Recommend
28 
 Thumb up
 Hide
86 Posts
1 , 2 , 3 , 4  Next »   | 

Terra Mystica» Forums » Strategy

Subject: Faction opening data rss

Your Tags: Add tags
Popular Tags: [View All]
Juho Snellman
Switzerland
Zurich
Zurich
flag msg tools
Avatar
mb
Edits:
- Added two more openings (TE+SH, 3xD+TP)
- Fixed the 2xTP opening (was actually stats for 3xTP on round 1)


The following data is from all 4p online games that are included in the rankings. It shows for each faction how well the faction did on average compared to the other factions in play when using different opening strategies. The way you should read this is for example that Alchemists scored 2vp above average when opening with a SH and 1.3vp below average when using a temple.

The openings I consider interesting are as follows. (If multiple buckets are applicable, pick the one that's earlier in the list).,

- SA: Faction built a sanctuary on round 1
- TE: Faction built a temple on round 1
- SH: Faction built a stronghold on round 1
- 4xD: Faction built at least 4 new dwellings
- 2xTP: Faction build at least 2 trading posts
- other: None of the above. The player is probably not very good. This bucket can be mostly ignored.

Any bucket that did not have at least 10 entries in it will be completely excluded from the stats. So for most factions there will be a few games that won't be in any of the displayed data buckets (e.g. there were under 10 games where Alchemists built 4 or more dwellings on round 1 -- those games are excluded completely, rather than appear under "other").

These statistics include two ways of looking at the data. First we have "raw" numbers, which are just a difference between the actual VPs the factions got compared to the average VP of the other factions. After that we have "skill-adjusted" numbers, where the player skill is taken into account as well (a player with a ranking 100 points higher than an opponent is expected to score 10 points higher).

There are a bunch of interesting things here. I'll point out the ones I that caught my eye sometime tomorrow, unless somebody beats me to them.

And finally, I would really appreciate it if this thread didn't degenerate into yet another boring balance flamewar. We've got plenty of those already. I'm hoping this could be a discussion about what openings work / don't work for various factions, and why. That kind of thing can teach us all how to play better, while bitching about a perceived lack of balance will achieve nothing.


3778/9877 games matched criteria.



