I think once you can get the never roll a three option with a few dice manipulation ones you can just be unstoppable, but on it's own it isn't great

exactly

Manipulating rolls is better.

I had my first 2 60K advanced games this weekend. Nice to see the purple.

I like the look of this

My new goal is to get to level 13 and wrest 3 of 5 advanced from TH10 one day.

Zac,

Can you post the nine dice rolls from the 8/16 Daily Challenge?

I think it would be interesting to analyze why one approach (flex strategy) worked over the other (classic strategy).


Flex Strategy: 33,840




Classic Strategy: 31,968 (max of 32,940 if all ten 3s are connected).




The classic strategy yeilded two five-zone chains (6s and 5s) but only a weak three-zone chain of 3s.

The flex strategy also yielded two five-zone chains but had to use 3s instead of 5s. As compensation, the third chain was in four zones.

As a result, the classic strategy got a stronger Chain 2. But, the flex strategy got a much better Chain 3.

I'm wondering if we can figure out some theory on how to decide which strategy to use in different circumstances. What are the variables which go into making the decision?

in

yesterday's DC had me musing about exactly this topic, too, and whilst i'm very far from having a concrete conclusion, i'm definitely interested in pursuing it further.

edit: although the fact that the two styles lead to similar scores means that this specific case might be an interesting starting point but not the most obvious proof

Funny that you mention it, I was having similar thoughts yesterday.

I usually revert to classic strategy when I don't have many early 2s, but I try to delay the choice until the 4th roll if I can. I know there are other factors that should swing the decision, but that's the main one for me.

Here are yesterday's rolls:

55135
51662
61565
36313

23451
33164
56363
12644

24435

i play a classic strategy when i am convinced that the outer chain is going to be one of my 2 key ones, so i want to maximise the chance that it snakes.

but it occurred to me yesterday, that by committing to the classic style, i then actually don't really want a lot of the outer-chain number in the second set of 4 rolls, or else i would have been able to snake it under the flex-mode also and therefore the flex mode would be superior due to the better third chain possibilities.

So perhaps i should decide to classic when i am convinced that the minor-inner chain is going to be crap, rather than i am convinced that the outer chain needs to be good.

i mean, i think that the reason that the flex mode was better yesterday was that there were a lot of 3s and few 5s in the second round of rolls. And there's no way to predict that.

lolpagethreeaments

Today looks like my 22nd DC win
Need a sick TMI read to beat that score imo

Thanks for making me fps and do badly.

I can't blame you I do so ****ing badly every DC that it makes me want to cry.

Can we ban TH10 or something? It's hard to get motivated playing advanced when I know I'll have to hit 63K-65K to even break into the daily top 5.

He just broke the phucking advanced record. Alex is out of the books.

In before TH10 is a bot.

Come on now, that's the wrong attitude. Shouldn't that motivate you to get even better scores yourself? Anyway, now that I got that new record and also average of top10 scores above 70k, I don't think I'll be playing as much advanced games anymore, so getting to daily top5 should be easier. Although I would still like to get at least 73k+, since it's been close quite a few times. Also, you should try finding motivation in getting better scores for your personal top10, after all, the daily top5 scores only last for 24 hours. Maybe this post will help you get to those 65k+ scores more often.

Ok, slowly moving on to the main point of this post. As some have noticed, I've recently played my advanced games with a different strategy, that I feel is more flexible (and thus better) than the normal one, when using only two chains. As there has been interested for it from a few people, this post will introduce how it works. First of all, this isn't any magical strategy, that immediately gives 70k scores, but it should help getting higher scores more consintently, if I'm correct. Obviously a lot depends on luck too. I haven't made any calculations to prove anything, so it is entirely possible that I'm biased in my opinion of feeling it works better than the normal strategy. That said even if it turns out that it isn't better than the normal setup, hopefully this post will be helpful in some other way. Although I figured the strategy out myself, I wouldn't be surprised if someone has used it or a variation of it before, but I would be surprised if anyone has used it with this specific planning. Also, I should mention that I started using the premonition ability at around same time I switched to this strategy, so it might be clouding my judgement on how good this strategy is. I think that certainly with premonition this is better since it has more routes to snake the chains. Without it, I think it should still be better, but I'll let you try and decide. The reason for using premonition is, that I've been playing with the only goal being to get 70k+ scores, and the 1->4 change is very, very rarely useful in trying to get that high scores, premonition let's me always start with ten 1/2/5/6 dice if I'm patient enough, usually I'll start with 9 too, if the only manipulation I'll have to use is re-roll 3, and it doesn't change to 4. This helps save the manipulations later, they will be needed. It is also sometimes helpful when trying to decide whether to use re-roll rack or not. Not sure how much it helps when playing according to the perfect/basic plan (not a lot I would guess), but it does help when using some variations to the basic way of playing the chains/bonuses.

