Tinkering Desk: combining blueprints
Everything on this page comes from our own testing on the cluster: more than 2,500 combines on our test server with our engine tools, plus every mode run through the real desk. If we haven't tested it, it isn't here. We repeated the 1,000-combine Stable and Unstable runs on ASA 93.38; their numbers here come from that re-run. Everything else, including the Flak Chestpiece combine, is from our first round on 93.19. The desk's settings match on both builds. The Tinkering Desk combines blueprints; the Helicoprion makes them.
What the desk does
The desk takes exactly three blueprints. Blueprints only. Items are refused. It needs power. The combine takes 3.5 seconds. The result lands in the desk.
There are two modes, Stable and Unstable, and nothing else.
Quality
The desk averages the quality of the three blueprints. Call that the average quality. Each blueprint's rating is divided by its stat multiplier, those values are averaged, and the output rating is the new quality times the same multiplier. For the same item, averaging the three ratings is that math.
Stable and Unstable
Stable: the result is always between the average and 1.2 times the average. It never fails: 0 losses in 1,000 combines. It never comes out below the average: 0 in 1,000. Most results land around 1.1 times the average, and landing within a hair of 1.2 is rare: 1 in 1,000. The game's own info text says Stable "can have Quality as low as 50% that average". On 93.38 it can't. We never saw it.
Unstable: the result is exactly the average quality, and there is a 25% chance the desk destroys all three blueprints instead (measured 23.7%, 95% confidence 21.2 to 26.4, 1,000 combines). The game's info text says Unstable has "only a 25% chance to successfully combine". It's the other way around: 75% of Unstable combines succeed.
So Unstable trades a one in four loss for a guaranteed exact average. Stable gives you at least the average and up to 20% more with no loss. There is no reason to use Unstable on 93.38.
| Mode | Result quality | Loss chance |
|---|---|---|
| Stable | Average to 1.2 times average | 0% (0 in 1,000) |
| Unstable | Exactly the average | 25% (23.7% measured) |
Unstable's 23.7% is from 1,000 combines (95% confidence 21.2 to 26.4).
A combine we ran through the real desk
We ran three Flak Chestpiece blueprints through the real desk, ratings 2, 3 and 4. Average 3. Stable gave a rating 3.26 blueprint. Unstable gave exactly rating 3.0.
| Blueprint | Rating |
|---|---|
| Flak Chestpiece | 2 |
| Flak Chestpiece | 3 |
| Flak Chestpiece | 4 |
| Average | 3 |
| Stable result | 3.26 |
| Unstable result | 3.0 |
Stats and durability
The result's stats are rolled fresh from the new quality. They do not carry over from the inputs. Three blueprints with a great armor value and poor durability don't give a great-armor result. The average quality decides everything.
Durability of the result is full. There is no durability bonus for combining blueprints. The 30% durability bonus you may have heard about only applies when items are combined, and the desk only combines blueprints.
Mixed inputs
Three different blueprints combine, and the result's item type is picked at random among the three, each equally likely. We ran 600 combines of Flak Chestpiece, Pike, and Machined Sniper: even split. Same quality math.
What the desk refuses
Non-blueprint items. Zeppelin and Helper Bot spawner blueprints. Unique blueprints (Scout remote, Enforcer spawner) only combine with copies of themselves.
What does not matter
Your crafting skill. Where you stand. What you're wearing. The blueprint's quality color. Only the quality number matters. The color follows the number.
How to use it
Combine equals. One Ascendant with two Primitives averages down. The result is at most 1.2 times that average. Three similar blueprints step up together.
Then craft from the result with crafting skill. Or run your three ratings through the desk calculator.
Myths
Unstable only works 25% of the time. It works 75% of the time. The game's info text has it backwards. We measured 23.7% losses in 1,000 Unstable combines, so three in four succeed.
Stable can drop your quality. It never comes out below the average. 0 in 1,000 Stable combines landed below the average of the three inputs.
Combining boosts durability. It doesn't for blueprints. The result comes out at full durability. The 30% durability bonus you may have heard about only applies when items are combined, and the desk only combines blueprints.
Crafting skill or the crafter matters. It doesn't. Your crafting skill, where you stand, and what you're wearing do nothing at the desk. Only the quality number of the three blueprints matters.
The best blueprint sets the result. The average does. One high-quality blueprint with two low ones averages down, and the result is at most 1.2 times that average.