raw
alchemists
2xTP -11.92 vp ± 5.88, count=21
SH 2.01 vp ± 1.12, count=650
TE -2.53 vp ± 3.47, count=79
other -19.35 vp ± 2.89, count=117
auren
2xTP -27.55 vp ± 5.27, count=13
4xD 13.93 vp ± 4.57, count=36
SH -2.83 vp ± 1.43, count=420
TE -7.98 vp ± 5.93, count=44
other -13.41 vp ± 3.04, count=101
chaosmagicians
3xD+TP -15.21 vp ± 5.83, count=21
SA 8.94 vp ± 1.58, count=363
SH -36.99 vp ± 5.38, count=25
TE 1.73 vp ± 0.97, count=985
other -24.10 vp ± 2.64, count=124
cultists
2xTP -9.94 vp ± 4.59, count=48
SA -2.71 vp ± 6.33, count=22
SH -33.22 vp ± 6.77, count=20
TE 8.08 vp ± 1.37, count=536
other -11.38 vp ± 2.16, count=203
darklings
2xTP -12.86 vp ± 5.70, count=27
3xD+TP 1.22 vp ± 6.31, count=12
4xD 21.13 vp ± 8.32, count=16
TE 11.24 vp ± 0.85, count=1177
other 2.14 vp ± 1.44, count=450
dwarves
2xTP -5.34 vp ± 4.27, count=35
SA 8.53 vp ± 8.98, count=10
SH -15.60 vp ± 2.40, count=139
TE 9.90 vp ± 1.54, count=386
other -5.30 vp ± 1.62, count=370
engineers
2xTP -7.12 vp ± 4.24, count=60
3xD+TP -8.60 vp ± 5.81, count=30
4xD 8.97 vp ± 6.35, count=22
SA -4.03 vp ± 8.74, count=10
SH -32.58 vp ± 3.81, count=60
TE 5.12 vp ± 1.28, count=587
other -19.39 vp ± 2.37, count=180
fakirs
2xTP -11.49 vp ± 6.74, count=19
SH -17.68 vp ± 3.39, count=77
TE -2.19 vp ± 1.79, count=260
other -19.09 vp ± 3.27, count=78
giants
SH -1.01 vp ± 1.22, count=540
TE -7.72 vp ± 6.56, count=24
other -25.08 vp ± 4.66, count=41
halflings
2xTP -8.02 vp ± 5.02, count=44
3xD+TP 8.33 vp ± 7.10, count=17
4xD 15.44 vp ± 5.95, count=24
SH -13.36 vp ± 1.84, count=207
TE 10.52 vp ± 1.20, count=600
other -10.19 vp ± 1.36, count=470
mermaids
2xTP -11.48 vp ± 3.66, count=67
3xD+TP -0.08 vp ± 4.14, count=44
4xD 8.31 vp ± 2.32, count=149
SH -13.94 vp ± 3.00, count=109
TE 6.69 vp ± 1.17, count=605
other -10.65 vp ± 1.99, count=242
nomads
2xTP 1.90 vp ± 4.01, count=55
SH 0.39 vp ± 0.94, count=980
TE 7.17 vp ± 2.34, count=160
other -14.63 vp ± 2.08, count=197
swarmlings
2xTP -27.27 vp ± 6.52, count=16
3xD+TP -1.93 vp ± 6.36, count=23
SA -12.30 vp ± 11.64, count=10
SH 1.42 vp ± 1.19, count=641
TE -11.91 vp ± 3.73, count=68
TE+SH 8.30 vp ± 1.86, count=234
other -11.59 vp ± 2.19, count=197
witches
2xTP -4.90 vp ± 4.16, count=49
3xD+TP 5.31 vp ± 5.51, count=26
4xD 10.35 vp ± 2.63, count=115
SH 1.19 vp ± 1.20, count=659
TE 4.70 vp ± 2.16, count=214
other -13.00 vp ± 1.98, count=242


