Arimaa Forum
        (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
      

        Team Games >> 2009 One vs TheMob >> 2015 Move 3
        
(Message started by: deep_blue on Jul 16th, 2015, 9:12am)
Title: 2015 Move 3
Post by deep_blue on Jul 16th, 2015, 9:12am
While we wait for Sharpvector to play their 2s, let's discuss what our near future plan is.
While in the other thread many people wanted to attack quickly, there just came in this:

on 07/16/15 at 03:30:35, chessandgo wrote:
I've experimented over the years with attacking as early as possible in all kinds of openings. In the end though, against good play you have to take time to develop pieces and shore up home traps soon enough. Attacking too early is not the right thing to do in my experience (unfortunately)...

I agree to that. It may just be my style but I don't like instant attacking out of the opening. Also against Sharp that might be sort of dangerous while I would like to keep the position relatively calm for some time first.

Quote:
...so any reasonable development step is fairly equivalent in moves 2 and 3 imo.

This clearly disagrees with early Horse -> a6 ideas or similar and again I agree.
Title: Re: 2015 Move 3
Post by browni3141 on Jul 17th, 2015, 12:53am
I do not believe all developing moves are close to equivalent early on. Each choice has subtle differences that can affect the opponent's optimal decisions. My preference is typically to play "big" moves first, which means our main plan of advancing a horse before general developing moves. I believe that it is very hard for horse a6 to be a bad move in the absence of a silver threat. As long we do not play such a move silver has the option of playing things like a7s a8s, trying to force us to pull a rabbit to make progress when we want h6.
Just as a simple example, suppose silver plays 2s ed7s ed6s hh7s hh6s
and gold responds with the general development 3g Hg2n Md2n De1n Rg1n (Dd1n is better, but this is just an example) Now silver plays 3s ed5s hh5s hh4s hb7s and gold is already having a problem.
If
4g Dd1n Md3e Me3e xxxx, then
4s ed4e ed5e hh3n Rh2n, and silver has won a good rabbit pull.
If
4g Dd1n Md3w Hb3n Hb4n, then
4s a7s a8s xxxx, and silver has achieved a better trap control racing position, although the evaluation of this depends on whether a rabbit pull is good for gold. I tend to think not.
Finally, if
4g Hb3n Hb4n Hb5w Ha5n, trying to refute silver's idea
4s Md3s ed4s mg7s rh8s seems annoying for gold, somewhat pacifying our camel.

Whatever we do, we should consider seriously the subtle differences between different developing steps.

If that last variation is truly an annoyance, then silver has the same thing if we move horse a6 on 3g, so a more solid developing move might be in order.
Title: Re: 2015 Move 3
Post by chessandgo on Jul 17th, 2015, 4:14am
If your 4g H->a6 transposes to 3g H->a6 lines, I don't see what's lost in not playing 3g H to a6. Also, in your variations, the "problem" for Gold is the caMel getting pushed back to d2, which is easily solved by playing Dd1n before the other less useful development steps.

I am not trying to argue that all development steps are equivalent, but rather that spending four attacking steps on one move and then four develoment steps on the next move typically results in fewer options and possibly weaker position than doing the converse. As a very general 3g philosophy.

We are also free to mix and match, if we think De1n and Rg2n are less important (obviously the latter is), we can play the two or three meaningful steps (3g Md2n Hg3n Dd1n) and play one or two western steps (Hb3n).
Title: Re: 2015 Move 3
Post by chessandgo on Jul 17th, 2015, 5:15am
Another interesting point you are raising is where do we stand on rabit pulling? Silver's would be the Rh2 and Gold's the ra7 presumably. Silver's rabbit pulling prospects are better than ours given that his horse is one step closer (from h7 instead of b2), and his camel in the right place (while ours can usually be cut off from c3 after our Horse advances to, say, a6).

For example:
2s ed7s ed6s hh7s hh6s
3g Md2n Dd1n Hg2n Hb3n
3s hb7s hh5s hh4s ed5s
4g Hb4n Hb5w Ha5n De1n

Both sides can pull a rabbit. I am not aware that sharp evers pulls rabits from the second row, so that's not an issue there. If we use the Rh3 Rh2 defense I don't know how sharp treats these positions. Are we willing to play Rh2n Rh1n as early as on 3g to avoid the above position?

Should sharp play a blank move on 4s, say 4s mg7s rh8s ed4e Md3n, are we willing to consider pulling the ra7 with our remining two steps (after withdrawing the caMel)? Same question after different types of move 4s if this specific position has too much tactics.
Title: Re: 2015 Move 3
Post by ikalyoncu on Jul 20th, 2015, 3:41am
I prefer, in general, solid developing moves to an early horse advance to the sixth rank. The latter leads to tactically involved positions, whereas the former requires more positional understanding. Given Sharp's tactical superiority, I suggest we do simple development, at least for as early as 3g.

