User:Wyriel

Sailing Advice

Wyriel

Viridian

.

.

Advice:

1. Build Links First
2. Build Normal clears only to
   1. cleanup
   2. link platforms
   3. setup, or extend, a starting trigger
3. Preserve Access
4. Fill, and Drop, Vertically
   1. Fill the left or right column of a 2x2, drop into the other one.
   2. Horizontally fill the bottom of a 3x2, reducing to 2x2.
   3. Don't wait for double-color pieces to fill triangle platforms, unless your 100% positive you have the time/space.
   4. Often leave 2 spaces unfilled in a platform instead of just 1.
5. Drop the Wrong Color (and then fix it)
6. Obstacles are worth just as much as Platforms
   1. Don't feel bad about getting rid of them
   2. Try to combo them if it's easy to do so
7. Avoid having the same color in 4 consecutive columns
   1. pay extreme attention to this situation if you have to make it
   2. ... or don't finish off the board so you can cleanup should a horizontal clear mess everything up

Meta-advice:

1. Pay attention to grapple
2. Abandon to modify standing faster
   1. Also to help with 1).
   2. If your doing well

      ...and your not doing well until well after the first league

      ...no matter what the indicator or DR says.

3. Stay on station after figurative or literal booches. Same thing as 2.
4. The indicator lies
5. The DR lies
6. Know where the next piece goes
7. Remember what colors/pieces your waiting for
8. If you could place a piece in 1 second, play it slow instead

   and think about the next piece

   and what waste needs cleaning up

   and what pieces are needed to ensure the combo works

   and where horiz. clears might accidentally occur

   ...

   and then drop it

   but don't take longer than 10 seconds

Counter-advice: (contradicts in-game advice or other players' advice)

1. Don't fill up platforms first (Build links first)
2. Don't fill up from the bottom (Build links first)

   If you can't build a link, then consider filling:

   1. Difficult spots (underneath obstacles, poor access...)
   2. Link-enabling spots
   3. Lower down vs higher up

      ...because lower down *can* become difficult eventually.

      However, less priority than links, a), and b).

3. Starting combos with platforms is fine
4. Play slowly
   1. no bonus for fast play
   2. fast play == less efficient
      1. obvious errors like misflips, accidental horiz. clears, etc.
      2. subtle errors like 4:42 in Geologist's video MOAR SAILEZ
   3. except if grapple is coming up
      1. and you can clear the combo by playing fast
      2. otherwise just bite the bullet and play 1 mv every 10 sec.
      3. or straight-up abandon if you can get away with it
5. Do not break a combo (just) because the indicator starts flagging
   1. finish building it the way you intended
   2. in general, never panic
6. Simultaneously clearing non-platform balls is worth a small bonus
   1. if it's no trouble, clear 5,6,7 instead of 4s (or 8, 9, 10, ...)
   2. don't try to separate cleanup clears into independent steps
   3. but also don't waste time trying to make simultaneity happen
   4. do try to separate normal clears at the beginning, of course

Note that if you follow 5) you will get an excellent on the next duty report in normal pillaging scenarios. This is not an issue standings-wise, and is only a ship performance issue if your the only sailor in the middle of a battle. In atlantis/flotilla/blockade/etc., following 5) will ensure you maintain top incredible. That is a restatement of the prior sentence: DRs don't lie when they are accumulated over a long period of time (like 10 minutes).

Technical stuff:

1. First-league performance
   1. efficient combo in the first 8 moves (source for 8: Shuranthae)
   2. drive excess balls on screen down to 0-2, frequently
   3. no normal clears: just platforms/obstacles
      1. don't even bother with a starting trigger -- takes time
      2. unless, of course, it's the only way to get rid of the pieces
2. Do not deliberately add normal clears to the end of a combo
   1. when you could get away with not doing it
   2. cleanup is a different matter
3. If not board-clearing, don't aggressively cleanup
   1. Use excess as starting material for the next combo
   2. Abandon with only 1 platform left (and little waste)?

      I don't do this, but it could be worth it.

