orchard target
orchard difficulty
monero target
monero difficulty
orchard window
monero window
| seq | chain | emitted | accepted | ratio | old | new | clamp |
|---|
| ratio | raw factor | applied factor | difficulty change |
|---|---|---|---|
| 0.50 | 0.8750 | 0.8750 | +14.29% |
| 0.75 | 0.9375 | 0.9375 | +6.67% |
| 0.90 | 0.9750 | 0.9750 | +2.56% |
| 1.00 | 1.0000 | 1.0000 | +0.00% |
| 1.10 | 1.0250 | 1.0250 | -2.44% |
| 1.25 | 1.0625 | 1.0625 | -5.88% |
| 1.50 | 1.1250 | 1.1250 | -11.11% |
| 2.00 | 1.2500 | 1.2500 | -20.00% |
| 3.00 | 1.5000 | 1.3200 | -24.24% |
the controller is a damped proportional step. after seventeen accepted shares on a chain, divide the intents that were emitted to that chain in that span by the shares accepted. that ratio minus one is how far the chain is behind. a quarter of that is applied to the target.
the clamps are asymmetric on purpose. the target may loosen by up to 32 percent in one step but tighten by at most 16 percent, so a chain that stalls recovers in a few windows while a chain that is flooding is reined in slowly enough not to swing.
the two targets never read each other. the balance between the chains comes from the routing byte, which sends half the intents each way over any long run, and from each controller keeping its own chain's fill ratio near one.