skill-adjusted
alchemists
2xTP -8.60 vp ± 4.42, count=20
SH -2.76 vp ± 0.92, count=628
TE -4.24 vp ± 2.58, count=76
other -8.84 vp ± 2.20, count=110
auren
2xTP -13.16 vp ± 5.51, count=12
4xD 8.98 vp ± 3.92, count=35
SH -2.90 vp ± 1.18, count=402
TE -9.74 vp ± 4.46, count=43
other -7.06 vp ± 2.33, count=96
chaosmagicians
3xD+TP -12.53 vp ± 4.80, count=21
SA 6.36 vp ± 1.31, count=351
SH -24.69 vp ± 4.52, count=22
TE 1.86 vp ± 0.79, count=948
other -14.29 vp ± 2.43, count=110
cultists
2xTP -4.97 vp ± 3.62, count=47
SA -2.92 vp ± 4.95, count=21
SH -15.24 vp ± 4.87, count=18
TE 5.25 vp ± 1.11, count=527
other -1.51 vp ± 1.84, count=194
darklings
2xTP -2.50 vp ± 4.57, count=26
3xD+TP 10.33 vp ± 5.42, count=12
4xD 15.44 vp ± 6.50, count=16
TE 9.08 vp ± 0.67, count=1136
other 6.78 vp ± 1.13, count=426
dwarves
2xTP -3.03 vp ± 3.42, count=34
SA 6.54 vp ± 5.12, count=10
SH -9.91 vp ± 1.94, count=131
TE 2.35 vp ± 1.22, count=374
other -1.53 vp ± 1.31, count=354
engineers
2xTP -9.62 vp ± 3.40, count=58
3xD+TP -9.60 vp ± 4.11, count=29
4xD 1.22 vp ± 5.47, count=22
SH -22.66 vp ± 3.31, count=59
TE 0.44 vp ± 1.02, count=572
other -14.03 vp ± 1.88, count=164
fakirs
2xTP -15.86 vp ± 5.11, count=19
SH -13.72 vp ± 2.70, count=74
TE -10.23 vp ± 1.50, count=253
other -12.23 vp ± 2.61, count=72
giants
SH -3.36 vp ± 0.98, count=527
TE -7.06 vp ± 5.42, count=23
other -15.12 vp ± 4.10, count=34
halflings
2xTP -4.68 vp ± 4.15, count=38
3xD+TP -0.26 vp ± 5.46, count=17
4xD 10.60 vp ± 4.66, count=24
SH -10.36 vp ± 1.54, count=201
TE 6.22 vp ± 1.01, count=576
other -4.53 vp ± 1.13, count=441
mermaids
2xTP -3.84 vp ± 2.89, count=65
3xD+TP 2.25 vp ± 3.32, count=43
4xD 3.10 vp ± 2.01, count=144
SH -10.78 vp ± 2.42, count=107
TE 1.68 vp ± 0.95, count=589
other -4.59 vp ± 1.60, count=225
nomads
2xTP 1.87 vp ± 3.14, count=52
SH 1.23 vp ± 0.76, count=941
TE 2.06 vp ± 1.86, count=156
other -4.79 vp ± 1.76, count=179
swarmlings
2xTP -18.99 vp ± 5.26, count=16
3xD+TP 6.02 vp ± 4.82, count=22
SA -12.52 vp ± 7.78, count=10
SH 0.56 vp ± 0.94, count=617
TE -8.75 vp ± 2.92, count=66
TE+SH 3.60 vp ± 1.60, count=226
other -4.34 vp ± 1.84, count=182
witches
2xTP -1.48 vp ± 3.13, count=46
3xD+TP 0.59 vp ± 4.21, count=25
4xD 3.97 vp ± 2.25, count=112
SH 2.84 vp ± 0.99, count=625
TE 1.78 vp ± 1.62, count=206
other -4.72 vp ± 1.52, count=229
26 
 Thumb up
6.25
 tip
 Hide
  • [+] Dice rolls
Robert
Germany
Bocholt
flag msg tools
badge
I paid 100 Geek Gold so that you can read this! :-)
Avatar
mbmbmbmbmb
Great data and certainly interesting food for thought!

Just so I understand "raw" vs. "skill-adjusted" correctly, some questions taking the Alchemists as an example:

Quote:
raw
alchemists
SH 2.01 vp ± 1.12, count=650
TE -1.26 vp ± 3.41, count=84
other -18.08 vp ± 2.59, count=140

skill-adjusted
alchemists
SH -2.76 vp ± 0.92, count=628
TE -4.04 vp ± 2.53, count=81
other -8.71 vp ± 1.95, count=132


a) The count for skill-adjusted is smaller - does this exclude games with Alchemists who never played enough to get a ranking?
b) The average to which the delta is given (e.g. 2.01 VP for SH in raw data) is the average of that race in 4p, or of all races in 4p?
c) Could you explain how the VP and average for the skill-adjusted data is calculated? For the Alchemists, the delta is negative in all cases - how does that happen? Does it mean that Alchemists are often played by high-ranking players who do well (which bumps up the raw data) but not as good as expected from their rank (which lowers the skill-adjusted data)?
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Robin Zigmond
United Kingdom
Durham
flag msg tools
mbmbmbmbmb
Cool, thanks for doing this, Juho! I'm too tired right now (about to go to bed) to comment on anything (or even really notice anything, if truth be told), but I have one question about the pre-amble.

You say that a rating difference of 100 equates to a difference of 10VPs, on average, but is there any basis for this - be it either theoretical or empirical, based on the data - or is it (as it appears to me, but I hope it isn't) just a guess produced from thin air?

