# Happy New Game – Round 3

Summary of round 2 can be found here

## Initial Position:

xxxxx6d5c9c8dJd8c/Kd4c/xxxx8hKc8h7h6h5d4c3c2c0s/x3h2hAh6s5s7d/Jd0d9d8d6c/xxx6dKsJc0c9c8sQd/xxxx3h2cAcJs0c2d/xxxKhQhJh0s9s8s6s5s4c3s4s/xxxKs9h/xxx3d2d5h

Checksum: (11 + 2 + 14 + 7 + 5 + 10 + 10 + 14 + 5 + 6) + 2*10 + 0*13 = 104

Monkey Recommends: ic,ac,je,jh

Actual play (date = 19/Jan, score = 459): ed,ef,df,jf,jb → Qh

Spider GM comments: Monkey reasons that even if we refuse the empty column there is still a turnover in column 4 – and hence the same minimum guaranteed turnovers (it’s either one or the other since there is only one Seven and two Sixes).

Monkey Recommends: ic,ac,ce,ji

Actual play (date = 21/Jan, score = 452): ic,ji → 4d

Spider GM comments: A few more good cards and the Dow Jones Happy Star Index could reach stratospheric heights before we know it. Game on!

Monkey Recommends: ae, hd, gh, cg, bj, gc, jf

Actual play (date = 23/Jan, score = 450): bj,jf →7d

Spider GM comments: Not sure if this decision counts as trivial – there are a lot of options to consider. We also need to think about our overall game plan, and let’s not forget to Mind That Stupid Gap in column 8 😊

Monkey Recommends: ac,de,dh

Actual play (date = 25/Jan, score = 448): hd,gh,ac,ai,gi,ce,ci,ei,cg,ig,ai,g1=i1,ca,fc,fe,af,ca,ec,aj → 5h

Spider GM comments: Our card distribution is rapidly improving but the RNG favouring Bart’s choice three times in a row leads to a new problem: the Nasdaq 100 Rigged Index currently stands at 13 since three or more Barts in three trials occurred in exactly 13 trials out of 100.

Monkey Recommends: be,ec,ge,fc,ca ,ji,fe,eg,aj,fe, jg,if,fe,jc,aj

Actual play (date = 27/Jan, score = 423): jg,f5=g5, bc, fb, he, h3=j, ah → 4h

Spider GM comments: IM Bug hasn’t suggested anything, so the Nasdaq 100 Rigged Index remains unchanged.

Monkey Recommends: jd,dj,gc,jd,cg,gc,dj,cg,ah

Actual play (date = 29/Jan, score = 414): ah (trivial)→ 0d

Spider GM comments: IM Bart made an interesting play, sacrificing in-suit builds for the greater good (e.g. not polluting potential empty columns). He also proposed to turnover column 4 if the turnover was useless but fortunately, we have a good card.

Monkey Recommends: jd,gc,cg,ba,dj,ab,ba,jd,gc,ab,ba,ed,cg,he,jh,ab,ij,ai

Actual play (date = 29/Jan, score = 413): ej,(cehceh),(feifei),(jejheh),de,dc→ As

Spider GM comments: At least Monkey remembered to tidy up the in-suit builds, thereby justifying Bart’s earlier judgment.

Actual play (date = 30/Jan, score = 401): dh, (g7=i1), ad → Ad

Spider GM comments: Any Ace is a good card. We have plenty of Twos and exposing an Ace now means less risk of exposing several Aces when dealing a new row.

Monkey Recommends: hb,bi,ib,bi,ah

Actual play (date = 30/Jan, score = 401): ab (Trivial) → 8c

Spider GM comments: And some Aces are better than others! Monkey reasons that our Diamond suit isn’t running away any time soon. That’s technically correct but I’ll let you judge this for yourself 😊

Monkey Recommends: ia,ei,ha,hj,gd,de,dg,cg,bj,gd + jb,bj,jb,dg,gd,bj,jb,bj,bh,dg + jh,hj,jh,hj,hb,bh,gd,dg,gd,jh

Actual play (date = 31/Jan, score = 490): c7=b0, deal

Spider GM comments: Nup, I don’t see any turnovers either ☹

Bart has made a very interesting decision, breaking an in-suit build in order to maximise the chances of winning back the empty column. Not a play I would have considered, but I can’t say it’s terrible. I guess it’s easier to judge the correctness of Bart’s decision after the game is completed 😊

Round 3 summary and Round 4 coming soon to a place near you.

