Average Points Per Turn: THE Scrabble statistic!
Donald Sauter
When I meet someone, and it comes out we're both
Scrabble players, the next thing I usually hear is, "How
many points do you score in a game?" Then I find myself
explaining once again that points per game is not a very
telling statistic under the best of circumstances, and
in the case of someone like me who plays a mix of two-,
three- and four-person games, is completely worthless.
The Scrabble statistic that always does the job is
average Points Per Turn (PPT), and how the Scrabble
world has not cottoned to it is a major bafflement to
me. Just as the batting average in baseball tells how
likely you are to make a hit in any trip to the plate,
your PPT in Scrabble tells how many points you're likely
to score on any turn.
Points per game, besides being a very coarse statistic,
is not too interesting because the total points that can
be scored is more or less limited regardless of how good
you get. After all, a Scrabble game always uses up the
same tiles on the same board with the same pattern of
premium squares. That yielded about 620 points per game
back in the days of the first OSPD. (It would be
somewhat greater now with the addition of QI, QAT, ZA,
etc. I don't have data to work with for play with
OSPD4.) Two novice players may take 21 turns to get
there, while better players may take only 16. Heck, if
two chickens could be taught to kick Scrabble tiles onto
a board, they would score in the same ballpark. (Be
advised: sometimes I exaggerate ever so slightly to make
a point and wake people up.)
Yes, there can be exceptional games where every power
tile gets counted six times, but such games owe far more
to the luck of the draw than developed skill. Expert
players score significantly higher, but that's mainly
due to their higher bingo production. Take away those
50-point bonuses and their final scores would look much
like ours. Put the other way around, if you did detect
a small upward creep in your average points per game, it
would be mainly due to increased bingo production. And
that's something which would be more fun and useful to
track directly.
Scrabble is about grabbing up those limited, available
points in as big handfuls as possible - and that's what
Points Per Turn measures. Just add up all the points
you score and divide by the turns taken. Ignore the
adjustment for the leftover tiles, but you must count
your 0-point turns! If you traded tiles, you did so for
the sake of _future_ points. If you got caught playing
a bum word, well, that was a strikeout!
I made the analogy between batting average and PPT, but
there is one fantastic difference. A batting average
only makes sense in the context of the "league" in which
you play. A 12-year-old slugger batting .564 in his
Little League probably wouldn't do so well against Major
League pitching.
Your PPT in Scrabble, on the other hand, is almost rock-
steady no matter what level of competition you find
yourself in. After all, what difference does it make
whether the board you are stewing over was cobbled
together by experts or by duffers? Ironically, if you
are a recreational player with a 19.6 PPT, say, you
would probably see it rise a hair against experts, since
their longer words open up more scoring opportunities
for you.
Since there are so many turns in a game, a very
meaningful PPT can be calculated in just a few games.
It doesn't depend on games always being one-on-one. It
would not be affected by a game being cut short, or
games using more than 100 tiles. (In my club, we scoop
a fresh set of about 110 tiles for each game from a
mixture of three standard sets.)
I am not proposing that Points Per Turn replace the
National Scrabble Association rating system, although I
do think one would be hard-pressed to explain how the
player with the best PPT average wouldn't also have the
highest rating. If you score more points per turn than
anyone else, then you are the greatest Scrabble player
on earth. (Congratulations!)
PPT and rating can coexist easily. But while the rating
is for tournament players only and involves a convoluted
calculation, PPT is for Everyman and is simplicity
itself. It shows how your point production improves
over time. It gives the most meaningful basis of
comparison between yourself and other Scrabble players.
It provides a benchmark to gauge the play you just made.
When the Scrabble world embraces PPT let me propose a
few auxiliary statistics. Scrabble points come from two
totally different places, so to speak. There are
"regular" Scrabble points which derive from the tile
values and premiums on the board, and there are "bonus"
points for making bingos. It would be useful and
interesting to keep separate track of them. We might
call regular points per turn "RPT", and bonus points per
turn "BPT". Then your total points per turn, PPT, is
simply the sum of the two: PPT=RPT+BPT. Another very
interesting statistic, Mean Turns Between Bingos (MTBB),
falls out directly from BPT: MTBB=50/BPT. In practice,
you would just keep track of your total points scored,
the turns taken, and the number of bingos played to work
up any of these statistics.
Now, then, what's your PPT?
A few comments on the current state of PPT:
In case the Points Per Turn statistic sounds so obvious
that you have a hard time believing it's not used by
Scrabble players, consider a few observations culled at
the time of writing this, May 2009.
A Google search on - scrabble "points per turn" - turns
up only 62 hits (after Google weeds out "very similar"
pages.) Several of these pages are, in fact,
duplicates. Eleven of the hits are my own writings,
either from my own web site, or pages that draw from
postings to rec.games.board. PPT mentions of any
interest are with respect to a single game, or perhaps a
set of games, never a running statistic indicating a
specific player's skill. There are a couple of mentions
in the book "Everything Scrabble", by Joe Edley, and one
of those was lifted into Hasbro's Scrabble FAQ. It
doesn't appear to get mentioned in everybody's favorite
Scrabble book, "Word Freak", by Stefan Fatsis. For the
record, I calculated the PPT statistic for members of
the Bowie Scrabble Club (Maryland) right from the start
in mid-1985. And discussion of PPT was part of my
original Scrabble page, which went up on the web as soon
as I got internet access in 1997.
