Forum
xC's rule #7: mid-month report
|
skrumgaer wrote
at 10:04 AM, Friday May 16, 2008 EDT
This month I have been testing xC's rule #7, which is to avoid attacks of less than two-dice advantage during the connect phase. I have operationally translated this into eschewing one-die advantage moves during the first two rounds, otherwise playing my normal (or abnormal) game.
Results: 1. My profile shows a marked decrease in percentage of seventh-place finishes. Under the current scoring system, this amounts to little because seventh is scored no worse than sixth for place. In a number of games I was able to finish in a high place because I narrowly survived attacks after my first or second turn. 2. My profile shows a small decrease in first and second place finishes. In a number of games I did not finish in a high place likely because I did not take advantage of a connect that would have resulted from a one-die advantage attack. Overall assessment: xC's Rule 7 creates a disadvantage if it is followed for two turns. My observation of play at the higher level tables indicates that players routinely make one-die advantage attacks in the first two turns. For the rest of the month I will test the rule during the first turn only. Also, I will feel free to do 2 v 1's at any time. For your information, I submit the following table, which shows the expected number of dice that need to be given up to win a land. 2v1 0.19 3v2 0.57 4v3 1.03 2v2 1.25 5v4 1.57 6v5 2.14 3v3 2.41 7v6 2.74 8v7 3.40 4v4 3.52 5v5 4.62 2v3 5.58 6v6 5.70 7v7 6.79 8v8 7.86 3v4 8.41 4v5 10.64 An attack should be considered if the expected gain in dice is greater than the expected cost. For example, if a 2 v 1 results in an additional restack die, the payoff is +1 and the cost is 0.19 for a net gain of 0.81. If the 2 v 1 does not result in an additional restack die (for example, if you are attacking from an island), there is a net loss of dice on the attack but it might still be justifiable for some other reason, such as improving position or getting out of an opponent's way. In a series of attacks, add the dice costs. For example, a 4 v 3 followed by a 3 v 2 followed by a 2 v 1 would cost 1.79 dice. The series will have a positive net benefit if the payoff in additional restack dice is at least 2. Next month, I will pick another rule to study. One rule I particularly need to study is Do Not Annoy the First Place Player. |
Replies 1 - 5 of 5
|
XCRobin wrote
at 3:41 PM, Saturday May 24, 2008 EDT Skrum, while I will agree with you that following rule #7 will give you less 7th and less 1sts, you also need to look at the net point gain if following the rule. From my playing days, I would be led to believe that the decrease in losses significantly offsets the decrease in winnings.
Let me know what you find. |
|
Danny_DCB wrote
at 5:39 AM, Sunday May 25, 2008 EDT I'm not sure I totally agree with you regarding the first paragraph below the table. I mean shouldn't the newly conquered territory be taken into calculations as well? For example I'm disconnected and have 5 territories and gain 4 dices with restack. But if I connected and won I'd have 6 dices in restack but sprinkled over 6 territories.
I don't know what to take from that but maybe you will be able to explain it to me. :) |
|
|
skrumgaer wrote
at 9:11 AM, Sunday May 25, 2008 EDT xC: I can work up an estimate of the change in points by comparing total points lost/gained between the two months less the total "draw" of points made when negative score is reset to zero. That will take some time but I think I can do it.
|
|
|
skrumgaer wrote
at 9:18 AM, Sunday May 25, 2008 EDT Danny: If an attack will _ultimately_ result in a connect, go ahead and count the new restack dice. But if that connect does not happen until a later turn, and you attack from the smaller island, your win will not result in any new restack dice the first turn.
|
|
jaquibosox wrote
at 3:14 PM, Sunday May 25, 2008 EDT Often times, I'll attack with a one-die advantage if I know that a 4v3 or 5v4 will take out a threat, rather than connect around the threat. Now, I believe this would be nearly impossible to calculate, but it is one more consideration during the connect phase.
|