Now the reason I feel this is more flexible is that it has imo better positions for 2s/5s/6s without blocking pretty much any options for the late rounds. With the perfect setup it also needs only three of each 5s and 6s in rounds 5-8 to complete snakes the basic way. I remember a few games with the normal strategy, where everything else went according to plan, but I got only three of the inner chain in rounds 5-8, and wasn't able to snake that chain, that costs a lot of points. In the normal strategy the outer chain can be built even with getting only two of it in rounds 5-8 (with the 3rd one in rack 9), so that is slightly worse with this strategy, but I don't think that is a big difference. On the other hand if we get a lot of the outer chain in a single rack in that normal strategy, we have to block some bonus possibilities, not with this setup. I think 1s are a small weakness, with the perfect setup we shouldn't really play 1s to the middle zone before rack 9 unless there are no other options. Depending on how we play 2s in racks 1-4, there are still 9 or 10 places total for 1s in the outer zones with the perfect/basic chain routes, so it isn't that terrible. The chains can be played in differents ways too, if we need space to play 11-12 1s total (again depending on the 2s were played in racks 1-4), while still leaving the middle zone free of 1s, this will be shown later too. Also it is quite easy to play variations with this setup. I guess that should be enough of an introduction.

First, here's an example of a game, where inner chain didn't complete in a game played the normal way, had it been played with the setup/strategy this post is about, it would've scored 70k+. We can see that 2->5 wasn't used, this was played a long time ago, so I don't know for sure, but I would assume that rack 8 didn't have 2, nor 5. Imo 2->5 shouldn't be used before it is absolutely necessary, since we don't want to get rid of 2s, unless they give us negative bonus or block snakes. That means it is usually available to us in rack 9, where it is needed more often. So imo not using 2->5 in this game was only a mistake if I somehow failed to do it in rack 8, or left 8th rack needing 2 or more 5s, which is a special case of using 2->5 early.





Now let's finally take a look what the new strategy looks like. Below are two games played with the perfect setup and basic chain routes. The difference is the way the 2s are played, the one with four 2s in one zone can support a bonus up to 25, the other one up to 24.





In the next one is the setup after round 4, with some additional markings. The 5 in the left zone should rather be a 6, since we might snake the 5s another way than the basic way, and it would need help from 9th rack in that case, but if the plan is to play the basic way, it is not a problem. In the bottom zone I would suggest to first try the setup, that has been added by mspaint instead of the one that was played, if the plan is to connect the snakes the basic way, as they should be in the beginning, before being more familiar with this setup. The four 2s zone (if used) should be either top or bottom zone. The important spots for the chains are the ones directly outside the middle zone and the top and bottom spots in the middle zone. The left and right spots in the middle zone, which have been circled, are not essential, so they should be the last ones to be played. They need to be there for what I call the perfect setup, and are needed if the snakes are to completed with three dice in racks 5-8.

But without them, the snakes can still be completed with four dice, and one of them can be in rack 9, that's why the last one connected should be the one that can be replaced by 9th rack, if it is played with an imperfect setup, if the 9th rack is needed to connect a snake, obviously there is no space to play that 1 to the middle zone in 9th rack. If we can't get the perfect setup, those two can spots can be replaced by the other chain, 1s or 2s, just have to be aware of the amount of 2s needed to pair all of them in racks 5-9, also imo 2x1 to the middle is better than 1x1. Like I said earlier, it is often the plan to leave the middle zone free of 1s for 9th rack. Depending of how the 2s are played, the 1s in racks 1-4 should be played like the picture/drawings show, only 1 to top and bottom zones, the left and right zone can have two 1s if necessary. All the 1s should be on the same sides of the zones (unless we play 4x2 in one zone or 2x1 in the same zone), so if we stand in the middle square of the whole board, and look towards any zone, the 1 is on our left, red circles in the picture. It might be a decent plan to leave another zone free of 1s too, but why that is, will be looked at later.