## 29 thoughts on “Happy New Game – Round 3”

1. HAPPY NEW GAME – ROUND 3 – Position 459

.ed,ic,fi,ei,di,ji,jb = T.O col 10

I don’t know if this is good or not but what I do know is that even if it contains errors or is ill advised, I could not have done this four days ago, so hopefully things is improving.

Critical Critiques Cheerfully accepted.

Like

1. Blockchain has been updated, and the Random Number Generator likes Bart’s option 🙂

Like

2. Bart Wright says:

1/19/2022
Score 459

Whine factor up to a respectable 22%, so “whine” is now replaced by “grumble”.

They say “build it and then will come”, but our nesting boxes to attract 7s attracted only one. Similarly we got only one queen.

The obvious way to get a space in column e uses up both that 7 and that queen. If I was allocating only my usual amount of brain power to this problem, I’d start with that. But thinking a little harder…

Turnover candidate columns are a, d, g, and j.

The one column we can expose without using up vital cards is j — but is there other stuff to do first? In terms of using up very scarce resources, column a requires both a receiving queen and jack. d requires a receiving 7, g requires just a queen. That means that if we look only at those scarce resources, instead of using up both of them to get a space we might be able to uncover both d and g by foregoing the space. When I follow that plan, I end up short one three — we need that three to take the 2 in column g, but then we need it again for a steppingstone a little deeper in the column, and (since we have no space) there isn’t one. So that doesn’t work.

So the plan is to uncover j. We can’t just do that without committing to the space, since otherwise the pile-up of 56 can’t be removed. So we’re going to need to move that 6 first. Details? The move ed followed by jd means we can no longer move that 7 in column d, something we’re likely to want to do very soon. So where to move that 7 first? The immediately available home for the 7 is the 8 in column a, but that jack underneath can be a vital steppingstone, especially for column g. Other 8s? Right below the surface are 8s in e and f. We have kings for that queen, but whether to use b or i isn’t clear yet, so we’d rather defer that decision if we can — and if we get two queens soon we might be happy to leave that queen right where it is for some time. Burying the queen of diamonds doesn’t sound too bad. Due to our work last round exposing multiple 8s, the 8 beneath that queen is not of expected great value soon. So here’s a plan:

ed, ef, df (super), jf, jb.

Another plan that also works, and seems just about as good is making c a garbage pile:

ic, ac, dc, ec, jc, jb. The space is available with an ef whenever we want it.

We can then reconceptualize the earlier plan as making column f a garbage pile. I think the first plan makes fewer irreversible commitments, so that’s the one I’ll go with. I think both plans avoid creating obstacles to work on columns d and g, which we are (with luck) hoping to do before long. There are “twisty little passages, all different” that keep me from feeling confident of an analysis here of just how to prepare for clearing column j.

Looks like the monkey is close to my second plan, but makes the grave mistake of junking up column e so it can’t be reclaimed as a space.

Oh, and now it looks like the bug hopped in before me! Let’s examine what he did… More like my 2nd plan and the monkey’s plan than like my first one, but he avoids the monkey’s terrible mistake, and he commits to moving the queen. Well within the range of the twisty little passages I can’t figure out — and no evidence of brain damage there!

Final answer: ed, ef, df (super), jf, jb

Like

1. Blockchain has been updated, and the Random Number Generator likes Bart’s option 🙂

Like

3. Bart Wright says:

Score 452.

Grumble index from 22 to 34. All it took was one queen. Can’t even grumble about it any more.

For the last move, I see monkey’s defense as far as it goes — same number of turnovers. But a space allows more flexibility depending on the card that comes up? The turnover in column 4 might be something that’s not useful. With the plan actually followed, I’m not sure I’m going to use that space now that we’ve uncovered the queen, but at least intuitively I feel like I have more options.

So a queen is pretty good luck (at last). It allows us to get another turnover by moving the queen, and it is a home for a jack, which will surely be useful. I’m not sure it’s useful right away, and I think I’d like to defer the choice about the jack until we see what’s under the queen. Column a is still not such great news, since it requires using up a valuable 7. We can use that queen to dig in column g, which seems the most likely follow-up. But no hurry. Just move the queen and see what happens next. There are two places to move it, b or i. The first requires using up a 5, and the second using up a 10. I see very little to distinguish them. There are 2 5s showing, and while there’s only one 10, there’s another close by in column g. Leaving an atomic 9 over our other king could be handy later if we come up with a queen when there is no space, perhaps. On the other hand, we expect to use up that second 5 right away to uncover the 3 under it. So on balance I’ll choose:

My suggestion: ic, ji.

The monkey’s ce is reversible, so can be ignored. He ends up moving the same 9 I do to put the queen in the same place, but also commits to moving the column a 8, which I do not think we need to do yet.

