Strategy?

[Becki M.]

Is this game mainly luck of the draw or is there some strategy I am missing? I have played maybe a hundred games over the last couple of weeks (training only) and most games by now I may lose occasionally but I win pretty consistently but this game is completely random. I think I am doing pretty good and then look at the score and realize the bot still has twice as many points as me. Sometimes I have absolutely nothing to put down and the bot is snatching all my discards and making huge paths. Other times I have lots of stones but no paths, or I concentrate on paths and still lose. I've tried making lots of shorter paths, one or two really long paths, more wishing stones, and it doesn't really seem to matter which strategy I use. What am I missing?

[Stuart L.]

G'day, Becki.   

It's probably hard to say without any examples to study...

I presume you know that paths can start either high or low?  That you need at least a couple of stones to get out of the negative zone?  That you usually shouldn't bother starting a path until you're confident that it can be at least four steps long (but sometimes it may be better taking a small hit than gifting the opponent discards)?  That you can repeat numbers on paths (up to three times, including grey)?  That there are only one of each of the three highest and three lowest numbers and two of everything other colour (one of each of grey) but that thirty cards are removed randomly from the game before starting?  Where the bonus extra-two-points breaks are (at 4, 7, and 9 cards)?  That more than 9 cards of a colour are pointless (except to avoid an only-other-option scenario of gifting discards)?  That you can potentially end the game prematurely if you're ahead by finishing a fifth path with an end card?

The following is a chart of how many points the stones are actually worth:

1st: 3
2nd: 1
3rd: 4
4th: 2

5th: 4 (x2=8 effectively, as this crucial one is a swing, capping opponent's max score from 14 to only 10)

6th: effectively 2 (because it caps your opponents max possible score down from 10 to 8 points)
7th: effectively 4 (same reason ... from 8 to 4)
8th: effectively 1 (same reason ... from 4 to 3)
9th: effectively 3 (same reason ... from 3 to 0)

Equalling or bettering the opponents stone count can be very important because for that mid point, scores aren't just a plus for you but rather a swing of points, from them to you.

Do you have any game examples that you can share?

If we're ever online at the same time, I'd be happy to play you, give pointers.

(Feel free to add me to your friends list, if you like.)

I hope that at least some of this might actually help a little, anyway.

Best regards,

Stu.

[Becki M.]

That does actually help some -- especially the "swing" part. I am not really an "analyzer" as far as figuring out the best strategy. I just know that generally I will do the training mode multiple times until I feel I have a fairly intuitive idea of what I need to do and can win at least half the time. I have done that with other games (Siberia, Cartographer, LevelX, Glastonbury) and don't win every time but enough that I feel confident in my game play. For Keltis, even when I think I am doing well (lots of paired stones, long paths) I look at the score and I still lost. I keep trying different things but don't seem to feel that I'm "getting it". I understand not starting a path unless you can make at least 4, but a lot of times I start out with nothing to play so do I keep discarding (and then the bot has all the tiles) or place a beginning path hoping to get something to play on it? I've done that and sometimes it works and sometimes it backfires. I've tried putting my pairs out immediately and other times holding them back in case I need them for paths. I try not to put my gray ones on the pad but use them for paths but sometimes I have no alternative plays.
     I appreciate the help. I don't play against other players (I've had bad experiences in other online games) but may try as I think it would be more useful. How do I add you?

[Stuart L.]

G'day, Becki.   

Thanks for the reply (and for the kind words).  I'm not the strongest of analysers necessarily either, it's taken me a while of playing this game to figure out all that stuff I told you.

Yeah, playing pairs of cards for stones (and as early as possible, before your opponent gets the chance to) is usually the better play, particularly as it lets you draw two new cards instead of just one.  These could be one or both previous discards (to improve your paths or possibly even claim another pair) or from the draw pile to increase your chances of getting more pairs of cards for even more stones.  Avoid as much as possible, particularly at least earlier on, playing on your paths numbers for which stones have not yet been claimed and especially avoid discarding them (which will not only hinder your chances of stones, but increase your opponent's chances of claiming them).