So, as we know by 2s that silver did not commit on the east wing, I suggest we do the same and go for something like

3g Md2n Dd1n De1n Hb3n.
Title: Re: 2015 Move 3
Post by Hufflepup on Jul 20th, 2015, 1:10pm

Quote:
3g Md2n Dd1n De1n Hb3n.


This move would give silver the option to confine our camel to the west via something like: ed4s Md3w hh7s hh6s, or ed4e Md3n Md4n ee5w. This might not be such a problem if we are aiming for an EMH attack on C6, but do we want to commit to that this early? If we did I would probably play something like Md2n Dd1n Md3w Ee5w instead.

Would we prefer to play a game with the camels on the same or opposing sides?
Title: Re: 2015 Move 3
Post by SilverMitt on Jul 20th, 2015, 2:58pm
I think I'd be opposed to any move not having 4 steps north here.
Title: Re: 2015 Move 3
Post by clyring on Jul 20th, 2015, 4:25pm
Some loosely organized thoughts:

On general principles I'm not opposed to early advances such as Ha6 and might even object to calling such an attack, as no other pieces are committed, unless the silver elephant decentralizes itself. Having thought more about our response to quick silver aggression in the southeast, however, I view ensuring our camel's activity as a higher priority. This means waiting at least one move.

Actively moving our camel to either wing prematurely is likely a mistake. Not necessarily because either alignment is bad for us, (It would be symmetrical.) but rather because doing so loses a lot of time either way. Matching a horse on the a-file and camel on the b-file is three horizontal steps and horse on the h-file and camel on the g-file is four. The magic number in the opening is two steps, so unless you believe we already stand objectively worse, we should wait until given some enticement to choose one wing over the other.

However, this doesn't at all mean we should fear our camel being pushed to either wing, especially to the west. Doing so costs silver one step and saves us the trouble relative to moving west ourselves, making it cost us a net of only one step to match silver's alignment of mhh. Another, more likely enticement for our camel to go west would actually be the silver elephant involving itself in an attack around f3 before our elephant leaves, when the presence of both elephants on the wing where only silver's camel lies provides us some compensation for our time. Our camel being pushed east is more interesting not least because the positions where our camel is not eventually attacking in the west are much harder to evaluate.

Regarding pulling rabbits, we might want to consider this if the camels end up facing in the east or if silver preemptively advances quite far on the a-file but I think we definitely have higher priorities in Jean's sample line. After 2s ed7s ed6s hh7s hh6s 3g Md2n Dd1n Hg2n Hb3n 3s hb7s hh5s hh4s ed5s 4g Hb4n Hb5w Ha5n De1n 4s rh8s mg7s ed4e Md3n my first instinct is to play something like 5g Dd2n Md4w Mc4w Rg1n and thank silver for helping us develop. I'll have to sleep on when moves such as ra7s ra8s and Rh2n Rh1n might be good.
Title: Re: 2015 Move 3
Post by arimaa_master on Jul 20th, 2015, 7:13pm
My suggestion for 3g is as follows:
3g Cc2w Md2w De1n Ee5n

Rationale behind this move:
Cc2w Md2w - in spite of moving our camel west seems to be unnecessary it helps strengthen our west against eventual E+H attack by sharpvector (and additionally moves our camel to safer place).

De1n - it helps protect the vulnerable f3 trap

Ee5n - threatens cat capture at f6 and central rabbit pull/dog flip

That much for supporting my move.

For other proposals I don´t like advancing g2 horse much (I still fear that opposing camel) nor advancing our camel to d3 (fearing of camel flip to d5 and then to be committed at west wing attack/defense very much, though it doesn't necessary mean it is bad for us).

AM
Title: Re: 2015 Move 3
Post by ikalyoncu on Jul 21st, 2015, 12:53pm
I strongly disagree about Cc2w Md2w steps, based on two grounds. First, it is a really slow plan. Moves with north steps should be prioritized over such side steps. Second, Camel would be much more active on d3 than c2 and Camel flip to d5 on 3s should not be a concern for us. If silver can afford to flip our camel to d5 after developing only a horse, then there is something wrong with this game :)
Title: Re: 2015 Move 3
Post by browni3141 on Jul 21st, 2015, 10:45pm

on 07/20/15 at 19:13:37, arimaa_master wrote:
My suggestion for 3g is as follows:
3g Cc2w Md2w De1n Ee5n

Rationale behind this move:
Cc2w Md2w - in spite of moving our camel west seems to be unnecessary it helps strengthen our west against eventual E+H attack by sharpvector (and additionally moves our camel to safer place).