4. 1 move every 10 seconds (source for 10: numerous forum posts)
   1. idle penalty assessed every 10 seconds
   2. probably the penalty is equal to 1 move
   3. could also be a constant penalty to points (for a 3 minute window)
5. 3 minute window in 10 second periods (source 3: forum, myself)
6. 30 sessions go into computing standing (source for 30: Jeda and alts in SW)
   1. possible to entirely reset standing in about 1-2 hours
   2. especially if you can cut board resets out of the equation
   3. ... e.g., swabbied ship, blockades, flotilla
   4. ... but not usually atlantis, since navvers leave so often
7. Step multiplier increases without bound (source: myself, c.f. "70 minutes of sailing")

   (last step of a Vegas^2 is a multiplier of 8 on that step)

8. Approx. weights:
   1. normal ball = 1

      (everything is relative to this; may as well pick 1)

   2. extra ball = 2 (5 in a row worth 6)

      (c.f. "70 minutes of sailing")

   3. target ball = 3 (or 4, maybe even 5)

      (compare "70 minutes of sailing" with "#1 Sailing")

   4. obstacle = 3 (or 4, maybe even 5)

      (Subtle. Junk-clearing obstacles doesn't tank the overall average; see any ult video on youtube.)

9. Efficiency is Points / Moves

   (regardless of whether or not you use the move)

10. DR is average of Efficiency sampled every 10 seconds

    (since the last DR, maybe with small overlap)

11. Indicator is the current Efficiency sample
12. Standing is mysterious
    1. Child's theory is that:

       Efficiency is computed once for the whole session

    2. Could also be 1 huge DR

       (average of Efficiency samples over the whole session)

    3. Scores are submitted to standing computation
       1. Grapple
       2. Touching land
       3. Leaving blockade/atlantis/flotilla board
13. Abandoning probably carries a small penalty (c.f. bilge)
    1. but retaking doesn't forget your Point and Move history
    2. which is way better than a cold start
14. Pausing
    1. Does not forget your move history for contests: no exploit there
    2. Does forget your move, and point, history for the purposes of DR
    3. But I think this is only to inform your captain that you were lazing
    4. That is, it literally acts like a real pause for the purposes of standing (and contests)
15. Contests/Competitions
    1. I, and Jeda, (and others...) have experimented with exploitation strategies that flopped.
    2. Best performance seemed to be from completely normal puzzling
    3. Jeda's theory: submission is what your session score would be if you abandoned at that DR

       Which is very different from your DR score, and is dramatically less exploitable.

       However, super-long sessions of sailing/puzzling tends towards the average, not maximum.

    4. Conclusion: Get warmed up on the puzzle and abandon frequently.

Notation

transcription format:
(3:12)........3n10
.3n10
.4n1o
..6n2o
.2n2o
...5n
.......3n1o
..3n1o
..5n
.............4n
..............(4:51)4n6p3p14n8p6p15n2o4n(4:56)

104 seconds; 11 period; 16 seconds shy of 2 minutes

windowed transcription format:
(3:12)...#.....3n10
.3n10
#.4n1o
..6n2o
.2n2o
...5n#
.......3n1o#
..3n1o
..5n
..#....#....#...4n
etc.
problem:pausing, esp. DR pausing.

calculation format 1 (last column gets computed from first two):
8: 3n1o 6
9: 3n1o 6 
10: 4n1o 7
12: 6n2o 14
13: 2n2o 8
16: 5n 6
23: 3n1o 6
25: 3n1o 6
27: 5n 6
40: 4n 4
54: 4n6p3p14n8p6p15n2o4n 647

format 2:
54: 4n 6p 3p 14n 8p 6p 15n2o 4n
format 3:
12: 4n2e2o
54: 4n 6p 3p 4n10e 8p 6p 4n11e2o 4n
format 4:
54: 1*4n  2*6p  3*3p  4*4n10e  5*8p  6*6p  7*4n11e2o  8*4n
format 5:
    1  2  3    4   5  6     7    8
54: 4n 6p 3p 4n10e 8p 6p 4n11e2o 4n
format 6:
54: (20*4)n (40+77)e (14)o (12+9+40+36)p
format 7:
54: 20*4, 40 77, 14, 12 9 40 36
format 8:
54: 4*(1 4 7 8), 40 77, 14, 12 9 40 36

weights:
1 1 1 2 (existing)
1 2 2 2 .
1 2 2 3 ,
1 2 2 4 ,
1 2 2 5 X
1 2 3 2 .
1 2 3 3
1 2 3 4
1 2 3 5 X
1 2 4 2 .
1 2 4 3
1 2 4 4
1 2 4 5 X
1 2 5 2 .
1 2 5 3
1 2 5 4
1 2 5 5 X

X - platforms being worth 5 is probably too dominant
. - platforms being worth the same as extra balls is very unlikely;
    also, platforms being worth only double normal balls is somewhat unlikely
, - obstacles being worth the same as extra balls is also unlikely
    albeit, they come as free moves so they get an additional boost anyways to eff.

