Factorio Forums

Broken_splitter wrote: Thu Feb 13, 2025 5:32 pmWith constant fruit count flow (from box) it is ok, but the tree count decreases each time seed was delayed, less trees - less fruits - less seeds, so you can never farm 200 trees at all

This is why productivity bonuses are important. They make it far less likely that a string of bad luck will cripple your production.

Nemoricus wrote: Thu Feb 13, 2025 6:38 pm

Broken_splitter wrote: Thu Feb 13, 2025 5:32 pmWith constant fruit count flow (from box) it is ok, but the tree count decreases each time seed was delayed, less trees - less fruits - less seeds, so you can never farm 200 trees at all

This is why productivity bonuses are important. They make it far less likely that a string of bad luck will cripple your production.

I understand that. With few productivity modules I got ~2.1% per 10K fruits, even without biochambers

https://en.m.wikipedia.org/wiki/Gambler%27s_ruin

You have to do something to get a positive expected outcome or you will eventually run out of seeds (ruin)

computeraddict wrote: Thu Feb 13, 2025 8:24 pm https://en.m.wikipedia.org/wiki/Gambler%27s_ruin

You have to do something to get a positive expected outcome or you will eventually run out of seeds (ruin)

This doesn't apply in this case. "Gambler's ruin" applies when there's a negative net expected value. In this case, the expected value isn't negative... it's zero.

1 tree = 50 fruits * 0.02 = 1 seed/tree.

It's entirely possible to harvest trees on Gleba and come out ahead with seeds (or break even, or run out of seeds) with no extra productivity.

NineNine wrote: It's entirely possible to harvest trees on Gleba and come out ahead with seeds (or break even, or run out of seeds) with no extra productivity.

The reality is much harsher. With break even odds you’re virtually guaranteed to hit zero seeds if you go long enough. Such is the nature of random walks in one dimension.

Nemoricus wrote: Thu Feb 13, 2025 9:36 pm

NineNine wrote: It's entirely possible to harvest trees on Gleba and come out ahead with seeds (or break even, or run out of seeds) with no extra productivity.

The reality is much harsher. With break even odds you’re virtually guaranteed to hit zero seeds if you go long enough. Such is the nature of random walks in one dimension.

No, I don't think that's true. I think you're making the mistake in not considering the times when somebody gets more than 2% seeds. Since the seeds don't spoil, you can build up extra seeds half of the time, and they'll protect you in the other half of the time when you're getting less than 2%. It's entirely possible to continue indefinitely at 2%, and that likelihood increases with more trees that one has. The "random walk" goes in both directions.

That being said, I would suggest that anybody playing on Glega either uses production modules or biochambers or both so that there's less chance of running out of seeds.

That near certainty accounts for movement in both directions.

NineNine wrote: Thu Feb 13, 2025 9:14 pm ...
This doesn't apply in this case. "Gambler's ruin" applies when there's a negative net expected value. In this case, the expected value isn't negative... it's zero.
1 tree = 50 fruits * 0.02 = 1 seed/tree.
...

Please stop convince everybody in this thread. Just go and test it with assemblers only, without any productivity bonus

Broken_splitter wrote: Thu Feb 13, 2025 10:01 pm

NineNine wrote: Thu Feb 13, 2025 9:14 pm ...
This doesn't apply in this case. "Gambler's ruin" applies when there's a negative net expected value. In this case, the expected value isn't negative... it's zero.
1 tree = 50 fruits * 0.02 = 1 seed/tree.
...

Please stop convince everybody in this thread. Just go and test it with assemblers only, without any productivity bonus

I don't need to "convince" anybody. Math is math. Probability like this is generally covered in the first few weeks of an introductory statistics course. If you don't "believe" math, then there's not a whole lot I can do to help you.

NineNine wrote: Thu Feb 13, 2025 10:08 pm ...
I don't need to "convince" anybody. Math is math. Probability like this is generally covered in the first few weeks of an introductory statistics course. If you don't "believe" math, then there's not a whole lot I can do to help you.

Sometimes there comes a time when you need to stop and re-examine your theoretical base. Now is exactly such a moment

Broken_splitter wrote: Thu Feb 13, 2025 8:14 pm With few productivity modules I got ~2.1% per 10K fruits, even without biochambers

The challenge on Gleba is to unlock the unlockables fast, then use their benefits immediately. Getting a few biochambers fast is essential due to the built-in productivity bonus. Personally, I additionally add productivity modules, so the overall productivity bonus of one biochamber is +84% or 1.84 times the recipe output (prod modules 2). The built-in productivity bonus also applies to end products, so it applies for a biochaber crafting a biochamber as well, so you craft 3 biochambers for 2 input.