(It would also be interesting to now watch upcoming games, particularly those from the league season about to start, and see whether there is a sudden shift in popularity towards the "best" openings as shown here )
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Dave Eisen
United States
Redwood City
California
flag msg tools
mbmbmbmb

The most interesting thing I see off the bat here is the value of getting 4D for halflings and darklings, two factions I had considered to be automatics for building a temple. I realize this is in very few games as both factions in fact do almost always build a temple on turn 1, but might be worth pursuing.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Rhett Morgan
United States
Washington
Dist of Columbia
flag msg tools
Avatar
mbmbmbmbmb
For starters, 4xD is ridiculously strong for darklings no matter what skill level. Also new players don't get the engineers.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Robin Zigmond
United Kingdom
Durham
flag msg tools
mbmbmbmbmb
dkeisen wrote:

The most interesting thing I see off the bat here is the value of getting 4D for halflings and darklings, two factions I had considered to be automatics for building a temple. I realize this is in very few games as both factions in fact do almost always build a temple on turn 1, but might be worth pursuing.


Yeah, that's a good point. I can see the value of doing so (although am not convinced that it's better than a temple) but even if we establish that it's the "best" opening, I'm not quite sure how one goes about pulling it off (as far as I know it's only a common opening for the green factions, and that relies on the board configuration allowing it if you can get early shipping). It seems especially hard for the halflings - presumably it involves the double-spade power action and finding 2 more brown hexes in easy reach?
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Robert
Germany
Bocholt
flag msg tools
badge
I paid 100 Geek Gold so that you can read this! :-)
Avatar
mbmbmbmbmb
Some thoughts on the 4xD openings:
An opening round with 4 new dwellings seems the winner regardless of faction, but that's not really very surprising, if you can pull it off; or is it?

We've already discussed that building 4D this relatively easy for the green races with help of ship+1 and requires only one spade (C5). It seems more often used for Witches (115 out of ca. 1300) than for Auren (36 out of ca. 600).

Darklings need to dig at least three times (get from G5 to E10 or vice versa via G6 and F7), i.e. they must get two spades out of power actions (ACT2, ACT5, ACT6) or bonus tiles (priest or spade). More likely they will need to dig four times though. 16 instances out of 1600 isn't a lot.

Engineers pulled it off 22 times (of ca. 900) - they also need three spades, but must rely on ACT5, ACT6 and the spade bonus tile. OTOH engineers start with 9 power in bowl II, so ACT6 is achievable.

Halflings can do it with just two spades (G5 and E10) - 24 did it out of 1300.

Mermaids with ship+1 can do it with just one spade (C1), without the ship+1 bonus tile they need two spades (C1, F2). However the spread of dwellings doesn't seem very attractive.

As per Robin's remark above, I would bet that green 4xD openings will become more popular, as will TE openings for most other races (Alchemists will probably continue to go for SH, and Chaos Magicians for SA).

P.S.: The most surprising data points for me are
1) the (few) SA openings by Dwarves give such great results. Probably these are cases with SA/SH scoring in round 1, SA/SH bonus tile and little competition on the cults.
2) Swarmlings opening with TE do slightly better than with SH. Since Swarmlings are about the only race who can build BOTH an SH and a TE in round 1, it would be interesting to look at the data for this special case.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Juho Snellman
Switzerland
Zurich
Zurich
flag msg tools
Avatar
mb
DocCool wrote:
a) The count for skill-adjusted is smaller - does this exclude games with Alchemists who never played enough to get a ranking?

That's correct.

Quote:
b) The average to which the delta is given (e.g. 2.01 VP for SH in raw data) is the average of that race in 4p, or of all races in 4p?

It's not a comparison to any global average, but specifically to the average scores that the other players got in the games from which that particular statistic was calculated. This accounts for the different scoring potential in different games, as well as things such as things such as "the giants are good because they can lower the scores of the opponents" hypothesis.

So for example if the SH bucket for alchemists shows a delta of 2vp, it means that in a game where the Alchemists build a SH on round 1, they score on average 2vp more than the other factions in that game. Results from games which don't involve Alchemists at all or games where the Alchemists do something else have no effect on this number.

