Post

Stars Reach Crafting In Practice

Stars Reach Crafting In Practice

In this post we are taking a close look at crafting in Stars Reach. Especially at what the current formula for calculating item stats (PQRV-values) means for crafting in practice.

This post might be a short-lived one, as crafting in the game is work-in-progress and likely to change. Potentially soon even. Thus, what is being described in here, might no longer be true, depending on when you are reading this.

The devs had always planned for “artisan crafting” - a crafting mode that includes crafter choices, maybe events, and different was to influnce the outcomes. But that’s not in the game yet - despite devs teasing it would come “soon”, a long time ago.

The Basics Of Crafting In The Game

You can probably skip reading this chapter, if you already know how the basic formula for calculting the new stats of a crafted item works.

The first initial draft of crafting originally was added to the game in summer 2024. It was a simple as could be: choose a recipe, add the required materials, hit the craft button - and out comes an item. All materials are the same, all items are the same. There were no variations, no dynamic stats, no components or sub-recipes, nothing.

It wasn’t until a full year later, that in mid August 2025 the PQRV stat system was added to the game. Technically, PQRV stats had been there before, at least on raw materials harvested or mined from the world - but they weren’t used in crafting and did not have any effect on items yet.

In my post on the Twilight Update, I describe how crafting with PQRV works, and what effects that stats had on crafted items. Those effects have since changed - as have the recipes. But the basic gist of it still holds true.

In my post on the Dynamic Update, a few weeks later, I describe how the recipe for the Terraformer changed to now include components (sub-recipes). Those and their recipes too changed a bit over time.

What never changed though, was the formula used to calculate the stats of a crafted item - based on the stats of the materials used:

Each stat individually was calculated as the average of all values of the input materials, except for the worst one.

Back in the Twilight update, a Terraformer was made from three materials:

  • Common Metal
  • Reactive Gas
  • Uncommon Metal

That means, each stat of the resulting Terraformer would be the average of the best two values for that stat, as found in the input materials.

In my example back then, I used the following the materials for crafting:

Material usedPQRV
Manganese434735470644
Flourine310466721600
Tin333751603359

For the resulting Terraformer, only the best 2 of each would be used.

StatInput 1Input 2Result
P434 Manganese333 Tin383.5
Q751 Tin735 Manganese743
R721 Flourine603 Tin662
V644 Manganese600 Flourine622

That means, that some stats on the Materials went completely unused:

MaterialUnused Stat(s)
ManganeseR 470
FlourineP 310, Q 466
TinV 359

And that was of no effect right then - but that changed, when the next update introduced components.

After that change, crafting a Terraformer now required 1 Stock, 1 Barrel, 2 Power Cells and 8 Focussing Lenses. And each of those components had to be crafted from raw materials in turn.

This two-tier crafting process significantly changed things. And the plan is to eventually have most recipes be like that - or to stack even more crafting steps. (Some professions, like for example cooking, might potentially stay closer to just raw materials in - final product out, in a single step.)

The “best 2” average (or “best 3” average, in the case of 4 ingredients), now took on a real significance.

The Practical Effects Of That Formula

For simplicity, we are going to use a simplified recipe and simplified materials with simplified stats (only going from 0-100).

So, in our simple example, a Terraformer will be made of:

  • 1 Stock (made from 1 Hide, 1 Metal, 1 Wood)
  • 1 Barrel (made from 1 Metal, 1 Oil, 1 Gem)
  • 1 Power Chamber (made from 1 Gas, 1 Metal, 1 Plastic)
  • 1 Focusing Lens (made from 1 Gas, 1 Sand, 1 Gem)

And we have only the following raw materials available:

MaterialPQRV
Hide90101090
2x Metal A90109010
Metal B10909010
Gas A10109090
Gas B90101090
Wood10109090
Oil10909010
2x Gem90901010
Plastic10901090
Sand10901090

And from that, we now want to build a perfect Terraformer with all stats maxed at 90. If you want, you can stop reading here, and try to figure it out for yourself - it is possible!


