Good/better/best algorithm

[pat]

My daily goal, apart from getting the seed word, is to hit 'best' (nice to get a rosette as well, but to some extent that depends on how long I stick with a puzzle). Maven is generally out of reach for me as I'm not that bothered about rare words.

I'm curious as to how the levels are set within individual games. It's obviously not as simple as percentages yet somehow it seems to be spot on - I usually manage to hit 'best' quite easily and then have to work harder for the remaining words, whether they be three or twenty-three. How is it calculated?

[Jacki]

Who is Maven anyway?

[pat]

Good question! I'd imagine it's irrelevant to most of us mere mortals.

[anona]

I think an alternative definition of maven that applies to me (on the rare occasions I rise above the mere mortals) is "someone with too much time on their hands".

[rogue_mother]

Alan posted a general explanation of the algorithm in this previous post from 2014.

Like non-a, I feel that I've given the formula before, but I can't track it down in the forum index, so I'll spell it out here. With luck I won't contradict what I said last time.

Attaining each of the five target levels, Good, Better, Best, Maven and Cham, is based on the total number of words you have played, common and rare. So there's no automatic connection between getting to one of these levels and scoring a rosette. You could get every word in a puzzle except for one common word, which would put you above the Cham target, but without scoring the flower.

The Good target is the lowest whole number greater than 40% of the number of common words. All the other percentage based levels are calculated the same way, so let's just say "about 40%".

Better is about 65% of the number of common words, and Best about 90% of the number of common words. So the first three levels are based only on the number of common words, but, as I said, all the words you play count towards reaching each level. So typically, when a player reaches Best, they won't yet have got quite 90% of the common words, because they will have played some rare words.

The highest level, Cham, is about 90% of the total number of possible words, common and rare. So, at last, we come to Maven: it's half way between Best and Cham.

[pat]

Thanks, r_m. That explains why I didn’t think it was percentage based - I didn’t take rare words into account. But however it’s done, I think it works pretty well.

[Alan W]

I think the formula for Good, Better and Best has remained unchanged since day one. (Maven and Cham were introduced a bit later.)

Before Chi went online I obviously didn't have the benefit of any statistical information about player performance, so I just had to come up with something that seemed about right, based on my own playing and the comments of friends who tried it out. It does seem to have worked out fairly well, in that the peak of the score distribution is usually around the Best target.

[Jacki]

Well I've learnt a lot there! Here I was thinking Maven was a person who's comparative ability was a good yardstick - sort of like T-Rex or Alonzo Quixote but more within reach!

[Alan W]

What is a maven?

Also - What is a cham?

[mkenuk]

'Will your grace command me any service to the world’s
end? I will go on the slightest errand now to the Antipodes
that you can devise to send me on. I will fetch you a
toothpicker now from the furthest inch of Asia, bring you
the length of Prester John’s foot, fetch you a hair off the
great Cham’s beard, do you any embassage to the Pygmies,
rather than hold three words' conference with this harpy.
You have no employment for me?'

Benedick in Much Ado About Nothing begs Don Pedro to send him on any dangerous mission that will get him out of a meeting with Beatrice.

(I watched Catherine Tate as Beatrice a few days ago - very impressive, very funny; almost on a par with Emma Thompson in the Kenneth Branagh classic version.)

[Jacki]

While I'm being educated is there is method to the order of names for those who reach a rosette in the same number of words, ie today's 10am quiz was reached by lots of people in 35 words and yet they're all spaced differently. Do you know what I mean? Is it the quickest is placed first or is it random? It's not alaphabetical I don't think.

[mkenuk]

It's done chronologically, Jacki.
On the 'common words only' scoreboard, first to a rosette gets top place and keeps it until the next game starts.

On the 'all words' scoreboard the rosettes do get more 'spaced out'.

[Alan W]

Jacki, I take it you're asking about Friday's Standard puzzle (started at 10.00 am here in Eastern Australia, but at different times in other places).

As mkenuk says, if things are working as they should, the players on 35 words should be listed according to when they got to that score. I.e. the first player to reach 35 at the top. Exactly what you see depends on what view of the scoreboard is in force - including rare words or not. In the view that includes rare words, most of the players who got a rosette will be shown with more than 35 words, because they played some rare words as well as the 35 common words. But those who got the rosette playing common words only (and I'm surprised at how many of these people there are) are interspersed with players who got 35 words in total but got no rosette, because they found fewer than 35 common words.

In the scoreboard view that excludes common words, there is a big bunch of rosette winners at the top of the board, all on 35 (common) words, with the first player to win the rosette listed first.

[pat]

Jacki, you might also notice what appears to be a discrepancy while making your way up the scoreboard. If you’ve got, say, 23 common words but haven’t yet got the big word and there’s someone just below you on 22 who does have the big word, if that person gets another word then he or she will appear above you on the scoreboard, even though you got 23 words first, the reason being that within any score, those who’ve found the big word appear before those who haven’t.