Quote:
c) Could you explain how the VP and average for the skill-adjusted data is calculated? For the Alchemists, the delta is negative in all cases - how does that happen? Does it mean that Alchemists are often played by high-ranking players who do well (which bumps up the raw data) but not as good as expected from their rank (which lowers the skill-adjusted data)?

Continuing with the same example, -2.5vp for the Alchemist SH bucket means that for games where the Alchemists build a SH on round 1, they scored on average 2.5 vp less than expected (relative to the opponents), based just on the ratings of the players involved. So why is the skill-adjusted number worse than the raw number? There are a number of different explanations which we can't really distinguish between:

1. Strong players play Alchemists relatively more often than weak players. This means that the "raw" score is artificially inflated due to containing a relatively high amount of scores from strong players.
2. This particular opening is an avenue of play that only strong players attempt. (The inverse of this is visible in something like the a round 1 cultist SH. It performs absurdly badly, but a significant component of that bad performance is that a good player would never actually open with that).
3. Weak players perform at an average level when playing the Alchemists, but strong players do worse than average. (E.g. a strong player was expected to score +10vp delta but only got +5vp, a weak player was expected to score +0vp delta but got +5vp. The raw average is +5vp, the adjusted one is 0vp).
4. Strong players perform at an average level in this scenario, while weak players do worse than average. (E.g. a strong player is expected to get +10vp but gets +15vp, a weak player is expected to score 0vp delta but actually scored -5vp. The raw average is +5vp, the adjusted one is 0vp).

And probably others I haven't thought of.

My expectation is that the main value here is in situations where the raw and adjusted deltas are significantly different, not just a difference of a few vps. In that case you can suspect that the actual reason for the results is due to the distribution being skewed, for example due to it being a play that only weak or only strong players would do (point 2 above). You won't be able to distinguish between cases 3 and 4 with just this data. (You could by adding another dimension, and I've done it for some other things on which there might be another post later. But it'd slice this data into too thin pieces).
4 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Juho Snellman
Switzerland
Zurich
Zurich
flag msg tools
Avatar
mb
robinz wrote:
You say that a rating difference of 100 equates to a difference of 10VPs, on average, but is there any basis for this - be it either theoretical or empirical, based on the data - or is it (as it appears to me, but I hope it isn't) just a guess produced from thin air?

It's empirical.

The actual number is a bit smaller, and has basically fluctuated between 9.6 and 9.8 over time. It's just dumb luck that it's so close to a nice round number, the constants used for the rating calculation were chosen arbitrarily. I also didn't expect that there would be a nice correlation between ratings and scores (and IIRC even argued on bgg that it was silly to expect such a correlation). But when I ran the numbers, it turned out I was wrong. Which was nice.



5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Juho Snellman
Switzerland
Zurich
Zurich
flag msg tools
Avatar
mb
DocCool wrote:
Darklings need to dig at least three times (get from G5 to E10 or vice versa via G6 and F7), i.e. they must get two spades out of power actions (ACT2, ACT5, ACT6) or bonus tiles (priest or spade). More likely they will need to dig four times though. 16 instances out of 1600 isn't a lot.

There's also the shipping option. Start at E5 with BON4. Advance ships, build H5+C1+B3. Get an extra priest from ACT2. Dig+build an extra somewhere (could even be two extra, but you run out of workers). Don't know if anyone has actually tried this...

Quote:
Engineers pulled it off 22 times (of ca. 900) - they also need three spades, but must rely on ACT5, ACT6 and the spade bonus tile. OTOH engineers start with 9 power in bowl II, so ACT6 is achievable.

Or two spades and shipping. (For example start at E7 and G1, dig D4 and F2. H6/I9 are another option).

Quote:
Mermaids with ship+1 can do it with just one spade (C1), without the ship+1 bonus tile they need two spades (C1, F2). However the spread of dwellings doesn't seem very attractive.

It can work superbly well with a round 4 spade scoring -- the mermaids can have enough workers after this to basically pave the whole map blue.

Quote:

P.S.: The most surprising data points for me are
1) the (few) SA openings by Dwarves give such great results. Probably these are cases with SA/SH scoring in round 1, SA/SH bonus tile and little competition on the cults.