First, we will craft a Stock using these materials:

MaterialPQRV
Hide90101090
Metal A90109010
Wood10109090
Stock90109090

Then a Barrel from these materials:

MaterialPQRV
Metal A90109010
Oil10909010
Gem90901010
Barrel90909010

And a Power Chamber as such:

MaterialPQRV
Gas A10109090
Metal B10909010
Plastic10901090
Power Chamber10909090

And lastly a Focusing Lens like this one:

MaterialPQRV
Gas B90101090
Sand10901090
Gem90901010
Focusing Lens90901090

As you can see - by being careful and intentional about where we place our “dropped” values, we have gone from materials with 2 high values, to components with 3 high values!

Finally we have everything to craft a Terraformer:

ItemPQRV
Stock90109090
Barrel90909010
Power Chamber10909090
Focusing Lens90901090
Terraformer90909090

And as you can see - all the 10s are getting dropped - and only the 90s are used - leading us to a Terraformer with all 90 stats across the board. We’ve now gone from materials with 2 high values, to components with 3 high values, to a tool with 4 high values. Add another crafting step to the chain, and you could achieve the same, starting with raw materials that have only 1 high value each.

But that’s only possible from these half-half materials, in that one very specific configuration. If you swapped out anything, you would likely end up getting a 50 or 70 somewhere in the final Terraformer, instead of a full 90 everywhere. In other words: some low value would mix in to the average somewhere.

Depending on what materials and what stats on them you actually have available - the Power Chamber could end up being the “low-P” component as above, or it could end up being “low-V”, or “low-R” - shuffling everything around. It entirely depends on what stats your raw materials have.

Solving that little optimisation puzzle for a given set of materials isn’t too hard - but you can’t do it if you don’t understand that the formula means that one value is being dropped and you need only 3 (or 2) good ones for each stat. Having that mastery and understanding of the game, will allow you to create better items from the same materials than a crafter who doesn’t understand it.

I do not know if this is intentional design by the developers (might very well be) - or if it’s just accidential genius - but design-wise this feels like a pretty elegant way to add some player skill and mastery to the crafting system. Assuming that’s what is actually wanted. Outside of our simplified example, there’d be even more choices, since the good stats would be all over the place instead of being all 90s. The low stats don’t matter - they will get dropped - but with the high stats there will be more decision to make, like do you want more P or rather more Q? Do you optimize for greatest total across all stats, or do you value some stats higher than others? You’ll never have all high values be flat 90s - so there will be averaging happening, between the high values. But you can keep the low values out of these averages entirely.

The discussion question now becomes, do we want that form of mastery - or do we want everyone to be able to create the same items, so that only character skill and used materials matter - but no player skill is involved?

The Alternative: Regular Average Across All Values (No Drop)

The most logical other way to do things, would be to just take the average of all values - without dropping the lowest one. That would remove that optimization puzzle and the requirement for player skill. Everything would just be a plain average.

Here’s the challenge for you: Can you predict the outcome? What will the stats of the resulting Terraformer be, using the same materials as above - but with a regular average across all values, without dropping any?


A little hint: every Material above has two values at 90, and two values at 10. The average value on a single material thus is 50.

Here’s how that would turn out for example above:

MaterialPQRV
Hide90101090
Metal A90109010
Wood10109090
Stock63106363
MaterialPQRV
Metal A90109010
Oil10909010
Gem90901010
Barrel63636310
MaterialPQRV
Gas A10109090
Metal B10909010
Plastic10901090
Power Chamber10636363
MaterialPQRV
Gas B90101090
Sand10901090
Gem90901010
Focusing Lens63631063
ItemPQRV
Stock63106363
Barrel63636310
Power Chamber10636363
Focusing Lens63631063
Terraformer50505050

The average value on every material is 50. The average value on every component is 50. And the average value on the Terraformer is 50. And you can swap the materials around as much as you like - that’s not going to change. It’s always 50 everywhere. And that’s why I think this method would be considerably “dumbed down”, compared to the current “drop-lowest” formula.

In the real game, values are not gonna be just 10 or 90 - but all the many different possible values… You can still just add them together, and the material with the highest total across all values (or the highest average across all it’s values) is the best (overall). Period. There is no: “this material might be good in that one recipe, because it’s high P and low R”. It’s equally good or bad everywhere. The entire possibility space just collapses. Things become extremely “flat”.