Like

1. Blockchain has been updated. Random Number Generator has sided with Bart two times in a row, so I might need some scientific experimentation to check if it’s rigged or not 🙂

Like

4. HAPPY NEW GAME – ROUND 3 – Position 452

.bf(super),jb

Get a card while prepping to build out Col 2 as much as possible

Like

1. Blockchain has been updated. Random Number Generator has sided with Bart two times in a row, so I might need some scientific experimentation to check if it’s rigged or not 🙂

Like

5. Bart Wright says:

Score 450

The yawn index went to 49%. While we’ve been pulling out of the land of bad luck, it is possible to have luck that is far better than average — like 90%, but we don’t have that situation yet.

Before the 4 showed up, I had been planning to go back to that column g I had in mind before… hd, gh, gi,bj,gb. But now I think we expose the useful 10 but nothing else there, and we have a major simplification plan that’s worth more than the turnover in g. Preview: Put the 8-j now at the ends of columns a and c onto i. move the 9 with an “ag” to expose the 5 in column a, and do lots of simplifications.

hd, gh, ac, ai, gi, ce, ci, ei, cg. Pause for senility cheat — lay out real cards and do these moves so can more clearly see…
ig(super), ai(super), g(1)-i(1), ca, 3-way-shift: fe, af, ca, ec,
Since it’s so easy to make mistakes, I’d note the top sequences in each pile in order at the end of this are:
2-6 diamonds, 4 clubs, 5-8 hearts, 4-6 spades, empty, 2-6 clubs, 8 diamonds, 2 diamonds, 8-ten clubs, [newly uncovered card]

I originally planned this out to do it BEFORE the jf flip, but I think we can do jf first and still not lose the ability to do that reorganization — at the relevant time we have an extra receiving 5 so we can shift those 234 of diamonds around as needed.

A likely move sequence in the future is fe, af, bc, fb, creating a 2-K diamonds run in a space.

The reorg might not happen or be modified if some new card changes the picture, but for me it helps keep it in mind rather than reinventing it after every card.

It looks like monkey is doing part of the reorg before the turnover. Certainly not crazy moves at all.

Like

1. Blockchain has been updated, and we now have two empty columns!

Like

6. Perhaps moving the Club 4 onto the Heart 5 wasn’t a perfect fit for the lay of the cards.

This time I cove the Heart 5 with the Dia 4

.jf

Like

1. Blockchain has been updated, and we now have two empty columns!

Like

7. Bart Wright says:

Score = 448.

Index to 65%. Can’t complain any more — on this one metric, at least.

Background… We are missing only an ace of diamonds for a complete suit there, which would also free up a space in column b, which would be very nice. In spades we’re missing Q,7,2,A, in hearts 10,4, in clubs Q,7. So putting those suits together is a lower priority.

This 7 allows us to move one of 2 sixes, the ones in column a and column d, because getting rid of the gap in h doesn’t seem like a high priority. Moving the 6 in d and uncovering the last card there is trivial, but keeping in mind the “hole but no turnovers” problem I have learned from the master, and already having 2 spaces, and the desire to simplify things, I suggest going for the turnover in column a. That gives us five (5) whole cards to uncover.

I have two suggestions. First is in the spirit of the simplification plan I was holding in reserve last time, but makes fewer commitments. If I really, really wanted every last tiny fraction of a percent of winning this thing, that’s what I’d do:

hd, gh, ac, ai, ag(super), aj (super).

However, putting off decisions also means every time GM turns over a card it’s a lot less likely there are trivial decisions. So I suggest we do the simplification I had in mind last time (modified slightly, because I think I forgot the need for “fc” before the “3-way shift”).

hd, gh, ac, ai, gi, ce, ci, ei, cg, then
ig(super), ai(super), g(1)-i(1), ca, fc, 3-way-shift: fe, af, ca, ec. Finally uncover with aj.

I hope that if I made a mistake in there in some little way you can figure out what I meant.

Monkey goes for uncovering column d instead.

It looks like you got 13/100 from 1/2 x 1/2 x 1/2. Scientists wouldn’t accept p=.13 as significant. It also should be a two-tailed test, since if “the other guy” (IMBug or the monkey?) had been chosen 3 times in a row you would think that would be equally unlikely, right? It’s post-hockery, the sort of thing that led to the big social sciences replicability crisis. But on the other hand, you can drop the claim of bias and just say you want to even things out so they’ll be fair. 🙂

Now I see that the incentives are a bit different than I thought. Suppose I am convinced that my analysis skill is the best of those competing (not GM himself, but still arrogant on my part). Then putting off decisions is a bad idea. I should have suggested the full reorganization last time!

