SDA speedruns history:

* 4:44.13, Zach 'ZachSK' Kessler (February 2013)
* 4:06.03, Charles 'Arcanod' Nannan (March 2013)

Oh my, another platformer by Matt Thorson! Seriously, I would hate this guy if he wasn't also the author of An Untitled Story – a game you shouldn't miss if you're a metroidvania enthusiast. Give Up, Robot is a hard platformer where you can die basically anywhere, just like in the Jumper series. I suggest you give the game a try before watching the run; you can find it on Adult Games (the official website to play it). Before the game is loaded, you will see a small advertisement (Armor Games-style); please note that I HAD TO ENDURE IT BEFORE EVERY ATTEMPT!

So, why this run? Back in February, I was verifying Zach's run (his first submission). I'm easily attracted by platformers and it was enough to persuade me to play the game more in-depth. Still, I wasn't sure if I wanted to invest my time in this devious one-man production and risk an epilepsy (just kidding, but the game is excessively flashy and blurry), so I simply recorded some strategy guide to help anyone interested. By doing so, however, I began to have a good knowledge of the game, which I didn't want to be "wasted". So I finally changed my mind and started to play more assiduously. The hardest thing was probably to memorize the order of the 50 levels, but after around 15-25 hours of extensive practice, I was ready to record my first attempts (you know you're ready when you have died at least once in every possible location). The final result went out surprisingly quick, after gradually better attempts, in less than 24h (including a good night sleep!).

A remark related to the gameplay:
* The grappling hook system is absolutely not intuitive. I'd almost say it's a bit random. The only thing you should know about it, tough, is that the longer the rope when you release it, the faster you can go (assuming you don't release the arrow keys afterwhile).

The levels:

Note that the values indicated for rooms of improvement are only assumptions. I played each level individually to get an idea of what a "TAS" would look like and compared these times with my single-segment, nothing more.

01. Excellent. Jumping at the end is faster than using the hook again, but very hard to do at the right time because of the variable speed gained after the first release; though after some practice it works pretty well (about half the time with an average time of 5.3 seconds). The best time I ever got was 4.7.
02. (Pretty much) perfect.
03. (Pretty much) perfect. A room for improvement of ~0.1 second.
04. Good. More speed is possible (not guaranteed) if the hook is attached at the top of the block. Room for improvement: ~0.2 second.
05. (Pretty much) perfect.
06. Excellent.
07. Good. The rope could have released slightly sooner.
08. (Pretty much) perfect.
09. Very good. A TAS-like execution could be ~0.350 seconds faster.
10. (Pretty much) perfect.
Current time: 29.33.
I aim at 29.50 after these first 10 levels. Less than 29 is possible but hardly achievable.
______

11. (Pretty much) perfect.
12. (Pretty much) perfect.
13. (Pretty much) perfect.
14. Pretty good. The rope could have been released slightly sooner, obviously. Room for improvement: ~0.2 second. I don't like this level.
15. (Pretty much) perfect.
16. Good. Room for improvement: ~0.1 second, because of the second jump.
17. Good. The rope could have been released later, but the time lost is insignificant.
18. (Pretty much) perfect.
19. (Pretty much) perfect.
20. Excellent.
______

21. Very good. Using only the first spinner is harder than it looks. Still, a TAS-like execution would be ~0.450 second faster.
22. Excellent. Thanks to ZenicReverie for the spinner trick. It adds a new effective deaths generator to the game, but it is worth it! Room for improvement: ~0.250 second.
23. (Pretty much) perfect. To hit the top of the screen is probably the fastest and safest approach here.
24. Excellent, I suppose.
25. (Pretty much) perfect.
26. (Pretty much) perfect.
27. (Pretty much) perfect.
28. Good. Room for improvement: ~0.250 second.
29. Very good. It's possible to do this level in less than 6 seconds by attaching the rope at the very top and pass the middle section in one move, but it is insanely hasardrous.
30. Pretty good. To go straight is very hard and I really don't want to fall... Room for improvement: ~0.2 second.
______

31. Good.
32. Very good. Room for improvement: ~0.1 second.
33. Pretty good but I failed in maintaining the speed. This level is harder than it looks. Room for improvement: ~0.4 second.
34. Excellent. Room for improvement: ~0.1 second.
35. Pretty good, altough I could have released the rope later. With a TAS-like execution, it's possible to do this level in 1.3 seconds by going above the red block.
36. Excellent.
37. Decent. Room for improvement: ~0.3 second.
38. Very good, I suppose.
39. Decent. The end is a bit messed up. Room for improvement: ~0.450 second.
40. Good. Of course it is possible to grab the block earlier, but I find this way safer. Room for improvement: ~0.3 second.
______

41. This is the "plan B". Not bad, actually, but it remains the biggest time lost of the run. A bit raging because, surprisingly, I do this level right most of the time (i.e. in ~1.4 seconds). Room for improvement: ~0.750 second.
42. Excellent.
43. Excellent. This level is a laboratory, a source of experimentations. My most reckless strategy in the run is displayed here! And after some practice, it works pretty consistently, although it requires near-perfect execution. Room for improvement: ~0.450 second with a TAS-like strategy.
44. Excellent. It is extremely hard to preserve some speed at the end. Room for improvement: ~0.2 second.
45. Excellent, I suppose.
46. Very good, considering the difficulty. Also a level where the ghosting problem of keyboards can easily occur... Room for improvement: ~0.550 second.
47. Decent. I don't want to take any risk at this point. With perfect reflexes and coordination, it is possible to not touch the left side at all. Room for improvement: ~0.750 second.
48. Very good. With a TAS-like execution, it is possible to skip the second spinner and finish the level in less than 4.4 seconds.
49. Excellent. With some luck, it is possible to arrive directly to the exit, without hurting the wall.
50. Excellent, went as expected. The final level can be done in less than 24 seconds with a TAS-like execution.

Thanks for watching the run! I hope you enjoyed it. It was a true challenge.
If you want to leave me your impressions or ask any question, feel free to contact me at arcanod (at) gmx (.) com.