53 vs 54 point word rarity in 5x5

[BoggleOtaku]

I happened to score a 53-point word (PW) today. I was under the impression that a 53PW hadn't turned up previously, so I took a look through the monthly records.
I found the last time it was scored was apparently September 2024.

The best word records indicate only one 53PW and only one 54PW have been found. Previously players had predicted there is only one 54PW prior to it being scored (MegaWord IIRC). I am not sure if there has been a similar prediction for 53PWs.

Here the yearly totals (all sightings each month for all months) I found from Monthly best word records, for 2021 through 2025 to date:

54PW: 6 + 4 + 7 + 8 + 11 = 36 total all-time finds.
53PW: 0 + 0 + 0 + 12 + 1 = 13 total all-time finds.

It's not perfect, because double findings (same person finds 54PW twice in same month) are masked, and scoring of 53PW is masked if same person scores 54PW that month. The scores are sparse, so there are probably not many such cases.


These numbers make me wonder if the 53PW was added later than 2021, maybe in 2024.
It doesn't seem like it would be any harder to find than the 54PW...maybe it doesn't come up as often?
Also, I am wondering why there are so many scores for it in 2024 but almost none in 2025.

[Spike1007]

Congratulations on the 53 point word! I'm fairly certain I've never found that one, and I don't really remember getting any other 53 pointer either.

While I couldn't swear to it, it does seem that some words (or families of words) rotate in and out of favor. (One group that sticks out are the SEMIDEMI...QUAVER(S) or whatever. There were times in the past that they seemed a lot more common.) Anyway, I decided that even values for long words were more common than odd values, and I kind of attributed it to how words are scored and the fact that there aren't a lot of the longer words, so the presence of special letters J, K, and Z are somehow not as well represented as Q and X (or no special letters) there.

[BoggleOtaku]

Thanks for the reply, Spike1007.

I have also thought there might be odd even disparity in scoring. Looking back over the last year it is striking.
Here's the number of monthy scores Jan-Nov 2025 for even versus odd points between 47 to 52.

odd (47, 49, 51): 4 + 1 + 2 + 0 + 3 + 1 + 5 + 5 + 7 + 1 + 3 = 32 found (to date 2025).
even (48, 50, 52): 15 + 13 + 14 + 17 + 18 + 8 + 15 + 5 + 13 + 13 + 17 = 148

That gives an even/odd 4.6 : 1 rarity ratio.
This pattern appears to extend down to low 40's scores and further, with the ratio decreasing. [edit: maybe through 20's even]
(Note: Masking is more prevalent than for 53 & 54 PW, but probably affects both even vs odd somewhat equally.)
There looks like just a handful of odd point words in the above range, maybe 1 or 2 (47pw) I guess predate 2021 but didn't check.

I have also noticed words coming through in cycles. I was having fun finding -iferous words and trying to learn some of these when they seemed to become more scarce. Maybe that is true for -phobias and also large numbers (there seemed to be a drop for a while some time back but back now).

[BoggleOtaku]

BTW, I'm sure it was posted before, but I gather from your comment that the only way a word can have odd points is to have an odd combination of J+K+Z, each of which has an odd amount added to the score. That would tend to be rarer... Except for I'm thinking there are odd points for odd number of letters and vice versa, so I would think that would even things out... I should have paid more attention when this was discussed by you, lalatan, megaword and others years ago IIRC.

[Spike1007]

The value of a n-letter word is m*(n-1)+s, where m is a multiplier based on word rarity (m=1 for common words, 1.3 for broad, 1.6 for rare and 2 for ultra-rare words). Thus the value for any ultra-rare word is even before adding s, the special letter scores (1 for J & K, 2 for X, 3 for Z, and 4 for Q). (By the way, each special letter can only count once per word.) There are probably plenty of long, ultra-rare words without special letters, giving lots of even word scores to bias the stats in favor of evens. I'd expect things to even out for lower word scores.