If column d is exposed we can still do my proposed simplification next time, because that 7 gave us an extra space. Maybe that changes the calculus because we know we can solve the “one hole no cards” problem next time so it’s not really a problem. But I should stick with my original suggestion, uncontaminated by monkey’s suggestion and GM’s comments.

Like

1. Blockchain has been updated: 5 of Hearts turned in column 1

Like

8. This is a bit beyond my abilities right now

I feel like we have to look at what happens after we get a turnover in Col 4.

I want to do a major reorg here but need to convey my thoughts in text.

Move the Dia 7 to Col 1, creating two voids.

Move the Club 4 from Col 2 to Col 4

Cover the Dia K in Col 2 with the Dia Q in Col 6 and reassemble the 6-2 back under the Dia Q-7

Under the Heart Q in Col 9 put the Dia J from Col 1, followed by the 10&9 from Col 3 and the 8&7 that were blocking the Dia J in Col 1.

Move the Dia 7 from Col 9 to Col 1, then swap out 8’s Cols 1 and 9

Take a good look around before proceeding.

Fill one of the two voids with the Club 9&8 from Col 1

Move the Club 4 from Col 4 to Col 1

Move the Dia 6 from Col 1 to Col 9

We have a turnover and one void left.

Sorry I can’t do our normal protocol for moving stuff.

That really tired me out. I will read Esteemed Scholar Bart’s comments later

Like

1. Blockchain has been updated: 5 of Hearts turned in column 1

Like

9. Bart Wright says:

1/27/2022

Score 423

It feels so much more relaxing looking at the layout now, with so much organized.

Two things happened last time to make uncovering column d more appealing compared to uncovering “a” further. First, with our reorg we have left column a atomic, so future uncovers from there are easy. Next, we have an extra 4, so now when we move from column 4 we still use up a 4 it is a less scarce resource. It will be useful to see what that one card is in column d for planning purposes — can it be part of a plan to make a space, or not? But first put the diamonds together to further simplify.

jg, f(5)-g(5), bc, fb. Produces 2-K of diamonds in column b.

Now the problem is that if we uncover column d first, we then have an offsuit 45 in column a with only one space, so we won’t be able to move it. We still can uncover column 4 with one space. So I propose to keep going with column a for that reason. We will produce a 45 in a space, and we would very much like it to be atomic, so the first moves are to insure that. H is the column that will get partially “junkified” to serve this end.

he, h(3)-j, ah(uncover), ch, jh, eh. Cleaned up after turnover in a.

I wanted to have 4 of spades and clubs in columns c and h reversed, but my brain wasn’t quite up to figuring out how to do that. It is a tiny difference.

dc, d(2)-e, dc(2nd uncover)

Final solution below. Obviously the first uncovered card could change plans. I leave GM broad discretion in deciding something is obvious.

jg, f(5)-g(5), bc, fb.
he, h(3)-j, ah(uncover), ch, jh, eh.
dc, d(2)-e, dc(2nd uncover)

I can’t see much merit in monkey’s suggestion, and I found it quite hard to follow.

Like

1. Blockchain has been updated. I’m not sure if the change in circumstances justifies asking Bart to confirm the “2nd uncover” or to reassess the situation.

Like

10. Bart Wright says:

Score 413

We did get one good card, and that’s great, though I assume you didn’t mean to say the following 10 is a good card. I would say things are really looking up quite a bit! Strangeness index stuck at an unremarkable 47%.

The four of hearts means the only heart we are missing for a suit is the 10, but I am reluctant to get aggressive about putting the hearts we have together. That gap in h that GM has reminded us of several times is a sort of “virtual 7” that absorbs a 6, and digging there loses us ground.

There is no merit in some dramatic move putting some king in a space. I think things are looking good enough that we want to play it safe. Chances are good that good cards will show up.

We have uncovered 3 of the 5 cards in column a already. I think it’s time to uncover column 4, because knowing what the options are for regaining an empty column is good as we get to a deal (which might happen right now), and if there’s an unturned card there we don’t know. We dislike uncovering an ace, but we do have 3 different 4s to handle that 3 of hearts. A side benefit is creating the A-8 of hearts. In comparison to uncovering column a, this leaves us with a 6 in a space instead of a 10, with slightly better chances of recovering a space as 7s are expected over Js in ratio 5-3.

Along with a 6 in a space, it leaves atomic units on top of kings in columns c and f, which may help with expected queens. h is for now a certified junk column, but it is the least bad place to leave odds and ends.