De1n - it helps protect the vulnerable f3 trap

Ee5n - threatens cat capture at f6 and central rabbit pull/dog flip

That much for supporting my move.

For other proposals I don´t like advancing g2 horse much (I still fear that opposing camel) nor advancing our camel to d3 (fearing of camel flip to d5 and then to be committed at west wing attack/defense very much, though it doesn't necessary mean it is bad for us).

AM


I think that it is too early to advance our elephant to the sixth rank when we may want to play the defensive e5ee later, and the dog flip is not much of a threat. Silver can reply g7s h7sss, and our elephant feels too far from f3.

If we move our camel west, we will have to advance our Hg2 very soon due to silver's threat of an EH attack, so I think it is a very good step to play right now. I like your move better if we substitute Hg2n for Ee5n.

However, I still have one problem with this new move. I do not like the Cc2w Md2w maneuver either because we are spending two steps which might not end up being very useful (Cc2w and re-filling c2 when the camel moves, which it likely will want to do soon)
I don't think the camel flip should be a concern if we advance it now. After

3g Md2n Dd1n De1n Hg2n
3s ed4w Md3n Md4n ec4e
4g Md5w Mc5s Hb3n Mc4s,

we have gained one free step towards our horse advance plan and I wouldn't consider our camel movement as negative. I think flipping the camel would be a small inaccuracy for silver allowing gold to equalize. We could even consider much more aggressive options like 4g Md5w Mc5w Hb3w Ee5n which could yield an edge if they work out tactically.
Title: Re: 2015 Move 3
Post by chessandgo on Jul 22nd, 2015, 2:27am
I agree with clyring, Mitt, ikalyoncu and browni. So:

3g Md2n Dd1n De1n Hb3n (ikalyoncu)
3g Md2n Dd1n Hb3n Hg2n
3g Md2n Dd1n De1n Hb3n

look like the top 3 options.
Title: Re: 2015 Move 3
Post by deep_blue on Jul 22nd, 2015, 4:57am

on 07/22/15 at 02:27:00, chessandgo wrote:
I agree with clyring, Mitt, ikalyoncu and browni. So:

3g Md2n Dd1n De1n Hb3n (ikalyoncu)
3g Md2n Dd1n Hb3n Hg2n
3g Md2n Dd1n De1n Hb3n

look like the top 3 options.

Is it somehow possible option 1 = option 3? ;D
But I think I can agree to those options.
Title: Re: 2015 Move 3
Post by clyring on Jul 22nd, 2015, 7:08am

on 07/22/15 at 02:27:00, chessandgo wrote:
I agree with clyring, Mitt, ikalyoncu and browni. So:

3g Md2n Dd1n De1n Hb3n (ikalyoncu)
3g Md2n Dd1n Hb3n Hg2n
3g Md2n Dd1n De1n Hb3n

look like the top 3 options.

How do we evaluate the camel push to e2 in line II? Overall I feel like our minor pieces are in the way, which I estimate is worth about one facing-camels-step to silver, but I'm moderately unsure, which is at least more than I can say about the corresponding push to d2, and would like to hear your thoughts on this possibility. (Or those of anyone else, for that matter.)
Title: ===Re: 2015 Move 3
Post by Hufflepup on Jul 22nd, 2015, 2:36pm
I don't like move 2. Advancing the g2 horse  makes it more vulnerable to the camel if the elephants end up on the opposite wing, and we may end up having to retreat the horse on b4 to shore up C3 if silvers elephant pushes into d3.

Both the proposed moves appear to concede to Sharp the ability to choose the general direction of the game. Pushing the camel west means that we will probably be playing a game with the camels on opposite wings, if its pushed east we may be forced  to play a game with the camels on the same wing. Do we want to conced this choice to Sharp.
Title: Re: ===2015 Move 3
Post by chessandgo on Jul 22nd, 2015, 3:41pm

on 07/22/15 at 14:36:00, Hufflepup wrote:
Do we want to conced this choice to Sharp.


Given that not conceding that choice would be mean losing several steps, and that we're kind of indifferent, I'd say yes, absolutely.

Sorry for the list confusion, I guess the third choice should have been 3g Md2n Dd1n De1n Hg2n
Title: Re: 2015 Move 3
Post by ikalyoncu on Jul 22nd, 2015, 4:21pm

on 07/22/15 at 07:08:21, clyring wrote:
How do we evaluate the camel push to e2 in line II?