Note that it's a delta of 8vp with error bars of 8vp. It's a very weak signal, and one I'd ignore unless someone can come up with an explanation for why that'd be a good move. It seems insane :-)

Quote:
2) Swarmlings opening with TE do slightly better than with SH. Since Swarmlings are about the only race who can build BOTH an SH and a TE in round 1, it would be interesting to look at the data for this special case.

Good point, that didn't occur to me at all! This paints a very different picture:


raw
swarmlings
SA -12.30 vp ± 11.64, count=10
SH 1.42 vp ± 1.19, count=641
TE -11.91 vp ± 3.73, count=68
TE+SH 8.30 vp ± 1.86, count=234
other -11.72 vp ± 2.01, count=236

skill-adjusted
swarmlings
SA -12.52 vp ± 7.78, count=10
SH 0.56 vp ± 0.94, count=617
TE -8.75 vp ± 2.92, count=66
TE+SH 3.60 vp ± 1.60, count=226
other -4.39 vp ± 1.68, count=220


7 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Eric Hogue
United States
O'fallon
IL
flag msg tools
Avatar
mbmbmbmbmb
If you open with three dwellings and a TP, is that considered a weak opening? It would be an "other" on your list. I do see there is a wasted worker in this opening.

Also, the engineers can open with both a TE and an additional TP. Is that the same as just a TE entry, or should it be separate?
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
David Jones
United States
Wilsonville
Oregon
flag msg tools
mbmbmbmbmb
I have a friend who insists that every race should be rush building their stronghold because it usually enhances that races ability. Its interesting to see how detrimental this strategy actually is.

I'm a bit surprised to see the 2TP data in here at all. I think one thing that most people learn early on is that too many trading post in the early game leave you without the workers you need to keep your empire expanding. I'm actually curious to know how each race does based simply on whether they have or don't have a trading post at the end of the first round. There are always exceptions, but I'm often inclined to avoid ending the first round with a TP on the board. The other thing I can upgrade the TP into or build instead of building it in the first place is often better in the long term. I also often struggle with the decision to upgrade the shovel track, so knowing if this is a good first round move or not would also be useful information.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Rhett Morgan
United States
Washington
Dist of Columbia
flag msg tools
Avatar
mbmbmbmbmb
davypi wrote:
I have a friend who insists that every race should be rush building their stronghold because it usually enhances that races ability. Its interesting to see how detrimental this strategy actually is.

I'm a bit surprised to see the 2TP data in here at all. I think one thing that most people learn early on is that too many trading post in the early game leave you without the workers you need to keep your empire expanding. I'm actually curious to know how each race does based simply on whether they have or don't have a trading post at the end of the first round. There are always exceptions, but I'm often inclined to avoid ending the first round with a TP on the board. The other thing I can upgrade the TP into or build instead of building it in the first place is often better in the long term. I also often struggle with the decision to upgrade the shovel track, so knowing if this is a good first round move or not would also be useful information.


Don't think ever upgrading dig turn 1 is a good play, probably falls into the "other" plays here.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Chris H
Singapore
flag msg tools
mbmbmbmbmb
jsnell wrote:


nomads
2xTP 1.87 vp ± 3.14, count=52
SH 1.23 vp ± 0.76, count=941
TE 2.06 vp ± 1.86, count=156
other -4.79 vp ± 1.76, count=179


(skill-adjusted)

This is pretty surprising to me. First, apparently TE is comparable to SH for Nomads as an opening. Not a play I would have usually considered, although the very good raw score (+7) seems to suggest it's a play mostly done by good players.

Second, at least in this relatively small set of 52 games, even 2x TP performs on par with TE and SH! This is quite baffling to me. Is it really possible to build a strategy around this that I'm just not seeing?
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
David Jones
United States
Wilsonville
Oregon
flag msg tools
mbmbmbmbmb
Thrar wrote:
Second, at least in this relatively small set of 52 games, even 2x TP performs on par with TE and SH! This is quite baffling to me. Is it really possible to build a strategy around this that I'm just not seeing?


