This is interesting, i play it totally different. o.0 And i suppose im doing just fine, at least twice, no problem with anything.
My guide to victory and doing whatever you want:
I. Garage - use your 4 themes for good, trying every combination, 1st at PC (its cheaper) once G64 announcment move there and stick to it. Untill 1M. No research or custom research.
Just spam games as mad. Of course games should be decent, so the funbase will rise with each title. It could take around 2 years, the record is 9months. 
II. Even if you have just 1M move to other office, and instantly hire two cheapies balanced mini dudes.
Buy med games. Create your 1st custom engine with pad and mono sound. (yup)
Start looking at publisher deals. If there is none up just develop some good game, good topic/genre/console combo. Its not always about money. You need just enought cash to survive, and constant gain of funbase.
If its possible make (either own or published) game for matures, that gives free hype.
And repeat this untill you get 100k funbase. It shouldnt be so hard, just remember to to watch the deals and pick any which make sense.
At some point you can develop the 2nd engine, with some 2D v2 and some stuff but not too much addons.
Once you get 100k, and in fact you can start even little below start self publishing, like mad, again using good combos.
Generally use mostly the top 3 consoles, but it shouldnt matter much which one exacly.
Once you research sequels go for it. Sequels are good mostly since they use good combo once again without any penalty, and it seems that its not hard to impress.
And so on. Thats all mid game. Every 5th year develop some new engine more going for 3D than 2D. Game dont punish you for early 3D on gameling. Dont overengine with add ons.
Stick to med games. In fact i hardly remember doing large games in mid office. Meds are just fine. You get just enought cash, and the funbase grows nicly.
Anytime im moving to 3rd office around 16y/30.
Some good consoles to stick with: Gameling, PS, PS2. Dreamvast to some degree (but this is sad story).
You can train sometimes main hero, dont train mini dudes.
3rd Office:
You should have some nice cash like 200M so its time to fire your 2-4 mid dudes, and buy (for 2M each) more specific dudes. The final setup depends what you want in the end game (which genre) i generally go with all specialist exept Sound and WorldDesign. I suppose it gives the most versality in genres. The cheaper version is to get just Gameplay, LevDes, Graphic.
Simply retire old staff and get brand new, already trained (but not leveld up) dudes. Do some games, could be large even. Even pick some contracts works. At 1st your new dudes will be tired, and they need time to rest. And go back to your sequel self publishin routine. ASAP get R&D lab and start Net, Grind research.
After that they are few options:
Developing own hardware - im not exacly sure how good deal is that, since its hard to tell how much money gives own console totally, or whats the total cost of it.
I personally develop own console just for giggles and since i like the name HexGame.
Stick to PC, or even use some other consoles like PS4. This 100k extra development dosnt matter as much at this stage.
Go all MMO, just release any 1-4 MMOs (large), and expand them year after year.
Maybe it depends if you thing about sims to infinity, or just GG at 30.