- Details
- Written by HG
- Category: Strategy
- Hits: 3952
Recently there was an interesting endgame - can you work out the best play ?
Player A has just played BLOATER and BLUMEN, but because BLUMEN was not allowed , the move came off.
Player B is 11 points behind and knows Player A's rack (BLOATER) and therefore also what the last 2 letters in the bag are B, Y.
What should player B do ? (The blanks make TRIVIAL and INCUBATE).
Let’s summarise the situation. The game is close (only 11 points between A and B).
Player B knows Player A’s tiles (BLOATER) and what is in the bag (BY). From Player A’s perspective the unseen tiles are B, Y, Q, I, R, O, C, E, T and does not know which two are in the bag.
In close games it is normally advisable to play all your tiles before your opponent, so that you gain the point values left on your opponent’s rack, and they also lose that value too. (This process is known as countback).
Player A has better tiles than Player B for playing out in two.
If you don’t want to read the logic of every individual move, there are key learning points at the end of the article.
Player B also has the Q to get rid of, and the tiles in the bag (BY) won’t help playing the Q. Therefore the Q ought to be played now.
Where can the Q be played ?
There are 4 options :
i) QI across in row A, column 3 also playing QIN. This scores 24.
ii) QIN down in row A, column 3. This scores 12.
iii) QIS across in row G, column 2. This scores 12.
iv) QI across in row K, column 3. This scores 11.
Let us look at the merits of each of the 4 moves in turn.
Option 1 – QI.
This move scores 24 and puts player B 13 ahead. But player B will pick up BOTH the B and Y.
If Player A can play the tiles from BLOATER in two moves, then it is very likely that they will win.
One example is
- for Player A to play BOLT row G, column 2 down scoring 18. (Player A is now 5 ahead).
- for player B to play BOYS in row K column 7 down scoring 26. (Player B is now 21 ahead).
- for player A to play AREA in row H column 12 across scoring 12 (Player A is still 9 behind, but player B has CERT left, which will mean Player A will win by 3). REAK (if you know the word wins by more).
Therefore player B should not play QI and empty the bag.
Option 2 – QIN.
This move scores 12 and puts player B 1 point ahead, and leaves one tile in the bag.
From Player A’s perspective the tile in the bag could be a T, E, R, C, I, O, B or Y.
A good move by player A would be BOLTER row G, column 2 down scoring 20. This will leave an A + the tile in the bag. Whatever tile it is from TERCIOBY will play next time. In most cases there are at least 2 alternatives.
Player A, lays BOLTER 20 points, going 19 points ahead, picking up either B or Y to accompany the A.
If player B picked the B from the bag, the rack is TERCIOB, otherwise it is TERCIOY.
If the rack is TERCIOB, the highest scoring move is COBS (K7 down) for 22 (leading by 3 points, but leaving ITER on the rack). Player A only needs play out which can be achieved by playing YA row H column 13 down for 14 points.
If the rack is TERCIOY, the highest scoring move is COYS (K7 down) for 26 (leading by 7 points, but leaving ITER on the rack). Playing BA in row H column 13 will again secure the win.
Thus playing QIN is not a good option either.
Option 3 – QIS
This is a very good move, as it blocks the BOLT/BOLTER moves that score so highly for Player A in the previous two options. This move scores 12 and puts player B 1 point ahead, and leaves one tile in the bag.
From Player A’s perspective the tile in the bag could be a T, E, R, C, I, O, B or Y.
Player A now has to find two places to play 8 tiles (BLOATER + the one tile in the bag). The option of BOLT/BOLTER has now been blocked, and there are two obvious places – onto the K of KATA (perhaps BLEAK or BREAK) or onto the first S of DESIGNS (and example might be BOATS). But whichever option is chosen player B will reply by scoring on the remaining opening, preventing player A from playing out in two moves.
So player A needs to create another opening to enable a playout with whatever tile is drawn from the bag.
