishing my civ experience. I may quit for a while.

    Regarding the mechanic, I think there's a theoretical limit on how much you can exploit this, because there are many catch-22's. Your bpt affects the total bonus, but the more bpt you have early, the less techs you don't know. And, the sooner you bulb, the less techs the AI has had time to reseach.. So, basically there's a limit on how far you can bulb without waiting until Scholars In Residence passes. (Which coincidentally give more time for the AI to learn techs)

    You basically can't get an Industrial tech much earlier than t115, you can't get to Renaissance much earlier than t100, and you can't *really* get to Medieval early at all. Maybe a few turns at most, because you're wasting turns researching old techs first after you start the bulb... so that's a limiting factor.

    This doesn't mean you can't get an advantage by doing this early, but it's not a huge advantage.

    The thing is, you can't get RAs until you get Education, and don't complete them until 30 turns later. So while you can theoretically get Scientific Theory on t115, your RAs still don't complete until t130, and besides, you just rush-bought Universities... how are you going to afford to buy public schools? If you have to build them for 15 turns, it's not better than rush-buying them 15 turns later. Another catch-22. Why not wait 15 more turns and get more snowball? See my point?

    There are 3 factors that influence how many "bonus" beakers you can get out of this:

    1) How many techs the AI know that you don't.
    2) Your beakers/turn
    3) Scholars In Residence

    The base overflow, if I understand it right, for techs that all other AIs know, is 1.265x. (1 +0.3*(7/8))

    Scholars In Residence is an additional 1.2x on your total overflow, for a total of 1.4625x.

    For techs that are roughly the same cost as your base beakers/turn, the net result is:

    (Initial bulb beakers)*(1.4625^y) where y is the number of techs.

    Let's say, for the sake of simplicity, that there are 10 techs you don't know, that are all exactly 100 beakers. (So your beakers/turn have no influence on the growth rate other than determining the initial bulb amount)

    Let's also say for the sake of argument that your bulb is worth 800 beakers.

    800*(1.4625^10) = 800*(44.76) roughly, or 35813 beakers.

    Once you run out of techs that the AI already knows, you get zero overflow.

    So, in this scenario, you get 35813 beakers to play with. Of course, not all techs are 100 beakers. So, that makes it harder to calculate, which is why I don't have exact numbers. I might record them the next time I try to I can write out a table.

    The important thing though, is that your total output is proportional to your total input (A), and proportional to the number of techs you don't know. (B)

    It's hard for (A) to be big early and B cannot be big early. Also, it's important to realize how much Scholars in Residence affects the outcome. Without SiR, you get 800*(1.265^10) = 800*(10.49) or 8394 beakers. For 10 techs, you're getting only 1/4th the total bonus beakers without scholars in residence.

    If there were 15 techs at 100 cost that you didn't have, the bonuses would be 239622 beakers and 27192 beakers, or 1/8th the total bonus withour SiR. That's a huge difference.

    In my experimentation of bulbing prior to turn 100, I haven't been able to get my overflow to get higher than 3500 before it dips down, for a grand total benefit of somewhere around 6000 beakers. That's *mostly* due to not having Scholars In Residence. That's a 4x factor, and while I was at 70bpt, that's 70% of the "target" number, so it was a much less significant factor. The other issue is the size of your initial bulb has to be MUCH bigger than the size of your initial techs, in order for it to have time to grow before it starts shrinking.

    You basically can't reach the cap without Scholars In Residence or waiting so late in the game that you might as well have had it.

    So, let's assume that the theoretical earliest Scholars In Residence can pass is somewhere around t125-130. (If an AI researches Printing Press around t100 AND they've met everyone, + the turns before the vote)

    This means that you can't start snowballing the entire tech tree prior to t130 at the absolute earliest, which means you can't complete a space race prior to about t170, even if you somehow had the production/gold to get all those spaceship parts done. Which you won't, because you *won't even have the techs to build factories/hydro dams/etc. until t160ish...*

    If you try to bulb around t110, let's say with 100 beaker/turn since t100, well, that means you managed to get early education and have been researching 800 beakers worth of techs. So, you're not very far behind in tech. Which means you're getting closer to that smaller number, or roughly 30000 total beaker benefit. Which will barely get you to the Modern Era. And at 1 tech/turn, you're getting there on roughly t135. And you can get to the Modern Era on t135 WITHOUT this exploit.

    Also, the catch is that you eventually run out of expensive techs to research without having to dip down again, so that's another difficult thing to manage. You need to time completion of your expensive tech with getting all your GS. If you mistime it you wasted your cheap techs. So, all of this means it's just not *all that* advantageous to go for it much earlier than the time you'd get Scholars in Residence anyway.