Nomads don't need as many workers to terraform with their sandblast and also get one extra dwelling at the start of the game. I don't think this is so much a strategy as much as it is a reflection that they have a kickstart advantage and don't need workers as badly as other races do. I can see where the extra power income can be useful, but I don't know that I would build a strategy around it. JMHO.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
David Jones
United States
Wilsonville
Oregon
flag msg tools
mbmbmbmbmb
Prodigaldax wrote:
Don't think ever upgrading dig turn 1 is a good play, probably falls into the "other" plays here.


Well, I don't want to get too far off topic, but if you are going to perform two terraforms in the first round, its six workers for two shovels at three per shovel OR two workers to upgrade plus 2x2 workers to terraform. So in terms of workers, you break even. It isn't necessarily an "other" move because some of these goals are still achievable even after paying the cost of the shovel upgrade. The only added costs is the priest/money and priests don't build buildings (unless you downgrade them). For me, the question is more about whether that priest is more valuable in the temple than it is spent on the upgrade action.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jan B.
Germany
NRW
flag msg tools
badge
Avatar
mbmbmbmbmb
The factions that built 4+ dwellings in the stats from Juho most likely got the needed spades through the double-spade power action. I don't think anyone would win a game with investing 6 workers into 2 digs in round 1.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Robin Zigmond
United Kingdom
Durham
flag msg tools
mbmbmbmbmb
EricHogue wrote:
Also, the engineers can open with both a TE and an additional TP. Is that the same as just a TE entry, or should it be separate?


Looks like it'll be counted just as a TE here, but I agree it would be nice to get further stats for the engineers. My favourite opening with the engineers is actually to go for 2 temples (thanks to the Engineers' cheap upgrade costs, you can do this with only a few coins on top of your starting resources, which is easy to get from the round 1 bonus tile or power). I haven't done very well with this online recently, and perhaps it's not a good idea - you're certainly short of workers for round 2, and taking the 1 worker 1 power income as one of your favours is pretty much a necessity - but it would be good to see stats on it, assuming I'm not the only one mad enough to try it
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Petri Savola
Finland
Espoo
Unspecified
flag msg tools
designer
mbmbmbmbmb
I consider TE+2xTP+D and especially TE+2xTP+2xD a strong opening with many races (Nomads, Engineers and Cultists can most often pull it off). I guess these openings listed under TE? It'd be interesting to have a separate category for TE+2xTP with at least 1 dwelling.

I also think that SH+2xTP+D can be strong. Nomads and Swarmlings can often do this opening and Giants sometimes.

I think it would be more clear to name openings based on the buildings you have after first round, not the buildings you build on first round. Otherwise Nomads and Chaos Magicians are always separate cases. So when I have TE+TP+D opening, it would mean that I have those buildings at the end of the round (and have probably done TE+2xTP+D opening) instead of having built a temple and a dwelling.

I'd also be interested in 3xTP openings, because I've been trying those out (Dwarves and Yetis), but I guess there are too few games to get any reasonable numbers.
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Robert
Germany
Bocholt
flag msg tools
badge
I paid 100 Geek Gold so that you can read this! :-)
Avatar
mbmbmbmbmb
robinz wrote:
EricHogue wrote:
Also, the engineers can open with both a TE and an additional TP. Is that the same as just a TE entry, or should it be separate?