And that’s one of the more surprising consequences, that I don’t think many people are aware of right now. People think they understand the crafting system, because they know how to calculate the “drop lowest” formula. But if you truly grok what effects that formula has in practice - then you cannot be surprised or disappointed that going from [90,90,90,10] to [90,90,90,90] doesn’t change anything - as that 10 was previously getting dropped anyway. You should know or at least suspect in advance, that if you already have three 90s - that upgrading that final material from low-grade to high-grade, doesn’t do anything.

If you truly grok it, then you should also have been able to predict, that this second example Terraformer here would inevitably end up with all stats being 50. Since nothing is being dropped, every 90 and every 10 in the materials will be used - and since those are 50 on average, the averages of their averages will be 50 as well.

And that’s why I’m writing this post - despite the risk that it’s gonna be completely outdated/deprecated when one of the next patches/updates lands, and crafting changes. I just hope that after reading this, more pople will be able to actually understand what wide-reaching effects that “drop lowest” actually has in practice. And that we will be able to have a more informed discussion about it.

The Issues With Drop Lowest Before Average

The formula is far from perfect though. It has indeed some serious issues. Issues that I think will need fixing.

Like for example, if you use completely random materials without looking at their PQRV at all. On grand average, the expected outcome would be that you get roughly 500 points per stat. Because, that’s just the average value. Just like it happened with 50 in our simplified example.

But, with the “drop lowest” formula, what you do get from completely random materials, is higher than that. Because the lowest value is being dropped and not used in the formula - the expected grand average for a component crafted from completely random materials, is not 500 - but ~600. And when you then craft a terraformer from that, the expected grand average on that is expected to be around 680 to 700 rather.

That means, that with zero effort und just using whatever materials you randomly get your hands on - or using materials that have been “mixed” together over and over again, by combining stacks - you can still create a Terraformer with average stats of 700 for all values. And that’s then pretty much the “worst” Terraformer there can be. And that’s a problem - the worst, zero-effort terraformer should be 500. Or arguably even lower - but to achieve less than 500, you’d need materials to have lower stats overall, regardless of which formula you use.

But the plain average (the alternative from above) does achieve 500 - just the “drop lowest” does not. But a modified “drop lowest” could achieve 500 - or even 400 in rare cases. But that would require a more complex formula. (Replacing the lowest with the median instead of dropping it, then taking a harmonic mean).

In the end - whatever formula we get - I think that the crafting UI should show a preview of estimated result stats. So that you can put in different materials, and look what happens to the stats - without having to actually press “craft”. Allowing you to try out different things, before committing. Then you wouldn’t need so much to calculate the result ahead of time. And could over time also develop a gut-feeling, a rough idea of what to expect to come out.

Another issue or not with the current formula might be, that crafting something from just two ingredients turns into a “max()” function. No average at all - just taking the best value for each stat. That’s very powerful. But if that is an issue or not, depends on how and where these two-input recipes will be used (like alloys), and how many good stats a single material can come with. But the modified formula would fix that max() as well. And it’S not the only way to do that - shouldn’t be a problem to come up with other formulas that achieve the same thing. This is just the simples formula I could find, that achieves keeping the optimisation space alive, without that stats-creep up to first 600 and then 700.

Conclusion

I hope you do now have a better picture of the intricacies of what that “drop lowest” formula actually means for crafting in practice.

Not only why it means that only 3 out of 4 ingredients matter for a single stat, and the 4th does nothing - but also how it creates that possibility space of using different materials with different stats in different recipes, which allow for mastery and player skill to matter. And I hope you also noticed how replacing the current formula with a plain average that drops nothing, leads to rather bland results, where everything is just invariably average.

As said in the introduction - this post will probably be outdated pretty soon. Depending on what the devs have planned for the artisan crafting mode, that might cover all the needs for mastery, player skill and choices. But even then, I personally do find it extremely fascinating to really work out in detail, what effects that little “drop lowest” rule there has - and how much it affects things. It is suprising, isn’t it?

This post is licensed under CC BY 4.0 by the author.