I am not sure about the tactics but Horse to a6 and some fourth step looks good to me. If silver makes an emh attack on the east, gold camel would be on the correct side of the board and it can be shifted to f2 (maybe) if needed to defend. If silver goes for a rabbit pull on the h file, we would be quicker pulling the a7 rabbit. Finally, if silver goes for eh attack on c3, we would be quicker making an EH attack on c6.

All that being said, I still prefer including De1n just to make sure our Camel stays active.
Title: Re: ===2015 Move 3
Post by browni3141 on Jul 22nd, 2015, 9:56pm

on 07/22/15 at 14:36:00, Hufflepup wrote:
I don't like move 2. Advancing the g2 horse  makes it more vulnerable to the camel if the elephants end up on the opposite wing, and we may end up having to retreat the horse on b4 to shore up C3 if silvers elephant pushes into d3.

Both the proposed moves appear to concede to Sharp the ability to choose the general direction of the game. Pushing the camel west means that we will probably be playing a game with the camels on opposite wings, if its pushed east we may be forced  to play a game with the camels on the same wing. Do we want to concede this choice to Sharp.


I would not like retreating the horse after advancing it either, but what's wrong with defending the trap in a different way? One of b1nn, c2wn c1n or c1e e1n, etc. must be fine?
Title: Re: 2015 Move 3
Post by Hufflepup on Jul 23rd, 2015, 3:18am

Quote:
I would not like retreating the horse after advancing it either, but what's wrong with defending the trap in a different way? One of b1nn, c2wn c1n or c1e e1n, etc. must be fine?


We could, but it would take longer, and would potentially leave the horse somewhat exposed.


Quote:
Given that not conceding that choice would be mean losing several steps, and that we're kind of indifferent, I'd say yes, absolutely.


How would a move like Md2n Dd1n Md3w X cost us several steps? X would probably be either Ee5n or De1n.
Title: Re: 2015 Move 3
Post by chessandgo on Jul 23rd, 2015, 7:32am
that's two steps, one played by us, and the one saved by silver not having to push our caMel. Not counting the fact that silver does not need to place his elephant on d3 any longer, etc.
Title: Re: 2015 Move 3
Post by SilverMitt on Jul 23rd, 2015, 8:15pm
So, I'm just going to try to summarize to see how good my understanding is.

There seems to be a general consensus that the steps Md3 and Dd2 are two that should be played, so that if the silver elephant comes to d3, we gain from having the camel on c3 or e3, rather than d2.

For the remaining 2 steps, we have the 3 steps Hb4, Hg3, and Dee2 to choose from.  The first is towards an E+H attack around c6, which makes sense, given the silver camel all the way on g7, and I like it as the 3rd step.  The latter 2 are fairly general developing steps.  I don't think the horse is really vulnerable on g3, as it could realistically be captured only on f6, which would take 8 steps to do, but I'd prefer Dee2 as the 4th step.
Title: Re: 2015 Move 3
Post by chessandgo on Jul 24th, 2015, 2:57am
Sounds good to me, especially as we avoid the 3s M->e2 headache.
Title: Re: 2015 Move 3
Post by clyring on Jul 24th, 2015, 7:15am

on 07/22/15 at 07:08:21, clyring wrote:
How do we evaluate the camel push to e2 in [3g Md2n Dd1n Hb3n Hg2n ]?


on 07/22/15 at 16:21:20, ikalyoncu wrote:
I am not sure about the tactics but Horse to a6 and some fourth step looks good to me. If silver makes an emh attack on the east, gold camel would be on the correct side of the board and it can be shifted to f2 (maybe) if needed to defend. If silver goes for a rabbit pull on the h file, we would be quicker pulling the a7 rabbit. Finally, if silver goes for eh attack on c3, we would be quicker making an EH attack on c6.

All that being said, I still prefer including De1n just to make sure our Camel stays active.

As I look at this line more, I am liking our position more because it seems our camel is better able to defend our home traps without needing to immediately relocate than I anticipated. So I definitely don't fear this move any more than the simpler push to e3 anymore.

Regarding the simple push to e3, I don't fear it, but we can sidestep the issue with 3g De1nn Md2n Dd1n.
Title: Re: 2015 Move 3
Post by arimaa_master on Jul 24th, 2015, 5:55pm

on 07/21/15 at 22:45:02, browni3141 wrote:
I think that it is too early to advance our elephant to the sixth rank when we may want to play the defensive e5ee later, and the dog flip is not much of a threat. Silver can reply g7s h7sss, and our elephant feels too far from f3.