Unlocking the tech up to the biochamber is easy, you need to harvest a bit from everything and create a few nutrients manually or with an assembling machine you brought with you. You need spoilage, so don't hesitate to harvest some stuff. DON'T destroy any egg rafts you encounter yet.
Creae nutrients until you unlock the biochamber recipe. Import the ingredients for the biochamber from Nauvis (except the nutrients and pentapod egg). Additionally, create a few heating towers on Nauvis and bring it to Gleba. The recipe should have been unlocked from your "harvest everything" run. And bring a few assembling machines, a stack of heat pipes and a few stacks of steam turbines.
Create a preliminary power supply with imported solar panels and batteries that will last through a few crafts.
Now you have:
- electric power
- assembling machines
- ingredients for 3-10 biochambers except the nutrients and pentapod eggs
- enough spoilage to create enough nutrients for 10 biochambers (5 * 10 = 50 nutrients, requiring 50 * 10 = 500 spoilage for the "nutrients from spoilage" recipe.
- heating towers, heat pipes, steam turbines

Now go to the egg rafts you found. Destroy them and collect the pentapod eggs they drop (easy to miss, use the deconstruction planner to mark them and pick them up). Each small egg raft drops 1-3 pentapod eggs. They spoil in 15 minutes, so don't search too long for more rafts. Return to your base, create as many nutrients as required, build as many biochambers as you have pentapod eggs.
Bonus: immediately build the first biochamber you got and set the biochamber recipe. You will get 3 biochambers for 2 ingredient sets with that.
You need to supply a few more nutrients to make it work, but it is worth it.
Now you converted all pentapod eggs to biochambers. You should have at least 3. If not, go out and kill a few more egg rafts, preferably small ones.

The next thing you do is creating a preliminary yumako factory. Use one biochamber to create yumako mash from yumako. Filter the seeds this results and use them to supply your farm and plant more yumako.
Use the second biochamber to process yumako mash to nutrients, so they feed/supply each other. As fallback, add an assembling machine to create nutrients from spoilage to kickstart empty biochambers.
Burn surplus yumako mash in a heating tower and connect steam turbines. This will be your first real power supply. Kickstart it with a few rocket fuel or nuclear fuel you bring from Nauvis.

Now you should have your first production cycle: farm yumako. create yumako mash in a biochamber and get seeds. Bring the seeds to the agricultural tower to plant yumako. Create nutrients from yumako mash in a biochamber to feed the biochambers. Put surplus yumako into the heating tower to get electricity for the inserters and assembling machines you additionally build.
Next step is creating jelly from jellynut with the 3rd biochamber and kickstart the jellynut farm as well. You get more jelly per jellynut than yumako mash from yumako, so you will see your power production rise.

Now you created the first full farming cycle, start building a bioflux production. And so on.

If you create circles as I outlined, yumako and jellynut will be constantly harvested, mashed, seeds extracted, seeds planted, plants growing and harvested again. And it will be more seeds than you can plant. 2% plus productivity bonus seem small, but if you actually build it, you will see the seeds will accumulate and you will have more than enough of it. As long as you don't let the initial fruits spoil and process every single harvested fruit into its mash/jelly and use biochambers to do the mashing and not assembling machines.

To create infinite biochambers at this stage, use one biochamber to create biochambers and one biochamber to create pentapod eggs. Feed them with the nutrient production from yumako mash or from bioflux, depends on how long you wait and build more of your factory. Now create a feedback loop with the pentapod production biochamber, slightly similar to the kovarex enrichment process. After a few cycles, surplus pentapod eggs will appear which you can build biochambers with. You need the other ingredients of course, import them from Nauvis.

I mentioned "import from Nauvis" often - that's correct, without external kickstart from Nauvis, Gleba is a somewhat tedious start right at the beginning

NineNine wrote: Thu Feb 13, 2025 10:08 pm

Broken_splitter wrote: Thu Feb 13, 2025 10:01 pm

NineNine wrote: Thu Feb 13, 2025 9:14 pm ...
This doesn't apply in this case. "Gambler's ruin" applies when there's a negative net expected value. In this case, the expected value isn't negative... it's zero.
1 tree = 50 fruits * 0.02 = 1 seed/tree.
...

Please stop convince everybody in this thread. Just go and test it with assemblers only, without any productivity bonus

I don't need to "convince" anybody. Math is math. Probability like this is generally covered in the first few weeks of an introductory statistics course. If you don't "believe" math, then there's not a whole lot I can do to help you.

Hi, yes I'm sorry but you are incorrect.

Suppose I am a casino, and I offer you a game. You pay $1 to flip a fair coin. Heads, I give you back the dollar and one more. Tails, game is over. If you play this game a large number of times, the expected value is that you break even, and no one here is disputing that. Over the long term, the average should be a wash.