The chain that is started with more dice should be played to left and top zones, if we play 4x2 to the bottom zone, that is because only that chain will touch the right square in the top zone with the snakes, and that is a square that isn't a great place for 1 or 2, so it has to be the same as that chain, otherwise it needs to be connected in 9th rack, the same is true for the other chain and the square on the left side of bottom square, when not playing 4x2 in any zone or playing them to top zone. If we are not playing 4x2 in any zone, the board is symmetrical, so it doesn't matter which chain is started first. The decision of playing or not playing 4x2 in one zone has to be made by rack 2, it can be delayed through rack 1, if there are both 5/6 in the first rack. But like I said before, I think it should be better to start by not playing 4x2 in any zone. When not playing 4x2 in any zone, the "sweet spots" (marked ss) are the squares that are "natural" to any of 1/2/5/6, when the chains are played the basic way, so these are really good to leave without dice for as late as possible. Although since we might use variations to playing the chains, it is usually good to connect the extra 5s/6s right away in racks 1-4, if possible. The first 1s in racks 5-8 should be played to zones that can support 3x1 the easiest (top/bottom), so that even a rack with multiple 1s later isn't unplayable. I drew the basic way of playing the snakes too, hopefully the picture is still clear enough.



One good thing with this setup is, that it has a lot of quite easily playable variations. I'll go through the most useful ones, with some of these playing 4x2 is helpful, since the variations of the chains are usually played in the same spots that are normally used for extra 2s, and premonition helps using the variations too. The most minor one is, when one of the chains can't get the perfect setup, in this case it is a good idea to put one of the other chain in the middle zone, this leaves one more possible place for a 1 in either left or right zone, in this case the right. The last 5 that should be played to the middle with the imperfect setup is the one in the top right corner in this picture, since it can be replaced with 9th rack.



Another minor variation is to play one or both chains through other routes to leave more spots for 1s, now played this way they need four dice to complete the snakes, so they aren't as likely, however the 4th one can be played from 9th rack, so that should again be the last one to be played, unless it has already been played before switching. In this picture the 6s have been played an alternative route, leaving one extra spot for 1s in the right zone. It is also easy to see from this, how that 25x bonus is possible, switch the 6 in the top left corner of the right zone to the bottom left corner of the right zone removing the 1, and put 2s in the top left corner of the right zone and top right corner of the middle zone.



In the next one both 5s and 6s have been played different ways than the basic way to allow 11-12x1 (depending of 4x2 or not) in the outer zones, and 12th/13th one to the middle if it is not needed to complete either snake. This game had a severe shortage of 2s, and the 6s had to be connected with rack 9. The right zone shows how an unpairable amount of 2s should be played, this will be explained later more precisely. Also in this one there wasn't another 2 to bottom zone in rack 4, so it looks kind of strange, but it's not that bad, if it happens.



If there are a lot of 1s in racks 1-4, it is probably a good idea to play 2x1 in the middle zone, so that there are enough places to play 1s in later racks, this way we will need four of both chains to complete them. Below are two games that demostrate this (2s played differently). In the first one the perfect 9th rack would've given 74250 points, which is the highest I've ever had a chance to get.

In the second one, everything went well, but it just proves how terrible I am, making a mistake even in my best game ever. In my defense I was simply too excited to even consider making any plays, I just put the dice quickly on their places, when I saw them, just to avoid the game being wiped out of ever happening by a server error. Then when I looked at it after placing all the dice, I immediately noticed my mistake. That was a sick run of dice too, I didn't even have to use re-roll rack, the only challenge was to play the 1s correctly, which didn't prove to be too difficult in the end, since rounds 5-8 didn't have many 1s.





Can you see the mistake I made?

Hint


Answer


This isn't something that should be actively tried, but sometimes there may not be other options. Here is the weirdest variation I've played, going into last rack 18756 points, after 43 dice 24040 points, ended with 60160 points. That's the most points I've gotten with a single rack, I would think most anyone has ever gotten. If I remember correctly, this was actually all-scoring after 7 racks, the 8th rack had 334, I chose to gamble and re-roll it, and it gave three or four 4s, which was terrible.