If we move our camel west, we will have to advance our Hg2 very soon due to silver's threat of an EH attack, so I think it is a very good step to play right now. I like your move better if we substitute Hg2n for Ee5n.

However, I still have one problem with this new move. I do not like the Cc2w Md2w maneuver either because we are spending two steps which might not end up being very useful (Cc2w and re-filling c2 when the camel moves, which it likely will want to do soon)
I don't think the camel flip should be a concern if we advance it now. After

3g Md2n Dd1n De1n Hg2n
3s ed4w Md3n Md4n ec4e
4g Md5w Mc5s Hb3n Mc4s,

we have gained one free step towards our horse advance plan and I wouldn't consider our camel movement as negative. I think flipping the camel would be a small inaccuracy for silver allowing gold to equalize. We could even consider much more aggressive options like 4g Md5w Mc5w Hb3w Ee5n which could yield an edge if they work out tactically.


Thanks - now I like 3g Md2n Dd1n De1n Hg2n best.
Title: Re: 2015 Move 3
Post by deep_blue on Jul 27th, 2015, 3:13am
The poll winner was 3g Md2n Dd1n De1n Hb3n.
Unfortunately this time only 13/22 people voted.
Results here http://civs.cs.cornell.edu/cgi-bin/results.pl?id=E_89dd6769b4dd0a8e
Or in case that link doesn't work anymore:

1. 3g Md2n Dd1n De1n Hb3n
2. 3g Md2n Dd1n De1n Hg2n
3. 3g Md2n Dd1n Hb3n Hg2n
4. 3g Md2n Dd1n De1n De2n
5. 3g Md2n Dd1n De1n Md3w
6. 3g Hb3n Hb4n Hb5w Ha5n
1.2.3.4.5.6.
1.-81181011
2.5-1081111
3.13-689
4.356-98
5.2231-6
6.12334-

Don't ask me why 3. is better than 4. based on this matrix, not that it mattered.

Of 22 voters, 13 voted. The right to vote (those 22) had:
Algorias, arimaa_master, aurelian, browni3141, chessandgo, clyring, CraggyCornmeal, DeathCure, deep_blue, half_integer, Harren, harvestsnow, Hufflepup, ikalyoncu, Knedlik, phairland, RonWeasley, Samraku, SilverMitt, Skarn, supersamu, Therin8.

EDIT: Harvestsnow, thanks for pointing out. I messed that up.
Title: Re: 2015 Move 3
Post by clyring on Jul 27th, 2015, 6:32am

on 07/27/15 at 03:13:28, deep_blue wrote:
Don't ask me why 3. is better than 4. based on this matrix, not that it mattered.

The result link is very clear:


Quote:
3. Tied:
3g Md2n Dd1n Hb3n Hg2n  loses to 3g Md2n Dd1n De1n Hb3n by 11–1, loses to 3g Md2n Dd1n De1n Hg2n by 10–3
3g Md2n Dd1n De1n De2n  loses to 3g Md2n Dd1n De1n Hb3n by 8–3, loses to 3g Md2n Dd1n De1n Hg2n by 8–5

(...but the tie is arbitrarily decided in the order of the rows of the detailed table.)
Title: Re: 2015 Move 3
Post by harvestsnow on Jul 27th, 2015, 2:45pm

on 07/27/15 at 03:13:28, deep_blue wrote:
Of 22 voters, 13 voted. The right to vote (those 22) had:
Algorias, arimaa_master, aurelian, browni3141, chessandgo, clyring, CraggyCornmeal, DeathCure, deep_blue, half_integer, Harren, harvestsnow, Hufflepup, ikalyoncu, Knedlik, PerkofBR, phairland, RonWeasley, Samraku, SilverMitt, Skarn, supersamu.

So Perk is back and Therin8 is already gone?
Title: Re: 2015 Move 3
Post by PerkofBR on Jul 29th, 2015, 1:03pm
Hello Mobsters!

I'm back, you can count me in for the next votation deep_blue.

Cheers!
Title: Re: 2015 Move 3
Post by deep_blue on Jul 29th, 2015, 5:13pm

on 07/29/15 at 13:03:39, PerkofBR wrote:
Hello Mobsters!

I'm back, you can count me in for the next votation deep_blue.

Cheers!

Noted.
Title: Re: 2015 Move 3
Post by Algorias on Aug 2nd, 2015, 10:18am
I missed the last vote too since I was away for a week, but I'm back as well.


Arimaa Forum » Powered by YaBB 1 Gold - SP 1.3.1!
YaBB © 2000-2003. All Rights Reserved.