"reasonable" weights:

• 1 2 3 3

  simple (1,2,3)

• 1 2 3 4
• 1 2 4 3
• 1 2 4 4

  simple (1,2,4)

• 1 2 5 3
• 1 2 5 4

Obstacles are more difficult than platforms, so perhaps their weight is higher, but at the same time, they are "free" points. I.e., 4n and 3n1o are both going to take 2 moves, but not only is the 3n1o worth more points, but also there will be an extra, potentially useful, ball lying around. Or consider 4n vs 4n1o; even if obstacles are only worth 1 the latter is 25% better (4/2 vs 5/2).

Obstacles are a lot more difficult though. One cannot move through them like unfilled platforms, their placement rules are far less restrictive, and indeed they can completely block access. For that matter they can redundantly block very useful corridors, e.g., redundantly block drop paths between platforms. Nigh impossible to incorporate doubled off-color obstacles between platforms, as that requires stacking 6 more balls on top, "even more impossible" to turn this situation into a link between platforms.

Musings on Sailing and Scoring

Puzzle scoring is about averages. Not summation. Do a bunch of low value moves to clear up space -- tank your average.

That having been said, one should still find suitably fast ways to clear up enough room to get high scores -- just be aware that such moves should be considered 'painful'.

The puzzles are about thinking on ones feet. Doing something repeatable, mechanical, scientific, etc., goes less rewarded. Being creative and imaginative, adapting to the board at end, without having to sit for a minute thinking about it == mad bonus.

There are 1) raw points == combos 2) raw points averaged in some fashion == indicator 3) raw points averaged in some fashion and subsequently summed == DR 4) standing: some calculation on 3). Perhaps something as simple as continuing the summation for the entire puzzling session (but this allows one to sidestep the consistency thing (maybe)).

Theories about idle: 1) Spending longer than 10 seconds counts as making a move even though you didn't. [and then one has to play the move one is on] 2) Constant idle penalty, perhaps even as nasty as constant penalty straight to the efficiency, or even nastier and straight penalty to the performance level. Probably more like a -20 raw points thing.

Probably both 1) and 2), but pausing eliminates 2).

Theories about averaging: 1) Can't be over time -- playing faster would be a huge win, and it isn't. 2) Can't be over just moves -- Nemo: "All puzzles are played better faster" 3) Perhaps moves made plus a constant? So if you go faster, the constant has less effect?

Actually, it could be 2, but nominal analysis assumes perfect play. An imperfect player by a constant, not percentage, playing twice as fast, does better. That is, some player who makes a combo that doesn't use every ball given -- has 5-10 balls lying around afterwards. But suppose the next combo doesn't increase that, it just leaves it at 5-10 balls lying around (presumably different balls). One can think of it as the same 5-10 balls lying around as dead weight -- analogous to the idea of 3). Playing twice as fast (with all else remaining the same) halves the effect of those 5-10 balls.

Theories about factors: 1) Extra balls in a break (bilge style combos) -- I think these are beneficial, not just in a tactical sense, but in a raw scoring sense. Being able to get to 40s and 20s on viridian with my old style of sailing (only normal clears!) doesn't make any sense unless simultaneous clears are giving a boost. I think it is small and hard to notice without a multiplier like x5,6,7,8, being applied. Even then unclear ;). 2) Obstacle blocks. These are worth a bunch. Clearing these out with a bunch of lame doubles/triples (or even singles) does not tank ones score nearly as fast as normal clears do. Singly clearing them takes 2 moves at least (one of a double color), i.e., just as many as a normal clear does -- the extra 1 block is not helpful for making the break happen. Generally what happens is that one of the balls that was dropped just becomes a new 'free' block on the board, replacing the old one. It follows that if one is dividing by the same denominator but not getting the same effect, then the numerator is larger. 3) Booching boards. Maybe hurts score (above and beyond the moves which can no longer be cashed in)? I think its a constant score penalty so I don't notice it, as I play high score high moves, for which a constant has less effect. 4) Abandoning boards. Doesn't appear to hurt, at least as long as one is not suffering from first league woes. So if you sail for a long time, holding station, get grappled, and then after combat start with sparkly, its enough to build big combo and abandon -- that big combo is not helping the current contribution to standing, it is negating first league woes when you retake station. The combos before the grapple are what is contributing to the standing update at the time of the abandon. In fact, it appears to help one's standing (but no difference to ship performance) 5) No partial credit (like in carp).

