If possible, could you by mention later next week if any of the major implimentations will not be resolved?
Thinking about the following:
A) Mud drop should be below player (so that it still affects player directly behind).
B) Issue where First tile hit while turning gets ignored.
Looking at the Github issues, and the announcements these are the only points I saw related to “game engine bugs” that is not resolved.
Also feel Point B is major. As it leaves a fairly large “blindspot”. Which some including myself would like to account for (if left unchanged).
Secondly we also have the Anouncement:
C) Dual Match feature
If any of these will be updated (or not updated) by the competition, please let us know.
So that we can make sure our bots work in the “final” harness.
And not what it is now, or what we think it will be.
Just some random thoughts.
+1 on B. I was also wondering about it - it feels like there’s no doubt that it is a bug, and with the first tournament almost on us, I’d really like to know whether I should “adapt” my bot to the bug.
My 2 cents - make it configurable, if possible in your implementation. That way, if the bug gets fixed, you just flip a switch, boom, you are in sync with game engine again.
Of course, if you are using advanced techniques (actual ML ), then it might not be as simple as that…
I would like B fixed prior to the tournament, the bug that caused it was a fix that I had to cater for to actually not get hit by it but now it seems this one is going to make me code something silly into my code to cater for a new bug which to me seems like I’m panelbeating my bot to make it work with a faulty engine.
I had to do the same panelbeating for the previous bug that was fixed that caused this bug.
I already used up enough time adding fixes closer to the start of a tournament and then missed something in testing that caused my bug to “bug” out, I would much rather prefer it bugging out on the engine then making code adjustment without enough time to test.
I made a pull-request to fix B (issue with turning onto a special block) at https://github.com/EntelectChallenge/2020-Overdrive/pull/57 - since I don’t know Scala or the game-engine I’d appreciate it if someone else could also test it (working starter pack downloadable from the automatic builds), but from my own manual (albeit very limited) tests it seems to have fixed the issue.
Well, to date, lines of scala code written by me === 0.
And this 2020 challenge is also about the most I’ve read.
That said - looks good to me.
However, I must ask if it is a good idea to merge that bug fix so close to the first tournament.
I would guess pretty much everybody has coded to the engine behavior as it stands.
Maybe consider merging only after the first tournament?
Can we then get clarity on this today then whether or not this will be fixed?
Is this a better path for me to turn if I’m not getting penalised for hitting something directly on the left/right of me?
I wouldn’t want to be left at a disadvantage for this?
The first tile immediately 6o the left / right when turning gets ignored (Both powerups and Mud).
Had a few matches turn bad because my bot went for a powerup in its blindspot…
I’m with you on this, that’s why I want to know from the Entelect team if I need to code for this bug or not.
I seem to having hit and miss on this, it’s just 2 lines of code changes however, some matches my bot with the change in wins and some matches, my bot without it wins, this balance seems quite inconsistent, would like to enter my bot without having this problem, I do now however have 2 minds about it as either way, this seems to be a disadvantage.
Thanks @NakitaReddy. Just pass on a few things to take into consideration also: time might play a bigger factor than a bug fix here with respect to point B. The latest release was released 14 days ago. The first tournament is in 3 days. If a bot was trained with the latest game engine, it would have to be re-trained to cater for the breaking change. The new training would also be shorter. Also, by putting in a fix so close to the tournament, there’s a risk of introducing new bugs for both the game engine and people’s trained bots.
If everyone is disadvantaged in the same manner, is that not by default a level playing field? @marvijo I agree the risk of introducing change to a platform so close to the go-live might just outweigh the gain.
I’d rather spend the next few days optimizing my bot instead of stabilizing.