When using any variation it needs to be decided quite early, so it is a good idea to look at the remaining spots in rack 5, to see how you want to play it. If there are not a lot of spots for 1s after rack 5, it might be a good idea to try make more of them by playing the chains in a different way, obviously also the amount of 5s/6s needed to complete the snakes is important, so see how many are needed, what manipulations are remaining, how many ways to play 2s etc.

It is also possible to make a rack 4 switch to the normal strategy if it is the better choice, like I think it is, if we get only two of either chain total in racks 1-4, that way it needs four more dice in racks 5-8 to complete the snake, instead of five more, although the 5th one can/must (depending of where the two of that chain are) be played with rack 9. As you can see below, with the setup I had in that game, the places of 1s/2s are pretty good for that strategy too, the biggest problem is placing the extra dice of inner chain, if we get too many of them.



The other ways of playing only two dice of one of the chains in racks 1-4 are below, they aren't optimal, but not something that should be avoided at all costs either, so if they happen for some reason, it's not time to panic, nothing is lost yet. The first picture is one of the earlier games, when I hadn't figured out this much yet, so the 1s/2s are played differently in that one, ignore them, focus mostly on the chains. In that the last 5 to be played should be the one to the top right corner of the middle zone, since it can be replaced with 9th rack, but if that is needed, then there is no place for 1 in the middle for last rack. I've never had to use the setup in the second picture, so it is made with mspaint, doesn't look pretty, but should get job done. That is just an example, it shows that playing what is likely 2nd 2 to the right zone isn't terrible, although the 6s and 2s can be played differently too in rack 4, there are many options. The green 5s show that if there are enough 5s, the middle zone can be left free of 1s, and connect the 5s from the left zone in rack 9.





Overall, incomplete setups are not that bad, usually there are ways to get both snakes through, even if the starting positions are not that great.

Playing 221 with the 2s paired and 211 to the same zone are not optimal imo, since they force to leave something stranded to the corner waiting to be connected with the 9th rack. Unfortunately these plays must be done quite often. Now if the goal isn't to score as high as 70k, these are probably a good thing since they give easy places to put extra 3s/4s. Now some might think that leaving a 5/6 stranded in the corner is not a big deal, since we need to get 5s/6s in the last rack anyways. Wrong, how about the times, there are two 5s stranded in corners and the last rack is a normally perfect 66666? Yea, that would be highly annoying. Below is a game where something like this happened, notice that the 2s have been played in a non-optimal way, because there were too many of them in racks 1-4, we will take a closer look at 2s later. Because of the multiple 2s/1s in racks 1-4, I had to leave four dice stranded, still, when you look at the manipulations I had left, I thought there was still a good chance this would be a success. Alas, it was not to be, instead I ended up with five 5s and a failure. Also when trying to play 4x2 in one zone, the 4th in that zone should be a 1, so anything isn't stranded in that zone.



One thing that should definately be avoided is playing two 1s and 55/56/66 in the same outer zone in racks 1-4. This is very problematic, I didn't even realise how bad it was until I made that mistake. I hadn't thought about it before, and had to make a play like that in rack 4, so the first three racks should be played so, that the chances of this happening are minimal. To score 70k with this, we need a lot of the chain that is stranded behind the 1s, and can't even play 2 there to be paired in rack 9, which can be done if we play for example 5551 to a single zone in rack 4, and don't use that zone to any pairing of 2s between two zones. Below is a game that shows this problem, although it must be said that I made a mistake in that game, and should've scored 70k+ anyway. In case someone gets to the same situation, here's what should be done. If I remember correctly, I had to put either 2 or 6 to the left zone in rack 5, so I put the 2 thinking I would get two more 2s in the later racks, but that is the wrong play. Instead I should've delayed connecting the 6 chain and play the unpaired 2 to the bottom left corner of the middle zone and put a 6 to the far left, this way that 2 in the middle could've been paired in rack 9 if there was only one 2 in racks 6-8 (as there was), and that could've been used to pair the one in the top zone, that way I wouldn't have had to leave one 6 out of reach to the top zone.