Probably the best option is to play LOBE in row F column 13 down scoring 21 (putting A 20 ahead). This will create an opening as the E of LOBE will be played adjacent to the T of PATRONS creating TE. TE is a useful two letter word, as it can be backhooked with A and T both of which we will be retaining.
The letters being retained are RAT and will be joined by one of TERCIOBY.
If the T is picked up TRATS will be play in K7 down or TET/TART will play in column 14.
If the E is picked up, TET/TEAR is the sole place to play out next time
If the R is picked up, we have no play out next time
If the C is picked up, we CART/TEC
If the I is picked up, we have TEA/AIRT
If the O is picked up, we have RIOTS or TET/TRIO
If the B is picked up, we have BRATS
If the Y is picked up, we have TRAYS or TET/TRAY
So for three of the eight letters we have a guaranteed playout, and in another 4 cases we have at least one place to play. This is quite a good % chance, so player A plays LOBE 21.
If player B picks the B from the bag, then player A will always win, as the playout will be either TET/TRAY or TRAYS. (Player B scores COBS for 22, a lead of 2 so any TET/TRAY will easily win.)
However if player B picks the Y, then player A only has one playout BRATS (j7), and can be stopped in numerous ways :
TOYS (k7) for 24 (a lead of 4)
COYS (k7) for 26 (a lead of 6)
TROY (m3) for 23 (a lead of 3)
OYS (l7) for 23 (a lead of 3).
In the majority of these plays, player A can score 24 points by playing ABBA (h12).
For both the TOYS and TROY move, player B is left with CIRE, which player B has TWO places to play next time : TEC/CIRE (i14 down) which scores 13 or CIDER (j2 across) which scores 16. It is this CIDER scenario which will give player B the best result a DRAW :
To summarise all the moves :
Player B QIS 12 (+1) pick up the Y
Player A LOBE 21 (+20) pick up the B
Player B TOYS 24 (+4)
Player A ABBA 24 (+20)
Player B CIDER 16 (-4) with RT left on Player A’s rack, (+2/-2) the final result is a DRAW!
If player A plays BRA (20) k3 rather than ABBA, then player A wins.. but will player A get greedy and go for the higher point scoring move ? ! Sometimes the lower scoring move will enable a win.
Option 4 – QI
This move scores 11, and has much the same logic as QIS (option 3). Player A needs to create a space on the board so will play LOBE. Picking the Y will guarantee a win by playing TRAY or TRAY as detailed above.
Picking the B will also win as player B has to block BRATS, but the reply from player A is ABBA for 24.
This time the space in the bottom left hand corner if the board is restricted because of the QI, and moves like CIDER are impossible.
Some learning points from this exercise :
1. In close games, always try to leave one tile in the bag. Emptying the bag, gives the advantage to your opponent.
2. If your opponent plays an illegal move, write down the letters.. its useful to know what letters they’ve got, or in this case… what’s in the bag!
3. Some moves can be quickly established to be losing moves. Don’t play them. Even if you can’t see how to win with other moves, play those instead. Let your opponent make a mistake…and you are ready to pounce if need be!
4. Sometimes the highest scoring move won’t win the game. Think what you opponent can play, do some simple sums, and play a lower scoring move, if it guarantees the win.
- Details
- Written by HG
- Category: Strategy
- Hits: 4328
Imagine it is your start, and you draw the following rack :
G K G L H T S
What should you do ?
You have three alternatives. Play, Pass or Change. Which is best ?
PLAY : There are only 2 words you can play ST or SH. Both are wasteful of an S. (In general, a single S on your rack in the first half/two-thirds of the game is best used to score at least 30 points and more likely a bonus word.) Therefore PLAY is not a good option.
PASS : What are you hoping to achieve by passing ? Your opponent will do one of the following :
Play .. and therefore give you a vowel. But if the vowel is placed centrally, you may only score 8-11 points next time (e.g THUG or KITH).