Yeah, there does have to be a bit of risk involved when you don't have the greatest of hands, but not as much as there is in Lost Cities.  Playing the grey cards on the grey stones for one point each is often the best option you may have (but generally, the longer you can try to hold onto them for the chance of increasing your path lengths to those crucial bonus-point lengths, the better it may turn out, as long as you're not otherwise gifting your opponent points through discards (s)he'd want).

Starting a path even if it doesn't make it to four cards can actually sometimes be an okay choice, because, even if you do end up with a small hit of negative points, doing that can buy you a bit of time to draw more cards which, if you are lucky enough to draw the right cards, may mean you get to play more 'in between' cards on other paths, making those longer, and potentially making up some of (or all of, or even more than) the difference with extra points on those paths (all the while helping you to avoid gifting your opponent points or, even worse, swing points).

If you click on my name to get to my forum profile, there's an Add to Buddy List option on the left.  Given you asked, I've taken the liberty of adding to you to mine, which should make it even easier for you to do the same.  If you've changed your mind though, that's okay, I won't be offended, don't worry.

Thanks again and have a great weekend,

Stu.

[Jimmy V.]

Great points Stuart. Thx so much for your input!

Becky thx for raising the question. Very interesting. Maybe you could share the URL to one of your games (replay link) and let others comment on it. It would be a great way to use the feature. Everyone would learn.

[Becki M.]

https://www.happymeeple.com/upload/game/game_5675115_351156.jpg

Not sure how to do a replay link (can you do that with training games?) but here is a screenshot of a recent one. I thought I had some pretty good paths but still end up running out of time (could have added the green end but needed to add two to the blue path). The big difference is pairs but I didn't have any without using the gray stones which I've found is bad strategy since the other person immediately takes them (of course!).

It's fine that you added me -- I think I figured it out but maybe not. Thanks for all your help and advice.

[Stuart L.]

Thanks, Jimmy, and you're welcome.

Becki, hey 19 to 22, that's not bad!   

Getting those stones is important to the point where it's probably worth letting your opponent have any grey cards if that's the only option.  Remember, you also then draw two cards instead of one, which can put you in a stronger position straight away.

Have you tried Keltis Ór?  It's good too.

Okay, it looks like you still haven't figured it out (?) but anyway, hope to see you around. 

[Mike A.]

Unlike most games, endgame isn't overly important. Luck of the draw is huge, which is rare for this brilliant game designer. If you can't pair numbers for stones very early, you're probably going to lose. It's maddening. Keltis Or is far superior, IMO.

[Meri R.]

I agree with Mike, Keltis Card is one of the most frustrating games in terms of strategy. The stone-pairing is so heavily weighted, and luck of the draw, especially in the first 5-10 pulls after the initial spread, is close to impossible to counter. Your goal, as others have stated, is to hit 4-5 stones and then a couple of runs of 7+ cards and no less than 4 on any run you start, but that's it really, and if the draw doesn't favour you, then you're done, often by mid-game. Keltis Or has the luck of dice rolls but there's more scope in the manipulation of those, and it's a game that's more about what you can steal from your opponent at the right time rather than building high runs.

Win percentage is the best indicator of how important strategy is. Hanamikoji is by far the most strategy dependent, it's almost chess-like in how you have to defend foremost and then have four or five potential winning moves in your head based on how the cards fall, there's very little luck involved if you play it logically. Everything else (not including Carto, which I don't really play and can't comment on) is about incremental gains. If you can hit 60%+ win percentage on a game then it usually means you're strategising extremely well. I'm at 78% for Hanamikoji but between 50%-60% for everything else, no matter how often I play those games. I'm actually higher in Keltis Card then Keltis Or, but in terms of the latter that's not the game, it's because I'm not very good at it, and that will hopefully improve the more I play; I doubt my win rate at the former will ever get much better than around 55% regardless. That's just how it is with some games.