You are correct about the Torque Frame rates - if you try the Shader engine demo as it is currently you will find a big drop in frame rates compared to the Standard engine which in itself has questionable frame rates.

I can say from direct experience that I have a City level created in GS with hundreds of MDL entities one of which is a whole city block of 30 buildings. The models are additional to all the buildings created as level blocks. All of the MDL entities have little effect on the level frame rate which varies from a consistent 40 - 75fps in a level where there is a long line of sight along streets.

I have just tested the said single City model of 30 buildings as DTS model in an empty Torque level which when looking away from the model gives a fPS reading of around 160fps - if I turn around and look at the city model the FPS drops to 6fps - yes thats six frames per second.

Theres a lot of debate about these different engines around various forums and really reading the official hype, engine specs, forum debates, claims and counter claims and alike is really a waste of time - if you want to know which ones are best at what - more importantly which ones are even capable of making a "complete" game of any standing - then you will have to test out something thats been made with them thats extensive showing the capabilities of the various engines under stress. The trouble is you will be hard pressed to find any such thing made by any of them.

Most of the indie engines out there and some others which are referred to as more advanced engines which indeed they may be promote themselves as such with a simple one incomplete level - hardly a demonstartion of an engines capability or power. Theres nothing there to demonstarte such.

The Torque engine for instance (just using the example not picking on it) as you know has a demo FPShooter which despite its excellent terrain, water, sky combo has very little else - try filling it with some complex geometry and 30 different actors running around and see what happens.

In terms of generating anything like a quality modern commercial title content none of them currently have anything that I am aware of that can show us they can achieve that. There are a few exceptions to this I am sure but the general claim of many indie engines (and each do it) to be the best and easiest to make a game with requiring little programming is really misleading. If it were so there would be many more games made with them as there are combined a great deal of users worldwide but try as you may its going to be hard to find many completed games made by them or even playable demos of any stature. Yes admittedly you can make a game - "of sorts" with any of them if you want to keep it simple and primitive, but not many do. We are spoilt by modern commercial titles and want to aspire to achieve more and rightly so.

In truth all of them for one reason or another are either lacking the features needed or for differing reasons are extremely difficult for indie developers to be productive with. Whether it be difficulties with collision, AI, terrain or the bugs and more that some of them have - you are likely to spend a lot more time working out how to achieve what you want to do instead of actually doing it. By the time you get your game made if ever it will be outdated and you wont be happy with that.

GS at least as far as I have been able to ascertain in testing most of the engines discussed here without doubt is the fastest in production time and frames rates which are its strengths but it does have other poorly supported features and the continuing ongoing forward development of the engine is holding back small developers in my opinion while they await a stable production environment. At least it is is my case.

Using backwards compatibility to previous versions one could call upon a great deal of others experience and knowledge gained over a long period of time and a vast array of script examples and tutorials to help in your gamemaking production, most of what you would need or need to learn could be found somewhere here - with the move along the road to the new engine that is likey to be all lost as much of that content is to become obsolete - unless you can be the first to solve the new problems, create the new scripts, content, tutorials etc and lead the way then you may have to wait a long, long time before as a GS user you will be in such an enviable position again.