Your opponent may change or pass themselves. They know your rack needs something from them, in this case a vowel, so they could and should change to make their rack better. It might be they have a word like QUICK or ZEBRA which is worth playing for over 40 points, but if they can only score 16 points they may as well change and see if they can improve their rack. It follows that a PASS will by you will lead to a ‘defensive’ more from your opponent, so you will gain nothing. Therefore PASSING is not a good option.
CHANGE : The all consonant rack, at any time of the game, generally should suggest CHANGE (Unless you can score over 30 points or play 3-4 consonants). So what to change ?
There are a variety of options we shall consider :
Keep S and change G K G H L T
Keep ST and change G K G L H
Keep L S T and change G K G H
Keep H S T and change G K G L
We have decided on these specific keeps, as there is synergy between the kept tiles. (i.e they make lots of words).
Although 'K' is quite a good scoring tile, it is also a 'hard' tile, there are fewer words that contain it.
Although G is an okay letter, it really needs IN to be ultra useful.
With these facts in mind the GGK are to be changed in every case.
Probably the best way to illustrate what might happen with these various changes is to sample 20 different pickups. (i.e 20 different changes).
Here are the racks if just the S was retained.
S + NOSGME |
S + REROVE |
S + OEIPOA |
S + RRFILL |
S + GYVAFE |
S + VWEAOS |
S + NOLBMX |
S + LUAFOA |
S + PEFISS |
S + ICOYU? |
S + ETOIEL |
S + FSDZUD |
S + PAUIEA |
S + NONI?E |
S + INTDOF |
S + RTTEAP |
S + DAYJET |
S + MAGECH |
S + NCFIIE |
S + APATRH |
Interestingly 4 of these racks yield bonus words (20%) .. can you work put what they are ? (Answers at end)
Despite changing 6 consonants, three racks have one vowels or less (15%).
What if ST was retained ? Again 20 changes were undertaken.
ST + ADEPI |
ST + VPPOE |
ST + AATEE |
ST + NHAOI |
ST + EHVAY |
ST + LCOGI |
ST + TZTYR |
ST + JAOOL |
ST + CBEID |
ST + MOTSM |
ST + AEUJE |
ST + IQITO |
ST + NIDOO |
ST + VEXRO |
ST + EASMD |
ST + TRADO |
ST + IROTA |
ST + NDXAI |
ST + ?AVIY |
ST + OPYET |
3 of these racks yield bonuses (these are a bit harder, so don’t worry if you can’t spot them), and two racks yielded one vowel or less. (10%)
What if STH or STL were retained ? Here we used the same 4 letter changes against both sets of unchanged tiles.
STL or STH + AECI
|
STL or STH + HZOM
|
STL or STH + OUCO
|
STL or STH + ?PGW
|
STL or STH + HOII
|
STL or STH + AECY
|
STL or STH + WDLE
|
STL or STH + RNCO
|
STL or STH + IGIS
|
STL or STH + SNEA
|
STL or STH + NELR
|
STL or STH + HXEU
|
STL or STH + AOID
|
STL or STH + CITT
|
STL or STH + WEAM
|
STL or STH + EEOG
|
STL or STH + HNIE
|
STL or STH + EEIX
|
STL or STH + IEAI
|
STL or STH + UBCF
|
4 of the STL racks will yield bonuses, 3 of the SHT racks yield bonuses. (Which can you spot?)
But, and here is the big difference between these racks and the other two changes, 7 racks have one vowel or less. Thus there is a 35% chance of have only one vowel (or less) next time. This is a considerable jump from the 10-15% on the other two changes we have analysed.
We did not consider changing 3 tiles. Such a change, is more likely than the 4 tile change to yield a consonant only rack next time, and should therefore be avoided.
In summary, if G K G H L T S was your opening rack, it is recommended that you change either six tiles (keeping S) or change 5 (keeping ST). Both have a 15-20% chance of a bonus but more importantly a very high percentage of getting two or more vowels next time.
Please note – the sample size in each case was only 20 racks. Statisticians will argue (correctly) a larger sample size is required, but hopefully the sample size of 20 is sufficient to prove worthwhile.