Looks like it'll be counted just as a TE here, but I agree it would be nice to get further stats for the engineers. My favourite opening with the engineers is actually to go for 2 temples (thanks to the Engineers' cheap upgrade costs, you can do this with only a few coins on top of your starting resources, which is easy to get from the round 1 bonus tile or power). I haven't done very well with this online recently, and perhaps it's not a good idea - you're certainly short of workers for round 2, and taking the 1 worker 1 power income as one of your favours is pretty much a necessity - but it would be good to see stats on it, assuming I'm not the only one mad enough to try it
My thoughts exactly! Why stop at 1 TE + 1 TP when you can get 2 TE in rd1, giving you 5pw instead of just 1pw+2c? If you really want the 2c, the 4pw are (sort of) equivalent, but the power gives you other options as well.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Robert
Germany
Bocholt
flag msg tools
badge
I paid 100 Geek Gold so that you can read this! :-)
Avatar
mbmbmbmbmb
jsnell wrote:
Quote:
2) Swarmlings opening with TE do slightly better than with SH. Since Swarmlings are about the only race who can build BOTH an SH and a TE in round 1, it would be interesting to look at the data for this special case.
Good point, that didn't occur to me at all! This paints a very different picture:

raw
swarmlings
SA -12.30 vp ± 11.64, count=10
SH 1.42 vp ± 1.19, count=641
TE -11.91 vp ± 3.73, count=68
TE+SH 8.30 vp ± 1.86, count=234
other -11.72 vp ± 2.01, count=236

skill-adjusted
swarmlings
SA -12.52 vp ± 7.78, count=10
SH 0.56 vp ± 0.94, count=617
TE -8.75 vp ± 2.92, count=66
TE+SH 3.60 vp ± 1.60, count=226
other -4.39 vp ± 1.68, count=220
Thanks for the extensive answers, especially about the meaning of "average", and for the Swarmling SH+TE start. Seems to be the strategy of choice, as it combines full payback from the SH (with 6 free TPs, and likely SH scoring in rd1, plus bonus tile SH scoring opportunities) with the general advantage of an early TE (giving 5 priests and an early favor tile)!
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Chris H
Singapore
flag msg tools
mbmbmbmbmb
Petri wrote:
I also think that SH+2xTP+D can be strong. Nomads and Swarmlings can often do this opening and Giants sometimes.

How can you afford that? If you want to end with SH+2xTP+D as Nomads (starting with 3xD), you need 11 workers and 19 coin. You start with 6w and can get 2w each from bonus tile and power action. That's 10w, and you already spent most of your power.
Where does the remaining 1w 4c come from?
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Petri Savola
Finland
Espoo
Unspecified
flag msg tools
designer
mbmbmbmbmb
Thrar wrote:
Petri wrote:
I also think that SH+2xTP+D can be strong. Nomads and Swarmlings can often do this opening and Giants sometimes.

How can you afford that? If you want to end with SH+2xTP+D as Nomads (starting with 3xD), you need 11 workers and 19 coin. You start with 6w and can get 2w each from bonus tile and power action. That's 10w, and you already spent most of your power.
Where does the remaining 1w 4c come from?

Well I wanted to say that you end with SH, TP and 1-2 dwellings. But if you end with that, you build 2xTP, SH and 0-2 dwellings.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Robert Stewart
United Kingdom
Newcastle-upon-Tyne
flag msg tools
mbmbmbmbmb
Thrar wrote:
jsnell wrote:


nomads
2xTP 1.87 vp ± 3.14, count=52
SH 1.23 vp ± 0.76, count=941
TE 2.06 vp ± 1.86, count=156
other -4.79 vp ± 1.76, count=179


(skill-adjusted)

This is pretty surprising to me. First, apparently TE is comparable to SH for Nomads as an opening. Not a play I would have usually considered, although the very good raw score (+7) seems to suggest it's a play mostly done by good players.

Second, at least in this relatively small set of 52 games, even 2x TP performs on par with TE and SH! This is quite baffling to me. Is it really possible to build a strategy around this that I'm just not seeing?


Take a look at the error bar on that 2xTP figure - the true figure could be anywhere from -1.27 to +5.01, so, while it's more likely than not that a 2xTP opening with the Nomads will perform well, it would be premature to conclude that it does - what the data really tells us is that it's too soon to call it...
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
ben ben
France
flag msg tools
Could we add TE*2 and TE+TP+2D to the monitored engineer scenarios ?

Also would it be possible to extract the different building combinations by race and popularity, rather than using pre-designed overlaping scenarios ?
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
1 , 2 , 3 , 4  Next »   | 
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.