Preface

I’d been on a bit of a hot streak in division 1 for the past few seasons, finishing with winning records for the past 3 seasons. Was I able to maintain that streak?

The Good

Game 1: Lightning Strikes 4 Times

Against JcrON, after opening with QUENA# for 48, on my next turn I played (Q)UENA# again.

Then, on my next two turns, I hooked both instances of QUENA# with S’s.

Unfortunately, on my 5th turn, I could no longer play any QUENA# related words, but with high probability bingoes such as N(I)DATION and O(R)DINaNT# available, but placing a tile in the triple-triple lane with an over 100-point lead, I found my only higher scoring and positionally superior play of the disconnected 9 of I(N)DeNT(i)ON across two double letter scores for 18 additional points.

Game 2: My Best Endgame Ever?

tonyleah just emptied the bag with the high scoring LUZ# for 54, and I’m some deep trouble heading into this endgame.

I’m down 18 points here, and tonyleah is holding AAIPQX. A major problem in this position is that there are not a lot of good places to unload my 4 high scoring tiles. Additionally, I’m holding the C and the V, which are some of the most difficult tiles to play in an endgame. I’d love to play out in 2 turns here to catch tonyleah with either the X or the Q, netting me an extra 16 or 20 points, respectively. To that end, the first play I looked at here was M6 CH(I)V# scoring 20 points. This play gets rid of 3 of my high scoring tiles for a solid score and also blocks the high scoring 7K (T)AX for 37. The problem is that my leave of AFOT doesn’t make an outplay anywhere, and given another turn, tonyleah will offload the Q for at least 26 points, after which I will just be too far behind to win.

Realizing that the left side scoring spot was making it more difficult, I then looked at the play of 5B C(O)V(E)T for just 10 points. This blocks both of the best spots for the Q, and unloads my most unwieldy two letters in one turn. Unfortunately, this play just barely doesn’t win. tonyleah can once again play I8 (T)AX for 37, after which my best play is 13K (R)AF(T) for 29, then the game finishes with 14I QI(S) 14 and O10 OH(S) for 20, and even after collecting 8 points for tonyleah’s left over AP, I lose 482-484.

Going back to the idea of playing out in 2 turns, I realized that I didn’t necessarily have to be able to guarantee an outplay in two spots, because my likely outplays are usually in different places from tonyleah’s best scoring options, so tonyleah would need to sacrifice significant points to stop me from going out. This thought process is what led me to my actual play during the game of 5A AV(O)C(E)T for 11. This scores 1 more point than COVET, but more importantly, by also offloading my A, my leftover tiles threaten to go out with 11G FOH for 19. If tonyleah plays 7K (T)AX for 37, this play wins 475-470 because I’m catching them with AIPQ. Additionally, I’m still keeping some scoring tiles if tonyleah decides to block my outplay for minimal points.

To their credit, tonyleah plays their top endgame move after AV(O)C(E)T of I8 (T)AX for 15, offloading their X and blocking my only outplay.

I still need to be careful here. If I carelessly play something like my top scoring move of OH(S) for 20, saving my F to play 12D IF/FUG for 12 points, tonyleah actually wins by just 1 point by playing M7 Q(I) for 21, with a final score of 468-469. Instead, I played my F for just 9 points with M8 (I)F, blocking the best Q spot and forcing tonyleah to take many fewer points with 14I QI(S) for 14, after which my outplay of D3 HO(C)# [OH(S) is equivalent] for 20 puts me at 455-462, and tonyleah’s left over AP put me just over the finish line, winning by a single point, 463-462.

Game 3: The Q is the Best Tile?

Midway through my game with Taurus777, they’ve just played the massive (T)HEREMIN for 102 to equalize the game.

The most important thing to notice is the massive Q spot that their play created alongside the I in THEREMIN. I could play Q(I) now for a solid 31 points, hoping for one of the 3 Is or the blank from the bag (4/31) for (S)PRAYING and R(E)PAYING. Instead, I made the play of N2 RA(N)GY for 34, leaving NPQ. This play draws 4 tiles instead of just 1, and if I draw an I, I can play J10 QIN# for 65. Additionally, there are 2 Us and 3 As that could help me draw QUA in the same spot, and also 2 Ts available that could give me QAT. Macondo actually loves this choice with such a strong spot available for the Q, though N1 GRA(N)NY for 36 in the same spot for 2 more points and drawing 1 more tile is certainly a bit better.