Playing four 2s to one zone kind of takes away from the flexibility, since we should try to pair those 2s before the other ones (as they can't be paired in rack 9) and that might delay/risk the completion of the snakes. But this is only when using the basic way of completing the snakes, I'd say that with premonition, or maybe even without it and with more experience with this setup, it is better to play four 2s to one zone, since the alternative ways to snake the chains use spots for extra 2s in the basic way of playing the snakes, and it is easier to use the alternative ways with premonition, since they aren't better than the basic way, but to be used when needed. But with the basic way of playing the snakes it might be better to not play four 2s to one zone, since there are plenty of space to play extra 2s anyway, if we don't have to pair too much of them in racks 1-4. Assuming we don't play any 2s in the middle zone in racks 1-4 (can be done if needed), if we get 4x2 in racks 1-4, we can still place 8x2 in rounds 5-8, any extra 2s in rounds 1-4 (have to be paired) each take away 1 place to play 2s in racks 5-8 (down to minimum 4 obviously), if paired correctly. The left and top zone 2s should be paired to left and top spots and in right and bottom zones to right and bottom squares. If they have to be paired in other ways, just have to make sure that they aren't paired the wrong way in both left and bottom, nor both right and top zones at the same time. That way we will still have the same amount of places for 2s available in rounds 5-8, although they take away spots from 1s in that case if the zone where the wrong kind of pairing was made was either left or right zone, thus pairing them wrong should be rather done in top/bottom zone than in right/left zone. Below is a drawing of the worst case scenario, that still leaves 4 spaces open for 2s, here all the 2s are paired in racks 1-4, and they are paired the wrong way in both left and right zones.



One thing I just recently thought about and am not 100% sure about yet is, whether it would actually be a good idea to try leave an unpaired 2 after 8 racks (if there is an unpairable amount of 2s in racks 1-8), since it would improve our chances of getting an all-scoring 1/2/5/6 last rack. Since our plan is to leave the middle zone free of 1s, this should be either the left or top zone depending on how we play other 2s. The way it should be done is to not play 1s to that zone in rounds 1-4, only the 2 to the end of the zone, then in rounds 5-8 we should put 1s on both sides of the 2, that way we don't lose any places for 1s or have a spot we can't score points from. If we get unpairable amount of 2s in rounds 1-8, this is the way they should be played, however this is impossible to know in advance. The right way to do it would be leave the corner spot open for rack 8, so that it is possible to put either 1 or 2 there (or whichever 5/6 connects to the corner if there are no 1s/2s in rack 8). The only problems I see with this are if we get pairable amount of 2s in racks 1-8, we have one less place to put a 1 and having to play two 1s in the same zone in racks 1-4 (which is not terrible, but not optimal either), if there are 4 or more of them. Still, those all-scoring 1/2/5/6 last racks seem to be so hard to get, I think it anything that improves the chance of that happening should be at least seriously considered. I would suggest trying this if there are only three 1s in racks 1-4, I think it should be a good plan in that case.



It is also possible to make a snake charmer with this strategy, and it can be done without having any dice of the 3rd chain in racks 1-4. Another way is to have either one or both of the left and right spots in the middle zone replaced with the 3rd chain.



On a side note, that game gives 770 award points, which is tied for the most I've gotten from a single game. 840 is possible, but it is very difficult, since it has pretty much no flexibility at all, so it needs very precise amounts of everything. I've tried it about 15-20 times, below is the closest I've gotten, this is actually the only time I even got through to rack 9 (would've been complete if the last two dice were 6 and 4/5/6). It should be easy to see from the picture, why this is difficult. Has anyone been able to do it?



I hope I didn't forget anything important. Also, this was probably pretty confusing, and I'm not sure if I explained everything in an understandable way, so if anything is unclear, feel free to ask for a better explanation. And if you decide to give this a try, please feel free to give me feedback on whether it works or I'm crazy for thinking it is better than the normal strategy.

As a repetition, I would suggest to first try the perfect setup with basic chain routes, and no 4x2 zone. And then after some familiarity, to try variations if needed.



These have nothing to do with that advanced strategy, but here are a few other ways to play snake charmer, if anyone happens to be interested. With the first one it is also possible to do the 3rd chain only using racks 5-9. These aren't actual games, they've just been played to demostrate the strategies. I have used these in timed games sometimes, with varied success. I wouldn't recommend them as first choice strategies, but they can be used as backup plans if needed.







