Because you have gone trough code of both projects and actually know how ED stuff has been made? Like FD actually did build their own galaxy life cycle simulator for Elite sakes! SE concentrates on visual appeal, while FD strives for scientific perfection, visual middle road between realism and "how it should look like". And that's cool. Because SE goals are very different than ED, besides it's not being even a game.
Take the end results of two simulations, if one produces equivalent or better results from simple equations, than a more complex model, which is better? If one produces those results far faster, or allows for more flexibility or greater scope, which is better? If I painstakingly work out a distribution of masses from first principles, or just plug in a distribution I found from a scientific paper, and both give approximately the same results, are you as end user, going to be able to tell the difference? There's a lot of work done on producing quick running meta-models, from more complex 'accurate' models.
It doesn't matter how accurate you think, or FDev says, Stellar Forge is - look at the end results
at present. Space Engine's overall scope (many, many, many galaxies), lighting, variety of planets and objects (asteroids/comets, star clusters) and incidental and atmospheric effects are better than E: Ds. It also doesn't suffer from the 'cubes' of stars of varying sector density near galactic cores, which make E: D's skybox a bit fugly, and probably the opposite of 'realistic'. Down low however, E: D's rocky worlds, textures and terrain can look much better than SE's (plus E: D has geysers

). And E: D's suns do look quite nice close up, but SE also does irregular shaped, and oblate stars. Also love the big, hemispherical storms raging on, and icy backsides to, tidally locked worlds in SE. But rivers in SE often look terrible. And yes, all of this
is subjective.
You could also argue why would a game, which doesn't need to be
accurate, just a facsimile thereof, uses so many resources on an accurate portrayal of the galaxy, when gameplay might have been improved by using a less accurate, but 'more Hollywood'/'more controlled' representation. Take, for example the hand-placed/catalogue imported star clusters in E: D, that line up like arrows pointing back to Sol. Are they 'realistic', in the sense that those stars exist in reality? Yes. Do they look out of place and 'unrealistic', given the density of surrounding stars, and the artifacts of their creation? Yes again. That said, I am impressed by what
both Frontier and Vladimir Romanyuk have achieved.
[FWIW, I've gone through the (reverse engineered) code for galaxy generation for Frontier/FFE. Very interesting stuff.
http://www.jongware.com/galaxy1.html. Plus SE may eventually become a game - though again I'm not disputing the very different end goals for each bit of software.]
*Edit - I'd also be curious to here from people with more astronomical knowledge as to whether the number of stable, binary pairs (or larger) configurations of planets E: D throws up with alarming regularity, are viable? And some large satellites seem to quite happily breach Roche Limits (based on my shonky A-level physics.

)
*Edit 2 - I'll also concede that, in the absence of peer-reviewed expert comments on either Space Engine, or Stellar Forge, I've no basis to suggest either is more
accurate than the other. One is prettier, with fewer visual oddities and artifacts, IMHO though.