I drew DEG?, and Taurus777 played O6 AUDIO for 23, dumping vowels.

I’ve drawn none of the letters I need, but I do have the option of burning the blank to play Qi for 62 in that spot. Ultimately, I opted against this because of the unseen pool. I felt that some of the time, AUDIO could be a setup play for either the unseen K or J, which could score upwards of 50 or 80 points, respectively, if Tuarus777 can get a 5 letter word starting with one of those tiles, extending to the double word score. I elected to play N10 PENG#+ for 23 points, leaving DQ?. This takes out that scoring spot, and draws another 4 tiles into the pool that still contains the helpful AAIITTU for my Q. PENG#+ does not take an S (though it does take an O), so I don’t need to be worried about massive J plays hooking it. There’s one more hidden benefit to this play. (10J Q(I) 31 is within 0.2% of this play according to Macondo). Taurus777 plays M3 J(E)U for 41, having just drawn the J (they would’ve had better 50+ point plays like JUVE in the scoring spot had I left it open), and I’ve drawn ELOR.

At first, it looks like I’ve missed all the letters I need to play my Q for a good score, but with J10 Qi for 62 being tempting again, I found my best play of 1L QUoD making the sneaky O(RANGY) hook for a massive 79 points, vaulting me into the lead.

The Bad

A few turns later in that same game against Taurus777, I’m up 39 with only 1 tile remaining in the bag.

The unseen pool is AEEIIKNT. The only thing I saw in the pool was KAINITE, which did not fit on the board. Not trusting that there was no bingo unseen, I figured that it was most likely that if an unseen bingo was in the pool, it was most likely to be from the V or S of VISIT, so I played 2B OO# for 10 in that spot to block. Unfortunately, I was completely wrong. This play throws away 25% winning chances, because with that same rack of KAINITE, if an E is in the bag, Taurus777 can bingo out with AKINETI(C) to win. As such, the 100% winning plays are plays like 14A ORLE 31, blocking the C and scoring extremely well.

I was lucky here that the final tile in the bag was an I, giving Taurus777 no bingoes.

Hopefully, as I continue working on studying bingoes, I’ll make fewer of these pre-endgame errors.

Game 4: How Can I Score?

With 1 in the bag and up 16 points, I’m in a weird spot against shafiquet.

The unseen pool is ADEIIQRS. I have a lot of vowels that are hampering my flexibility to score. Additionally, I need to be careful about what I play, because if the Q is in the bag, shafiquet is threatening an outbingo of DRAISI(N)E# at the base of the board. Even if I block that bingo threat, I also need to make sure I can manage the Q if I draw it. The first play that I looked at was 13C AUA# for 17. This blocks the bingo threat and scores decently, while keeping my C to score with on my next turn. I was worried that if I drew an I after this play, then I would have a tough time scoring to try to outrun shafiquet’s 14F QI(N)# for 32, threatening the high scoring outplay of 2H R(U)EDAS#. This was a miscalculation. I failed to notice that if shafiquet does play 14F QI(N)#, I can actually end the game immediately with 15B OCUlI or COqUI, which easily wins.

Thinking that I needed to stop the QI(N)# threat to avoid occasionally being outrun, I made a poor play of I14 OU# for 13. This blocks the only good spot for the Q (or so I thought), and if I draw the Q myself, I’m threatening massive plays on opposite sides of the board of D1 QU(E)y and O1 QUAy hooking (FOG)y. This wasn’t the worst idea, but again, there is a now a draw that loses me the game, which is surprisingly the R. If I draw the R, shafiquet is holding ADEIIQS, and has just one winning play.

The winning play is O1 IDES for 27. this leaves AIQ, and threatens the sneaky D6 QA(D)I for 27 while also setting up a second place to go out with 2L QAI(D) to the D in IDES. My best response is 6E CURA(R)A for 24, after which QAI(D) beats me by 2 points, 407-409.

My initial play I was considering of AUA# is one of only 3 100% winning plays, while my actual play of OU# wins just 87.5%. Once again, in the actual game, I was lucky and pulled the Q as the final tile en route to an easy win, but I squandered more winning chances.