    I believe the theoretical limit without Scholars in Residence is achieved thusly:

    Having 2 bulbs instead of 1 makes a huge difference, obviously. Having 3 instead of 2 makes an even BIGGER difference, but to get 3 you either need to go for Pisa, which greatly reduces your unresearched techs, and getting Printing Press by t100 is hard enough as it is... or you need to have not gotten any GS before. Which rules out Babylon.

    If you've already gotten 2 GP (Liberty/Writing) you need 300 pts. Each university earns 6 pts/turn or 10 w/ secularism, so you'll never get a GS fast enough that way. Unfortunately with the Maya, if you took a GE, now it's 400 pts. But the Maya is still your best bet IMHO. But it's tricky. You have to *delay* finishing Liberty and not take a GE or a GS until you've spawned one naturally.

    The optimal way to do it is to somehow get a University up and staffed on t100 without completing Liberty, get the points GS on t115, bulb liberty on t116, and take the t117 Maya Long Count GS, then bulb all three immediately. If you're at 100bpt, this is 2400 beakers, which, if we plug back into that "15 tech" number, gives you 108K beakers. Of course, that number is made up, but you get the idea. That could, in theory, get you to the Atomic Era around t135.

    I think... haven't pulled it off yet. The reason being, it's really hard to avoid researching techs all the way to 117. If you skip the useful ones, there's a cost... like you can't build roads or make tile improvements. If you tech the Astronomy line, unless you're on a Water map or doing SV, that's basically a giant waste of time. Your best bet, as others have said, is to go towards Architecture.. That way you avoid Engineering and Bronzeworking. If you're trying to bulb on t115, you really need all the cheap unresearched techs you can get... Sailing isn't enough.

    Basically, I've found that with an 800beaker attempt, you stop growing when you hit the final medieval techs, so I beeline as soon as I run out of cheaper ones. That gets me one second-tier renaissance tech, IE chemistry. With a 2400 beaker attempt, you theoretically shouldn't stop growing until the second-tier renaissance techs, so if you bulb when you hit them, you can get one (1) Modern era tech before you run out, if you're lucky... I haven't timed it right yet. Maybe you could get to the Atomic Era if you waited to start until t120...

    The point is you can't really start the snowball prior to t130ish and finish out the tree, which creates a limit on turn to victory.

    You can, however, get pretty far up the tree in that case I describe.

    But that's best-case scenario, and again, it's a catch-22. If your tech rate is good, and have a university slot worked early, you aren't far behind in tech. In order to fall behind in tech far enough with science output, you have to wait longer. If you don't have good science output, you won't get as much snowball.

    You also can't really do this effectively with the bottom of the tree, at least not early. You need good tech output when you bulb, and you won't have it if you work the bottom of the three. You'll get more cheap techs, so you'll get more bonus beakers, but your bulb will be worth less.

    So, while you can get, say Rifling or Cannons, on t105, you can't get much farther than that.

    This all makes the more interesting question: Which tech do you beeline? t130 Dynamite? t115 Industrialization? T115 Sci Theory? Which tech is the best use of the exploit?

    Personally, I think Industrialization might be the most interesting use:

    Use GG to get coal, rush-buy factories... you have ideology on t115. But then what? What's the goal? For domination you'd be better off beelining a war tech.... Unless you've been leveling CBs... upgrading all your CBs to Gatlings on t115 is expensive, but if they've got double-tap, those are some nasty units. But, range-2 gatlings lose a lot of the tactical advantage of range-3 XBs... So, yeah, not sure super-early ideology is worth it. However, there are a TON of social policies in there... I'm sure someone could figure out something cool to do with early ideology.

    Then there's Sci Theory. Ok, so you're just doing an optimal science run, which you can do without the exploit. *yawn*

    Dynamite? T130 artillery is fun, but it's not all THAT different from t140 artillery, which you could achieve without the exploit.

    T115 Archeology on the other hand, now that could be interesting for a CV... getting all those archaeologists out earlier could be fun.

    Personally, I'm warming up to the idea of going for earlier tanks. Flight isn't a good plan because you can't upgrade into those units. So you're sitting around building air units or buying A LOT of them.

    However, with tanks that's a lot of expensive upgrading. But, by t140 you could definitely have the cheap upgrades from Honor, and the cheaper upgrades from Ideology when you get Combustion... see, now THAT is a fun idea. :D

    Anyway, this is turning into a rant. :p
     
  18. direblade99

    direblade99 Warlord

    Joined:
    Jul 19, 2013
    Messages:
    279
    Location:
    Adelaide, Australia
    Ugh, why would you do this
     
  19. Cromagnus

    Cromagnus Deity

    Joined:
    Sep 11, 2012
    Messages:
    2,272
    Played around with this with the Maya. Bulbed on t120, and got t151 Combustion. Full Liberty except the closer, then Full Honor, saved 2 GW to bulb Liberty closer when I needed it, then finished Honor.