Bonuses :
S only keep : REVERSO, ETOILES/ESTOILE, ENJOINS/PENSION/TENSION/VENISON + 16 others, SPATTER/PATTERS/TAPSTER
ST keep : ISODONT, DEMASTS, VARSITY/SUAVITY/VASTITY
STL keep : LACIEST/ELASTIC + 3 others, ACETYLS/SCYTALE, SEXTILE, LAITIES
STH keep: TOADISH, ACHIEST/AITCHES, HASTENS/SNATHES/SNEATHS
- Details
- Written by HG
- Category: Strategy
- Hits: 3707
Although the point value of S is one, and the value of a blank is nothing, it well worth considering their worth to your rack and in particular their usefulness in generating a large move score.
Many top players will rarely use an 'S' for less than 30 points during the first half of the game.
Many top players will rarely use a blank for less than 50 points during the first half of the game.
In all likelihood both will form the basis of a bonus word (ie 70+ points).
There are exceptions : perhaps you are holding 3 esses, or maybe a rack with no vowels, or there is a high scoring opportunity that needs to be blocked before your opponent uses it. In general (and probably 80% of the time) the 30/50 values should be considered.
As the game progresses the point worth will diminish as the board becomes more blocked, and letters are used up. However, the value of 'the last' S or blank is a useful weapon, so may be worth retaining for the end-game.
- Details
- Written by HG
- Category: Strategy
- Hits: 4097
Last week, I was told after a particular game…”I had the Q for ages and didn’t know what to do with it”.
The Power Tiles of J,Q,X and Z ought to yield high scores because of their high point value.
But another way is to consider them as, PROBLEM tiles. There are various problem tiles in Scrabble – a V, two Is… and if the JQXZ were added to this list, then it would be more obvious what to do with them.
My ‘gut feel’ for the PROBLEM POWER tiles was that I would play the Q off within 1-2 moves. Similarly for the J. The X and Z may stay on the rack a move or two longer.
I keep records of racks at the tournaments I play in. I discovered, through sampling 50 games I picked the POWER tiles 101 times (broadly in line with probability). 67% of those times I played the power tile on the VERY NEXT MOVE. 21% of the time I played the power tile on the next subsequent move. So 88% of the power tiles I picked up were played in the first 2 opportunities I had.
I also analysed 10 games played at the 2016 British Matchplay Scrabble Championships. (Frequently at this tournament, they have a recorder who writes down every rack and move so that the games can be subsequently analysed using a tool called Quackle). http://www.centrestar.co.uk/tsh/16bmsc/html/annotations.html
Those 10 games would have yielded 40 POWER tiles. (Actually 41 as there was one change and the Q appeared twice in a game).
In 33 occasions out of 41, the POWER tile was played/changed immediately. That is over 80% of the time.
So the top players play off the JQXZ very, very quickly.
But the top players know more words – they can obviously play the tiles quicker !
Here are the words that were played in those 10 games.
Most of the words are everyday English, or a 2, 3 or 4 letter word.
Knowledge of these short words will enable the JQXZ to played off quickly!
Don’t forget though to try to get a score when you play them!
JURA QAT REZ LEX
XI JEWED ZO QI
ZEA FIQUE JOWL DEXY
JAVELINA CHIZ OXEN QAT
NOX QIN JAI REZ
PUTZES TAX QIN JUROR
WAX JIZ change + QI (JIZ included both J and Z)
BISEXUAL ZEA JOE QANAT
RELAXINS JIZ QI (JIZ included both J and Z)
QAT JILT HOX RIZ
- Details
- Written by HG
- Category: Strategy
- Hits: 5348
How many vowels are there in a standard set ? How many consonants ?
Have a little think ? Are there 30 vowels ? 40 ? 50 ?
There are 42 vowels, and 56 consonants. (and 2 blanks).
That is a ratio of 3:4, which is very difficult to work out especially as the letters come out randomly.
One way to consider the ratio is a 50:50 ratio with a slight emphasis to consonants.
Page 4 of 5