Mathy thing. The maximum of x*y subject to the constraint x + y <= c is (c/2)^2 (setting x and y to be equal). In sailing one can only indirectly affect the products going into the average, and it is a series of products (each one with x_i increasing by one). y is the base value of the balls in that combo step, which is nominally 1 for a totally vanilla clear, and more for everything else. A V^2 on a normal ball is just 7*1. A bingo on a better ball (say a platform) is 4*k, where k is how much better a platform ball is than a normal ball. k >= 2 (most people say k=2), which gives the bingo at least 8 on that particular combo step. The shorter combo has greater contribution from the low terms, so the V^2 would be better than a vanilla bingo'd platform, but this sort of tradeoff is very important to keep in mind.

For example, P^4 would be 10*k/8, whereas N^7 would have value 28/14. Since k is at least 2, the P^4 is bigger. P^4,N is no better than P^4, unless k is exactly 2 and the resultant value/move is higher than one's normal value/move -- the last term is a measly 5*1 raw, so it would be (10k+5)/10, i.e., k+1/2, which for k=2, is the same as 5/4*k. Any greater k and 5/4*k will exceed k+1/2. Now, if k+1/2 is better than one's normal average, then this is a contribution of 10 moves above average instead of 8 moves at even more above average, so even if k > 2 the combo P^4,N might be good for you personally, but it depends upon where one's average is, and how much weight is given to one's past performance. The difference of contributing 10 moves worth of value above average would not, however, often beat contributing 8 moves of even higher value.

NP^4 is better than P^4 under any reasonable assumptions. This would be the series: 1*1,2*k,3*k,4*k,5*k=14k+1 and 10 moves, so, (14k+1)/10 = 7/5*k + 0.1 7/5 > 5/4, and the .1 only helps the already winning combo (but obviously not by much).

The value of that initial clear is to bump up the multipliers on the platforms. Any number of initial clears, for k >= 2, are superior to less initial clears. For N^mP^4, we have sum(1..m) + k*(m+1,m+2,m+3,m+4) = m*(m+1)/2 + (4m+10)k and 2m+8 moves, so, (m*(m+1)/2 + (4m+10)k) / (2m+8) = f which we want to compare to l=m-1: (l*(l+1)/2 + (4l+10)k) / (2l+8) = (m*(m-1)/2 + (4m+6)k) / (2m+6) = e and we want to show f > e so (m*(m+1)/2 + (4m+10)k) / (2m+8) > (m*(m-1)/2 + (4m+6)k) / (2m+6) <=> (2m+6)(m(m+1)/2 + (4m+10)k) > (2m+8)(m(m-1)/2 + (4m+6)k <=> (2m+6)(m^2/2 + m/2 + 4mk+10k) > (2m+8)(m^2/2 - m/2 + 4mk+6k <=> m^3 + m^2 + 8m^2k + 20mk + 3m^2 + 3m + 24mk + 60k > m^3 - m^2 + 8m^2k + 12mk + 4m^2 - 4m + 32mk + 48k <=> 0 + 2m^2 + 0 + 8mk - m^2 + 7m - 8mk + 12k > 0 <=> m^2 + 7m + 12k > 0 which is true since m,k > 0.

This is easy to see for large m, where the difference between 2m+8 and 2m+6 is neglible, but the difference between m(m+1)/2 and m(m-1)/2 as well as 10k vs 6k, is large.

Eventually P^4N^m is superior to P^4 (consider m=100). Breaking the two pieces separately would be almost as good as joining them; and doing P^4 and then N^mP^1 is probably better than P^4N^m, and almost certainly better than P^4N^m followed by P^1 (but who would do that?). If one's board has only 4 platforms, then adding clears to the end to get rid of any excess balls (as much as possible) is good, as those balls will be 0 otherwise -- any value is better than 0. If the board will have just one platform left, and the waste balls won't be clearable before that platform, then P^4N^m followed by N^lP^1 is better than launching the P^4 immediately and doing N^jP^1N^m, where j is presumably fairly small (because one has to fit the followup m clears for cleanup). If j is 4 or more, then the second way is better; for exampe, if there are very few waste balls, so that minimum m is something like 2, then P^4;N^4P^1N^2 is better than P^4N^2;N^4P^1, since the waste balls have greater value. If clearing up the waste allows one to pull off an N^5P^1, however, then P^4N^2;N^5P^1 is going to be better than P^4;N^4P^1N^2 -- increasing the value of the waste balls is not going to compensate for the decrease (by 1) of the platform in the second clear.

