ED Exo Compare 1.2.18 — portable

A journal-linked exobiology companion for Elite Dangerous. It reads your local journal, merges scans and organic signals, and predicts the candidate species on a body from FSS/DSS telemetry alone — with payout ranges, first-footfall tracking, a system map, a customisable HUD and a species encyclopedia.

This release is the single-file portable build. For the unpacked program folder plus the two console builds, see v1.2.18-zip — same version, same code.

Download EDExoCompare.exe and run it. Nothing is installed.

Fewer wrong species before a DSS, checked against a year of other commanders' finds. Every rule added in this version comes from that year on EDDN (342,000 bodies, each species logged on the exact body), and none of them hides a single confirmed find there, on the two samples held back to test it, or in the developer's own journals. On those held-back samples, about 31 % fewer wrong rows are listed before a DSS than in 1.2.17, and three species that 1.2.17 missed are listed again. Which genera a body carries now also depends on its pressure and gravity. Community Goals now come to you as a notice, the built-in browser can keep its page cache and remember passwords (both off until you turn them on), and the Encyclopedia's body and atmosphere bars are measured on 52,000+ bodies.

Predictions

1. "Never found here." A species is moved to the unlikely list where it has never been found although its usual rate says it should have been — at least ten bodies expected, none seen, in the whole year and the held-back samples. 86 such rules over 35 species: temperatures, orbits, parent stars, main stars and regions (Frutexa acus is never in the Inner Scutum-Centaurus Arm or Hawking's Gap, for example). The reason is shown on the row, and the species is still listed under "show unlikely".
2. Colour stars. A species is listed as unlikely under a colour star (the brightest in the body's sky) it was never found under, where its rate expects 20 or more: 47 such pairs.
3. Stratum under A-type stars. Stratum has no colour under an A star: under a plain A star the only Stratum ever found is araneamus (1,486 finds, none of any other Stratum). Other Stratum species are now unlikely when every star that could colour the body is a plain A star. A Herbig Ae/Be star is not an A star here: six Stratum species grow under one, in their own colour.
4. Temperature limits. Where finds reach past the codex temperature range, the app now opens the range only as far as the finds really go, and a single odd find no longer stretches a range for everyone.
5. A hidden row comes back when its genus is almost certain. The 1 % line used to judge each row before the app weighed how often its genus grows on bodies like this one; a Tussock divisa read 0.4 % on bodies that carry Tussock 91 % of the time, and was hidden. Now a row the line hid comes back when nothing else of its genus is listed and bodies like this one carry the genus at least 80 % of the time.
6. Concha labiata above 180 K grows only on close orbits (420 of 420 such bodies within 12 ls); farther out it is listed as unlikely, and renibus is the one shown.
7. Aleoida spica from 165 K (was 167.7 K), Fungoida bullarum also on high-metal-content bodies, Stratum tectonicas also on argon-rich atmospheres — each confirmed on more than one body.
8. Five more planetary nebulae from EDAstro's newest list, for Electricae radialem.
9. Pressure and gravity in the genus chances. Most wrong rows on rocky and high-metal-content bodies were a whole genus that was not there. How often each genus grows on bodies like this one now also reads the body's pressure and gravity — on a year of EDDN finds they tell the genera apart clearly better (Tubus, Clypeus, Recepta, Fungoida, Concha and Aleoida most), while a combination of values inside one species (does more pressure let it live colder?) did not. "Chance here" calibrates better for it, and for twelve genera a body whose pressure and gravity they practically never grow at lists them as unlikely.

Measured on the two held-back samples (8,000 bodies each) against 1.2.17: wrong rows before a DSS 32 % and 31 % fewer; species missed 1 → 0 and 5 → 3. On the developer's journals: 764 of 764 confirmed species listed before a DSS, with fewer wrong rows.

New

1. Community Goals. When a new Community Goal starts you get a notice: where to sign up, the objective, when it ends, how far along it is and its rewards, with a button that copies the system name. The app reads Frontier's own public list of goals once an hour; nothing about you is sent. Switch it off in Options → Notify me → Community goals.
2. Browser: Keep page cache and Remember passwords, both off until you switch them on in the browser's settings. With the cache kept, sites you visited load faster the next time (cookies and site data are still cleared). Passwords are saved only when you say Save, encrypted with the operating system's protection for your account (on Linux, the desktop keyring; without one the option refuses to save), and filled in again on the same site (on its own when one login is saved for it, with the key button when there are several); they never leave this PC.
3. Encyclopedia: the body and atmosphere bars now show the share of bodies logged on EDDN (Stratum on 52,265 of them), and the unlikely list names the genus above a lone species.

Fixes

1. Browser: F5, Ctrl+R and the Reload button use the page cache again. They skipped it, which made every reload as slow as a first visit (about 360-510 ms against 140-300 ms on five sites). Switching blocking on or off for a site still reloads past every cache, so the change shows at once.
2. Browser: F5 reloads the page you are on. While the app was in front, F5 cleared the notices instead; it still does when no browsed page has the focus.
3. Closing the app with browser tabs open could show a JavaScript error ("Cannot read properties of undefined"); it closes cleanly now.

Where your data lives

%LOCALAPPDATA%\ED Exo Compare\ — settings, your LAN key, your EDSM credentials if you add them, the species you have scanned on foot, your Fix corrections, the journal folder you chose, your HUD layout, where your windows were, the systems you looked up on Spansh (kept 30 days), the shared exomastery folder, and the prediction caches. Nothing of yours is kept in the program folder, so replacing it on an update loses nothing. On Linux the same lives in ~/.config/edexo-compare. Backups go only where you point the Backups tile.

Your journals are read, never copied out. Uploads to EDSM, Canonn and EDDN happen only if you turn them on yourself. Looking a system up sends only its name or number to Spansh.

Windows will say the publisher is unrecognised. The executable is not code-signed, so SmartScreen asks first: "More info" → "Run anyway".

Licence

Code: MIT. Data: CC BY-NC-SA 3.0 (credit, no commercial use, share alike; part of it is derived from EDAstro) — see data/LICENSE-DATA.txt. The species photographs are Elite Dangerous screenshots: Frontier's game art, under Frontier's terms — NOTICE.md names every contributor and says which images are borrowed from the ED-DSN community. The list of Frontier-populated systems and looked-up systems come from Spansh.