    I forgot to take into account that there were 31 techs between where I was and Combustion. So, it took a lot longer than I expected to get Combustion. This is the downside to not having Scholars in Residence... you *really* need to bottom-feed the cheap techs... you can't skip any tech that gives bonus beakers, and that takes a while. But all those techs add turns.

    Peaked out at 27k excess beakers, but obviously earned more than that total. I started losing ground when I started researching the 1500 beaker techs. I was still gaining excess beakers with the 1000 beaker techs. The AI hadn't researched the 2200 beaker techs yet on t120.

    Initial bulb was 2k (off 2 GS) on t120, for reference. That got me one tech short of Combustion, but with the help of Oxford it didn't matter.

    I found out the hard way that if you take a Great Prophet with the Maya, it's bugged and it counts towards the GPP for GS. (I needed 200 for my first GS as a result)

    With a Garden and National Epic, you earn 9 pts per turn, so you complete a GS in 11 turns. If you can somehow get t90 Education, you could choose the Long Count GS on t101. But, I don't think it's worth it... gotta give the AI time to research techs. So, t117 is really the earliest point you can burn 3 GS.

    a 3000 bulb vs 2000 would have really helped. 50% more beakers means I would have peaked out at closer to 40k beakers, which would have been exactly enough to get Combined Arms, and starting the process 3 turns earlier would have helped too. It also would have made a huge difference to get the Pentagon before upgrading my 7 horsemen to landships. :lol:

    It was so expensive even with Honor! I had a GE saved up by t150 anyway from the Long Count, but I had crappy production in my capital, so one GE wasn't even enough to complete Kremlin or Prora. And you really want that 2nd tier policy ASAP IMHO. Hard to say whether it's better to burn the GE on Kremlin, Prora or Pentagon... You could go for the reduced unit purchase policy instead of Honor + Pentagon, but it's much more feasible to get an army insta-upgraded rather then build or buy it, IMHO.

    Anyway, this could have been t150 Combustion & Flight if I had bulbed 3k and not needed to bottom-feed as many techs along the way. Landships by themselves take capitals too slowly though... so save some money to build GWB. ;)

    Spoiler :


    Also, yes, I know my capital has no production... and my city placement was very non-ideal... it was just an experiment anyway. So, I think t150 is the earliest you can achieve Combined Arms, just because of the sheer number of techs you need to research along the way.
     
  20. Dardis

    Dardis Chieftain

    Joined:
    Sep 6, 2013
    Messages:
    25
    Firstly just want to say that this is a great spot, definitely needs patching!

    Just been messing with the mechanics of this without scholars (at the start). You get a little bit more mileage out of your GS's if you research all available techs to within a turn of completion. For example I researched civil service to within 20 beakers (pre-education), then bulbed 2 GS's for 1000 beakers - I got 362.50 overflow & spend 20 beakers on the tech. You then have several techs with under 50 beakers to research, in fact usually less than sailing (so you want to research these first). Also as Philip said above, bulb with no tech selected for a free turn of overflow.

    For optimal impact from your first bulbs you should aim to have the largest number of techs available to you, without forgoing writing. You therefore wouldn't want to follow a path such as researching all techs to the end of classical, since you can then only reduce 4 techs to 1 turn, missing out on 2-3 turns of almost full beaker multiplier.

    For fastest route to a SV, I'm beginning to change my mind on how to achieve this. It's going to be using the Maya, beelining Theology (GL?) & closing liberty. However when you bulb you don't want to complete everything you can, since that reduces the available tech multipliers when you get scholars. You probably just need to accumulate enough overflow to beeline printing press, but no more than this (possibly leave the sailing line untouched). Getting education & unis before bulbing is unnecessary, since you can hit printing press without this. Propose scholars (which should pass in a flurry of yeas if you are now tech leader) & then use pisa for one GS. In the meanwhile rather than teching further up the tree, just try to tech everything else to one turn so that the AI's catches up & overtakes. I just need to experiment tonight to see whether you need unis before pisa finishes - logically you don't, since they prevent the AI catching & overtaking after your initial bulb. They are probably only need to boost the GS beaker output around completion.

    Just one final thing with regards to researching to 1 turn - & this is a bit broken as well. If you are the first to research a tech, but leave it unfinished at 1 turn, when the AIs research it the beaker cost to complete the tech falls below the beakers invested. When you finally finish the tech you are given the beakers as a credit, which is then added to any other overflow & put through the multipliers. With scholars, a couple of credit techs, & a tech line un-researched from the start, you may be able to trigger a game winning amount of science without a GS.
     
Thread Status:
Not open for further replies.

Share This Page