Donald Sauter
Current PPT=37.3 (RPT=23.4 BPT=13.9)
Dover Scrabble Club (Delaware)
http://www.geocities.com/donaldsauter/scrabble.htm
davi...@gmail.com
> Your PPT in Scrabble, on the other hand, is almost rock-
> steady no matter what level of competition you find
> yourself in. After all, what difference does it make
> whether the board you are stewing over was cobbled
> together by experts or by duffers? Ironically, if you
> are a recreational player with a 19.6 PPT, say, you
> would probably see it rise a hair against experts, since
> their longer words open up more scoring opportunities
> for you.
>
I disagree with the above assertion that your opponents will not
affect you PPT. I consider defensive play as one of the more
interesting aspects to expert play in Scrabble. It is much easier to
score triple word scores against a duffer; than against someone who
consciously avoids opening up the triples. I also disagree with your
claim that experts tend to play longer words. It has been my
experience, and disappointment, with Scabble that you will score far
more points playing short words in good places then long words in
mediocre places; the exception being Bingos.
I guess it bothers me that this discussion of PPT being used as a
rating will reward players that simply maximize their own points with
no regard for their opponents score. In most two player games it is
considered good play to score a point less if it will mean your
opponent will score 2 fewer points. Though, this comes from my
experience playing 2 player Scrabble games.
Donald Sauter
I'm just as certain that Scrabble is almost 100% an offensive
game. Sure, you don't want to do something stupid like
play BUSH right up against a triple-word score. If I thought
there were more people tuned in to the discussion, I would
gladly run a study on average tiles played per turn of
the experts vs. the masses. Also, number of triple-word
scores hit per game. I have no doubt the experts would
score higher in both categories.
That doesn't mean the experts are always making long words;
far from it. It's just that their bingo production pulls
the average way up. Of course, the reason for these posts
to rec.games. board is my disgust with the huge role tiny
words play in Scrabble. How many decades can a Scrabble
player get excited over playing HO/HA for 28 points without
starting to see the vapidness of it? (And HO and HA now
rank up with the classiest "words" on a tournament
Scrabble board.)
Scoring a point less to make your opponent score two points
less sounds very hypothetical to me. What a Scrabble player
is weighing all the time is, "How many points am I willing to
throw away on this turn in order to save good tiles for a
potential killer on my next turn?" Completely offensive.
Donald Sauter
Donald Sauter
> that you will score far more points playing short words in
> good places then long words in mediocre places; the exception
> being Bingos.
It sounds to me like were both on the exact same wavelength
concerning the sickening importance of short words in
Scrabble. Could I encourage you to give my Scrabble "fixes"
a try? You can visit my various web pages on the subject,
but here they are in a nutshell:
1. No plays consisting solely of two-letter words. They are
for BABIES. They have been DONE TO DEATH. Enough
already. Let's start using our brains. Isn't that why we
sit down at a Scrabble board?
2. Stepped bonuses for long words. If you are reluctant to
step up to the joy of an 8-tile rack, award 10, 30, and 50
points for plays of 5, 6, and 7 tiles, respectively. If you
go with the 8-tile rack, the bonuses are 20, 50, and 80 for
plays of 6, 7, and 8 tiles, respectively.
Those are the two basic rule fixes that will change your
Scrabble life. Now there is a running incentive for playing
long words. Now GOATEE out in the open is worth more than
GOAT on triple-word score. Your boards will *burn* with long
words.
If you can stand even more amazing fireworks, add three outer
rows on each side of the board - to be used for extensions
only! At least one of the newly played tiles on each play
must land within the classic, 15x15 inner board. Now you
aren't constrained to landing on that triple-word score, you
can blast right through it!
To get away from the staleness and skimpiness of the
conventional 100-tile set, mix three sets together and scoop
about 110 tiles from the mixture. At the same time, fix
the I overload by replacing three I's in the 300-tile mixed
set with A E O.
A suggestion for further cutting back Scrabble's two-letter
nonsense is to simply adopt a more sensible two-letter word
list. We use a list of 70 two-letter words based on the
American Heritage dictionary, but with a few rejections
(foriegn AA AE JO PE QI SI; obscure WO; stupid ZA):
AB AD AG AH AI AM AN AR AS AT
AW AX AY BE BI BY DO ED EF EH
EL EM EN ER EX FA GO HA HE HI
HO ID IF IN IS IT LA LI LO MA
ME MI MU MY NO NU OF OH ON OR
OS OW OX PA PI RE SH SO TA TI
TO UH UP US UM UT WE XI YE YO
Isn't there more than enough dumb stuff in that list? Nobody
ever has to memorize it; the list is kept out during play in
perpetuity.
I urge you to find some perceptive, or at least open-minded,
opponents and try some or all of these suggestions, then
report back and let us know how it went. Thanks.
Donald Sauter
http://www.geocities.com/donaldsauter/scrabble.htm