Game 5: What Am I Doing?

As I have been in many of my games this season, I’m in a close pre-endgame vs bnjy.

I’m down 24 and looking at a pool of AEEIJNRT. The first thing to notice is that no reasonable play will win if I draw the J because bnjy will bingo out with 1D ANTI(M)ERE. The next thing to notice is that bnjy often has high scoring plays in the bottom left using the J of I(N)JERA for 42 and (N)ARTJIE# for 45. Given that I’m already down before their turn, I’m in a tough spot if I let bnjy have these plays. With this in mind, I played 14G STAY for 41, scoring well and blocking (N)ARTJIE#. At the time, I thought this play never won but tied if I drew any of AEI. This was unfortunately completely wrong.

In actuality, this play wins with the E and the I, and loses with all other tiles. If I draw the I, bnjy does not have I(N)JERA available, and their best play is B1 JANE 32, which I can beat with either DALI# to the left of JANE or (C)ALID#.

In the actual game I actually did draw the E, but I did not make the right play. bnjy played their top move of 15A I(N)JERA for 42, leaving me in this position:

After looking for 3:15 for a better outplay, I shrugged and went with the play I had seen when permuting the pre-endgame of B1 LADE for 21, leading to a tie game at 418-418. This is the only game I’ve ever tied in league.

However, I should’ve spent more time here. There was a winning play that I missed. That play is the sneaky 13B (X)ED for 35. The value in this play is that it scores 14 more and with bnjy holding NT, their top outplay is shockingly 3B N(E)T for just 3 points, giving me a 3 point win.

Even though I blew the endgame here, it’s always fun to get a tie.

Going back to the pre-endgame, there was a play that was 12.5% better than STAY, which was 14H DAY for 35. This play also blocks the (N)ARTJIE# threat, and wins with draws of EIR.

If I draw the E, I’m threatening the high scoring O1 TESLA hooking (ZOL)S# which bnjy cannot block without losing to O10 (E)LATES.

If I draw the I, bnjy does not have I(N)JERA and I have outplays of 15A I(N)STAL and N1 TA(L)LIS that bnjy cannot block simultaneously (their best play is B1 JANE).

Finally, if I draw the R, bnjy once again does not have I(N)JERA available and my outplays of O10 (E)STRAL and O3 SLART(S) for 21 and 20, respectively, cannot be blocked.

The Ugly

Games 6 and 7: Ending the Game With a Whimper

In these positions (against kfraley and dragon, respectively), I threw away a total of 108 points of spread by missing my outbingoes (though the games were already decided). Can you do better than me and find them?

Game 8: Common Words are Hard

Against MGI, I’m in a lost position, trailing by 146 points with the bag empty.

I don’t have a playable bingo and I can see that MGI is holding DEEIKT?. During the game I saw that they were threatening 15E KIt(T)ENED as an out bingo. There are multiple ways to block, but I thought I found a decent one with the odd looking 14H (F)AINT, leaving LPS. My ideas was to save the high scoring 8A PL(U)S as an outplay, and MGI will need to sacrifice a lot of points compared to 1L KIsT for 49 hooking s(CURVY) to block, after which I can still play out with 3I SL(O)P. I had the right idea but the wrong execution. With KIt(T)ENED blocked, MGI instead ended the game with the simple 15C TIcKE(T)ED for 68 (15B DIsKET(T)E is also available), netting me a 70 point error. Better for me was to play the same letters with 15D TAIN(T) for only 6 points, which blocks all the bingoes.

Afterword

Despite all these errors, I had another good result, finishing in 3rd place at 9.5-4.5 +120. I had a decent 2.7 MiAOW score. Although I usually try to play a good endgame even when the game is already decided, I think this season really shows that I should be spending just a bit more time to preserve my spread.

Stats

Lost Endgame Points: 209 (7/14 Perfect Endgames, 1 Blown Endgame)

Missed Bingo Turns: 9 (7 Word Knowledge, 2 Playfinding)

Suboptimal Bingo Turns: 7 (1 Word Knowledge, 2 Playfinding, 4 Strategy)

Cumulative Pre-Endgame Win Percentage Loss: 61.6% (9/14 Perfect Pre-Endgames)

(Not counting Monte Carlo Simulation Errors)