No news from the bug (I hope recovering proceeds well!). I think the monkey is comprehensible maybe to the programmer who created it, but as for the rest of us…

Like

1. Esteemed Scholar Bart I have decided that this is your hand to play and I wish just to comment occasionally. So don’t wait on me.

I just am not capable of doing the deep mental gymnastics required at this point in time.

Master Chi-Yuen’s comment on Esteemed Scholar Bart’s move to give up an in suit build for an Atomic is something that I have been using more and more, and I think with good results.

Like

2. Blockchain has been updated. At least the choice of which column to turn over is not difficult 🙂

Like

11. Bart Wright says:

Score 401

I guess there is a bit of tweaking that can be done, though I thought maybe what we had was good enough. But you didn’t judge it as “obvious”, so… we would rather have column g relatively clean compared to column i, because with a Q and 9 showing up (for instance) we could uncover cards in g without putting a king in a space. So that’s my goal here. Ideally we’d match up the 10 of diamonds in “a” with the jack in i, but I don’t see how to do it without breaking up the 9T of clubs. I’ve changed what’s hanging off the jack in h from 4 pieces to 3 pieces, so it’s a very minor gain. I can’t see how to do better without messing up some other column we care about more. I have this feeling there’s something more important you think should be done, but if so I’m not seeing it.

I don’t know if “dh, g(7)-i(1), ad” would be explicit enough, but I’m sure you could figure out what I meant.

I agree any ace is a good card here. They also might be relatively good cards if they show up in the deal too as we will still have at least one receiving 2 for them. Better than K, T, 8, 5, 4, or 2 for instance. Of course if lots of any rank show up in the deal they could give us indigestion.

Like

1. The blockchain has been updated and an eventful round 3 is about to draw to a close. We have removed the Diamond suit, but at the risk of sounding like a broken record, we have one hole no card :-O

Like

12. Bart Wright says:

Score 490

We have one hole but no card, but I think we’re in pretty good shape. We will have 5 potential spaces occupied by atomic units. If we were back in the Kung Fu TV show, you can just imagine the master saying, “Patience” with a beatific smile.

We’re doing well, but one hazard is if we get a deal where we “can’t get the engine started” — for instance being unable to get a space. I’ve certainly been in situations where all kinds of good things could happen, if only I could get that first space. The way to minimize the chances of that in my view is to put something in the space that has the largest number of places it can go. If we were allowed to deal with an empty space, that would be our top choice (I saw somewhere an online variant where in fact that was an option — actual spider’s restriction in that regard is unnatural if you were designing a game and seeking elegance.)

“cb, deal”, where “cb” means the A-7 of hearts. (GM, if your notation system had a concise way of indicating how many of an in-suit run go into a space, I don’t remember what it is.) There are then four separate 8s that that A-7 could go on as part of our getting a space or two. Nearly as good is moving sequences headed by “3” in columns c or i. If forced to do something irreversible, I would do “fb” (moving from the jack down). That exposes a king for a likely queen that comes up. Getting one space somewhere would let you do that as a second step. The “fb” move is much worse in my view. A bonus is that the “head card” is different for each of our five atomic units in spaces.

Could you add some code to your monkey so it recognizes past arrangements of cards and never suggests something that reconstructs exactly an earlier state? This is standard in the relatively simple graph-walking applications I’ve written.

Final answer: “cb, deal”, where what we move from c is the A-7 of hearts.

Like

1. Round 3 is now complete. Summary and next round coming soon

Like

13. May I offer the suggestion that Col 7 is the only place we can get a turnover without filling a void with a K?

Perhaps give up the Atomic Void in Col 4 for an Almost Atomic Turnover in Col 7?

Like

1. My gut says Bart will disagree but you never know. Can you give an exact move sequence?

Like

2. Bart Wright says:

It’s wonderful to know you’re looking these things over and making suggestions! In this case you’re a bit late to the party, as GM already said the round is over — but as for the timing I would make an exception if he would. But I don’t think you can actually get a turnover. You can be left with just the A23 in that column 7. So I don’t think it’s worth making a potential space non-atomic just for that, as GM predicted.

Like

1. Master Chi-Yuen and Esteemed Scholar Bart thanks for responding. I wish not to be an influence on what you are doing on this hand, but still wish to comment occasionally. That is why I have refrained from stating an exact move sequence.

Here I wanted to point out that we only have one Col. with potential TurnOvers and without a K. and that I would give some thought to improving that Col. for future TurnOvers should we fail to get a void.

Esteemed Scholar Bart you are correct in that I can only reduce Col 7 to A23 and cannot achieve a TurnOver.

Like