Post Reply 
Programming puzzles: processing lists!
02-14-2019, 10:33 AM (This post was last modified: 02-14-2019 10:34 AM by pier4r.)
Post: #272
RE: Programming puzzles: processing lists!
(02-14-2019 04:10 AM)DavidM Wrote:  Nice, John! I had originally tried something similar to that (but using LSEQR instead when the scores started with 7). When you talked Pier into changing the scoring rules, I suddenly remembered that there was an equation for the sum of 1..n and decided to use that instead.

@Pier (and anyone else): is the progressive points on successive wins a common scenario? I obviously don't play anything enough to recognize that type of scoring. Just curious.

Yes indeed I wanted to start with an offset to "mask" the 1+..+n ; often known via the story of Gauss https://www.nctm.org/Publications/Teachi...-of-Gauss/ .

But then I said "well, people will recognize it in any case, let's make it simple. The offset can be always added". And I modified it according to John's request.

I am still waiting comments from those with better knowledge of RPL that tells me why vectors are so faster than list when modified (see post #269). I should have tested it before! Incoming "processign vectors!" thread.

About the scoring time for digression: I like to organize tournaments (or ratings between players), so every now and then I when I see a new tournament organization I try to analyze it. In the vast majority of cases you have those elements:
- normal round robin (or double round robin).
- 3-1-0 scoring system https://en.wikipedia.org/wiki/Three_points_for_a_win
- like in chess 2-1-0 scoring system.
- knockout format. This allows many upsets the less legs in a match there are. In the US they do right allowing bo5 or bo7 (basket or football playoffs), lowering the chances that one team lucks out a win. See also http://pier4r.wikidot.com/pierworks:arti...formatcomp
- swiss system. That requires less pairings than the round robin but it is better than the knockout format to identify range of players as those ends up being grouped by the same amount of points. Although upset may happen too. I personally love it. There is a lot of interesting little math investigations for the tiebreaks of the swiss system.
https://en.wikipedia.org/wiki/Tie-breaki...ournaments (and much more is there that is not listed)
- match up according to elo ratings or their variants (they result in quite balanced matches): https://en.wikipedia.org/wiki/Elo_rating_system
- I hope I didn't forget anything common.

The last year I noticed the scoring of the lichess titled arenas. In short it is 2-1-0 as normal in chess, but if your last game was a win, the next game has its outcome doubled. It is not bad as it rewards activity but with good play.
I never saw this applied elsewhere so far (surely someone else did but it was outside my sight).

Then from few months I noticed the tournament format in RTS game (art of war 3. Not a bad game ) and I like it a lot. It is exactly the #44 and I find it even better than the lichess arena because it has a cap on the amount of games that can be played (otherwise one would just play a lot hoping in streaks).

The nice part of identifying the score rules in tournaments is the reverse engineering part as the tournament organizer often forget to explain how the scoring works, even if the scoring is uncommon. So one has to crack the scoreboard trying to identify how such scores are possible.

Wikis are great, Contribute :)
Find all posts by this user
Quote this message in a reply
Post Reply 


Messages In This Thread
RE: Programming puzzles: processing lists! - pier4r - 02-14-2019 10:33 AM



User(s) browsing this thread: 1 Guest(s)