And last, just for you Gaby. I've tried to do that normal record too a few times, but I haven't been as lucky with the 9th rack as Popcorn was in his game (don't know if his had more chances to get that high scores). I mean, look at that game, if he could've chosen the 9th rack for that game, it would've been exactly the one he got.



Here are two games, where I could've broken that record with a 9th rack as perfect as he got (60933, 58443). Not sure how much I'll want to try to get that one, since trying to get very high scores in normal mode can be quite tilting. Though, it would be sick to get 60k in normal mode.





Oh, and in before tl;dr.

wow that is an epic post

Wow, that's going to take a while to digest.

I've only scanned TH10's post.

He made several references to ways in which to get very high bonus #s.

However, that shouldn't be a goal when going for a high score. We established in the LoL thread that the optimum bonus was 23. Anything more or less than that is sub-optimal.

To clarify, if you have a bonus of 23 going into the last rack, you are better off getting a 6 to fill any particular square (assuming a five-zone 6s chain) than you are getting a 1 and increasing your bonus to 24.

Likewise, if you have a bonus of 22 going into the last rack, you want to get one 1 to increase your bonus to 23 rather than getting a 6.

I'm a bit surprised the 23 bonus isn't the foundation of advanced play.

TH10, your high scores do motivate me to try to improve. My post was meant as a compliment, albeit a clumsy one.

I appreciate your strategy post and will definitely be going over it thoroughly.

Yes, 23 is the best possible bonus, however it is quite difficult to get exactly that. By looking at that table you posted, you can see that 21, 22, 24 and 25 are very close to optimal, so that is the range we should be aiming for, 21-25. That is the reason why it is good to have room to get the bonus at least up to 24, even better if it is possible to get bonus up to 25, although we would be perfectly happy to only get a bonus of 23. The thing is, we don't want to waste any 1s/2s/5s/6s, since getting a game where we have to play no or just a few 3s/4s doesn't come by all the time, so if that means taking a bonus of 24/25 instead of one/two non-scoring dice, so be it. The foundation of advanced play is to get bonus in the range 21-25 and all other dice 5s/6s, preferably with the bonus exactly at 23, and more 6s than 5s (or no 5s at all).



One more thing regarding 2s. We can pair the 2s between two zones (in this case top and right), instead of using the middle zone to pair them, usually this happens in the later rounds, most likely rack 7/8. The square marked with X should then be 5/6(/1) and the square with black circle can be another 2, it could've been played there earlier and then with 2x2 in a later rack (or vice versa), the 2s can be connected like shown here, leaving the 2 in the middle zone waiting to be paired in rack 9 if it can't be paired before that. In that case we can't play 1 to the middle in 9th rack though, unless of course we get no 2.





Don't know if they are optimal, but here are some guidelines I have for manipulation use, if anyone is interested. Don't start with a rack, where manipulation is needed. If two or more manipulations are needed in the 2nd rack, start a new game (can be frustrating). If a rack has for example 34651, and 3->any hasn't been used, use it first, if it becomes 4, re-roll rack. This is because not counting re-roll rack, there are only up to two ways of getting rid of 4s, and up to five ways of getting rid of 3s, so it is good to use re-roll rack as an extra way of getting rid of 4s. Of course something like a rack of 33335 should be re-rolled too. Save the 1->any, an 2->5 as late as possible, it is usually good to have them available for 9th rack.

Serious bizness itt

dynasty,

i've been a bit busy, but have been musing about the whole charmer vs. dynasty strategy decision in the gaps.

one quick thing that occurred to me. When I have to choose at some given point whether to choose strategy a or strategy b, from the point of view of the DC, it occurs to me that i should choose, not the strategy with the highest expected score, but the one which has the highest probability of being the right choice. is this correct, as you see it? Because I care not so much about maximising my total score, but by maximising the chance that it beats other layouts.

Zomg highest score ever

That is an awesome score, and you didn't even need to use all the manipulations. Since it is that high, now I don't even have worry about my mistake anymore, since it wouldn't be number 1 anyway.

Since you've clearly used that new setup/strategy, any thoughts on it? Now is probably a good time to ask.