However, this thought experiment fails to take into account one crucial aspect of real life - you have a finite amount of money to start. Say you entered the casino with $100 to your name. Odds are over the long run, you hold $100 BUT also the odds that at some point you have lost 100 more games than you've won go up the longer you play. The issue is not that you lose on average over the long run with a fair coin, the issue is that at some point you are bound to have $0 and you do not have the option to play again. This is the gambler's ruin.

The situation is the same on Gleba - you start with some finite number of seeds. If you cycle seeds --> fruit --> seeds without any production, at a certain point you go broke. This is what computeraddict and Nemoricus are talking about.

travvo wrote: Thu Feb 13, 2025 10:34 pm The situation is the same on Gleba - you start with some finite number of seeds. If you cycle seeds --> fruit --> seeds without any production, at a certain point you go broke.

Possible, but increasingly unlikely if you create a big enough farm and if you make sure you create mash from every single fruit you harvest. If you forget about the probability and just play the game, you will see you might need to manually harvest just a few stacks of fruit again, but after you make sure you mash the fruits with a biochamber, you will get more seeds than you plant. You can now create artificial soil to increase the farm density.

It's also useful to invent a farm system that will only harvest if there is demand. Not just harvest every ripe fruit and store them - this will just let them spoil, while unharvested fruit stays fresh infinitely.

Tertius wrote: Thu Feb 13, 2025 10:31 pm ...
without external kickstart from Nauvis, Gleba is a somewhat tedious start right at the beginning

Thank you.
This whole topic should not appear in "Gameplay Help"- moderator's sorting hat just fails.
In my opinion seed rate for yumako and jellynut should be increased a bit to stay always greater than 2. And that it.

Broken_splitter wrote: Thu Feb 13, 2025 10:55 pm In my opinion seed rate for yumako and jellynut should be increased a bit to stay always greater than 2. And that it.

Unlikely to happen. I would not wait for that change. The current ratio is part of the Gleba challenge and guide. It requires you to use a productivity bonus, which is an incentive to focus on biochambers, which is THE central tool on Gleba.

Tertius wrote: Thu Feb 13, 2025 11:02 pm Unlikely to happen. I would not wait for that change. The current ratio is part of the Gleba challenge and guide
...

Well, thats OK too.
The next person with a similar question will find this topic and get all the answers from it.

NineNine wrote: Thu Feb 13, 2025 1:41 pm Both fruits work exactly the same way.

You are right

Tertius wrote: Thu Feb 13, 2025 10:43 pm

travvo wrote: Thu Feb 13, 2025 10:34 pm The situation is the same on Gleba - you start with some finite number of seeds. If you cycle seeds --> fruit --> seeds without any production, at a certain point you go broke.

but after you make sure you mash the fruits with a biochamber,

Once you use a productivity bonus greater than zero you have a positive expected outcome and are now talking about something entirely different than the rest of the thread is discussing. Gambler's ruin is specifically about zero expected outcome or negative expected outcome.

But you don't do this (using no productivity). It makes no sense. If it's about using no productivity, the whole thread is pointless, as far as I see it. There is always a little bit of loss in your factory and with spoilage, even with the best precautions, so the overall net gain with no productivity is below 1, so you're always running out of seeds in the long run. It's also extremely tedious to manually collect enough Yumako to start a farm big enough to support a decent Gleba factory, so it's not feasible to go without productivity. So it's required to use productivity for 2 reasons, so you need to base all calculations with the use of productivity and not without.

Tertius wrote: Fri Feb 14, 2025 11:40 am But you don't do this (using no productivity). It makes no sense. If it's about using no productivity, the whole thread is pointless, as far as I see it. There is always a little bit of loss in your factory and with spoilage, even with the best precautions, so the overall net gain with no productivity is below 1, so you're always running out of seeds in the long run. It's also extremely tedious to manually collect enough Yumako to start a farm big enough to support a decent Gleba factory, so it's not feasible to go without productivity. So it's required to use productivity for 2 reasons, so you need to base all calculations with the use of productivity and not without.

Correct, you don't do this. I think you may have missed the point of the thread.

Broken Splitter started with a complaint that even though the fruit/seed recipe looks fair over the long run, if you attempt to cycle fruit --> seeds --> fruit without productivity or addition of more fruit or seeds you eventually go bust, and therefore the recipe should be tweaked to have a net positive result, on average. The response to this is: Yes, this is a well known consequence of mathematics called Gambler's Ruin. If you play a random game with 0 expected value over and over, you eventually go bust because you don't start with infinite money. The takeaway is therefore that you should never play fairly, i.e. you should always use productivity from Biochambers and modules. No one's arguing that you shouldn't use prod, it's about if they don't use prod, here's what happens, and that's why you should use prod.