I'm reminded of pool -- really expert pool players are not just contemplating the geometry of the current shot, but how to setup the ball for the next shot, and the shot after that...

I.e., given a 6 platform board, one could try to shoot for some variation on P^6. If the pieces are not in one's favor, then you might have to break it early at something like NP^3NP^2, and be left with one platform. It is likely better to deliberately do NP^3 leaving one with 3 platforms and more space -- perhaps enough space to pull off N^3P^3 as the second clear. NP^3;N^3P^3 is likely better than NP^3NP^2;N^3P^1, but that second clear isn't really realistic since with just one platform it wouldn't be hard to pull off at least an N^4P^1 if not an N^5P^1. If one, from the beginning, plans on doing 3 and 3 on this board, then one might be able to do N^2P^3;N^3P^3, and this is a very good set of combos on a board. They're also both smaller (than the monster NP^3NP^2), which is advantageous in pillage situations. But I haven't run any numbers; and certainly NP^3NP^2 is a combo to be proud of. Just trying to make the point that it could be advantageous to break earlier and leave oneself better placed platforms for the second combo, rather than trying to the make the first as good as possible. If one really practiced this you might be able to reliably do N^2P^3;N^4P^3, or maybe even N^3P^2;N^3P^4 -- either of which is amazing, the second combo set is probably a little better.

N^3P^2 as first combo on a board is really tough though. One would almost certainly have to do down up -- a triple from the top clears a platform on the bottom which has a stack leading up to a top platform. That stack is mostly dead weight (one could do N^3P^2N^2 if there aren't too many obstacles in the down-up fashion), but, one could structure it so that it serves as the backbone of a followup N^3P^4. After the first 3 minutes this would be good.

Of course, if you pull of the NP^6, or the more believable NP^3NP^3 (or perhaps NP^3NPNPNP, which is easier to construct (sortof) because it gives you more time to get just the right pieces for the platforms), that is huge. So ideally one wants to play so as to leave oneself the chance to backout and do an NP^3, if the P^6 variation doesn't work out. Also, one can be ultimate by just doing NP^2 over and over again. One can probably be #1 by just doing N^2P^2 or NP^3 or P^3 over and over. The thing about the P^6 way is that it takes time; time during which scores could be falling out of one's queue. This doesn't matter too much to one's standing, but it does affect the ship. So in order to have the time one has to regularly do big infrequent combos (so that only 0s are falling out of the queue). Backing out and doing two P^3 variations increases the frequency of scoring, making it more difficult to go the P^6 way on the next board. Playing faster helps a little, in that it minimizes the absolute amount of time spent at less than sparkly (if one is okay with continuing to build combos while not sparkly). Ramping up to infrequent scores from frequent scores is tough, and I think one can show that no matter what one has to pay an equivalent penalty of just building the vegas as the very first combo (but doing it as many leagues of excellent before incred is better for the ship, perhaps, then one league of booched followed by increds). Standings-wise I think it doesn't matter at all, although if one gets engaged within the first league this could be fairly awful (since it will force an early score submission...).

The ideal sloop probably has one sailor going for the first league excellent and one sailor going for the second league incred. Two increds isn't much better than one incred, so this gives max acceleration from the starting line, followed by max acceleration when the excellent starts wearing thin (and if one gets engaged, then the vegas builder will hit the incred rate of token generation first, providing max token generation as soon as possible).

It is unclear, to me, if token generation is tied to performance level or raw performance. If performance level, then a NP^2 ult can be just as good for the ship as a vegas-ult. If raw performance, there is a maximum cap that a triple-player can't just get past (the value of P^3), so that the vegas-ult will surpass the triple ult not just on the DR, but also in the absolute amount of token generation. An interesting test would be if the raw numbers from sails are adjusted by bilge levels before being converted into token generation rates, so that at no bilge, it doesn't matter if one is doing low or high incred -- i.e., there is a cap on token generation that a low incred reaches. If when there is bilge the number is adjusted before being capped, then a high incred has value. To my knowledge this test has not been done.

Two ult bilgers can keep the water out at full damage. I wonder if two, or three, ult sailors (on a sloop) can keep up max token generation at full bilge? And whether or not it matters if they are doing high vs low increds? That would be a nice symmetry. Although I'd prefer it if it just wasn't possible to be that good -- infinite move sloops, no matter the damage, are quite irritating.