2 minutes ago, Danwarr said:

After watching a few Crane vs Crab match-ups last night I can say that Crab are very good at any kind of defending. Their Stronghold means that defending political attacks with 2 Pol guys bumps them out of Shizuka Toshi range so Crane doesn't normally have good targets for that bow effect. Borderlands Defender is also a pita as is Kisada's universal conflict aura.

Cool! I'm probably going to try a few matches on Tabletop Simulator myself, it seems like a fun match.

Some Observations:

It's interesting to me that the Crane are played aggressively, like the Lion. In the CCG, this was seldom the case. Lion would be the de facto military aggressor. Their focus would be offense. Granted, other clans like Crab and Unicorn focused on military as well, but they would not be as purely focused on offense as the Lion were. In the LCG, Lion seems to maintain that focus, but the Crane are designed in much the same way, only with the Political skill in mind. I'm not sure how their balance issues will play out, but I find this a welcome change.

The Dragon characters have better skills (total) to fate ratios than any clan we have seen previewed. Again, compared to the CCG, this is a clear departure. In the CCG, they would 'give up' efficiency in force stats to be balanced in chi. Not ideal, but versatile. In the LCG, they can threaten military and politics with equal aplomb. What's more, they are not short-changed in either skill when compared to the Lion and Crane. Who would have thought they would have the best balance _and _skills efficiency across the clans? (Of the ones previewed thus far)

The Crab seem more akin to what they were: Heavy military with counter-punching abilities. So far, they are the only clan to have a higher average Military stats quotient (2.85) to their average Fate cost (2.62). So they have to do the least to win via that method. Further, their weakness in average Political stats (1.77) is still relatively better when compared to the weaknesses of Crane in military (1.47) and of Lion in politics (1.53).

At a glance, it seems like Lion and Crane will be more reliant on character abilities and speed to compensate for what they relinquish to Dragon and Crab in pure skill efficiency. Is that the early read most people have on these clans?

I've played Lion versus Crab a couple times and in my experience they've been pretty evenly matched on military for most matches. Lion was easier to play and managed the same score totals just with more guys on the tables. Lion got way more benefit out of being honored, and seemed to honor their characters way more often either through a card effect or via the ring of fire. Also Lion had many effects that allowed their characters to persist (Steadfast Samurai, For Greater Glory..) or recur characters for free (Spirit Caller, Eji) and managed to funnel those superior character counts into a huge benefit from their stronghold or from cards like For Greater Glory, Honored General or Strength In Numbers etc.

I think this game allows for a greater amount of asymmetry since the Dynasty decks are segregated and the influence splashing is so limited. that means doing a comparison across factions can be difficult.

20 hours ago, phillos said:

I've played Lion versus Crab a couple times and in my experience they've been pretty evenly matched on military for most matches. Lion was easier to play and managed the same score totals just with more guys on the tables . Lion got way more benefit out of being honored, and seemed to honor their characters way more often either through a card effect or via the ring of fire. Also Lion had many effects that allowed their characters to persist (Steadfast Samurai, For Greater Glory..) or recur characters for free (Spirit Caller, Eji) and managed to funnel those superior character counts into a huge benefit from their stronghold or from cards like For Greater Glory, Honored General or Strength In Numbers etc.

I think this game allows for a greater amount of asymmetry since the Dynasty decks are segregated and the influence splashing is so limited. that means doing a comparison across factions can be difficult.

Difficult, true, but possible. All clans share the same game mechanics that rely on Fate + Military skill + Political Skill. As a result, we can more acutely interpret how the disparate effects (the asymmetry you mention) allows each clan to close gaps in relative inefficiencies.

The comment about "same scores just with more bodies" is especially pertinent. This supports the theory that Lion are meant to close the Military gap between them and Crab through speed (cheaper characters) and character abilities (honouring effects and character retention abilities).

From what I've seen of Crane, they make up for their military shortfall to other clans by pushing their Political advantage, and by being very selective with their military conflicts. The challenge with them is threatening in Military. Almost makes me think that their Conflict Deck should be all geared to winning military battles because their Politics will take care of itself (save auto-includes like Ornate Fan and perhaps Height of Fashion).

I think looking at average totals is not super illuminating except insofar as it gives an idea of the overall clan's conflict orientation. It ignores the fact that only a subset of the characters will be in the deck, and it also ignores that many of the abilities are essentially raw skill bonuses.

For example, consider Lion. 1.53 average political value seems super low.... except, in practice, it is not nearly that bad. Why? Well, Venerable historian has only 1 political value-- but in a majority of games, she will in practice have a value of 4. Honored general has a value of 1, but starts off at 3. Kitsu spiritcaller can bring in the best political value in your graveyard to fight. Ikoma eiji has 3 pol, but also 3 glory, so court games or ring of fire brings him up to 6...

Another thing that the analysis overlooks is that fate expenditure and mono no aware means that each character does not spend an equal time on the board. More expensive characters usually are in play for longer. In practice this means that the statistics on 1 and 2 drops are less influential than they seem, since they are usually around for only a turn, whereas a Spiritcaller or an Eiji might be around for 2 or 3. Most of Lion's more lopsided military personalities (especially berserker, Deathseeker and Obstinate recruit) are around for 1 turn unless FGG is played.

In actual games, Lion can compete in politics just fine. They'll usually be at a disadvantage against Crane and Dragon, but there is a lot of variance in the draws. The game comes down to specific attacks and choices and plays, which do not often bear much relation to the skill values that have flipped up in the provinces.

The short version is, the game is much too complex and the situation is too fluid for average skill values to be worth much concern. Differences in clan wide skill to fate efficiency is utterly drowned out by the actual board situation.

3 hours ago, Eugene Earnshaw said:

I think looking at average totals is not super illuminating except insofar as it gives an idea of the overall clan's conflict orientation. It ignores the fact that only a subset of the characters will be in the deck, and it also ignores that many of the abilities are essentially raw skill bonuses.

For example, consider Lion. 1.53 average political value seems super low.... except, in practice, it is not nearly that bad. Why? Well, Venerable historian has only 1 political value-- but in a majority of games, she will in practice have a value of 4. Honored general has a value of 1, but starts off at 3. Kitsu spiritcaller can bring in the best political value in your graveyard to fight. Ikoma eiji has 3 pol, but also 3 glory, so court games or ring of fire brings him up to 6...

Another thing that the analysis overlooks is that fate expenditure and mono no aware means that each character does not spend an equal time on the board. More expensive characters usually are in play for longer. In practice this means that the statistics on 1 and 2 drops are less influential than they seem, since they are usually around for only a turn, whereas a Spiritcaller or an Eiji might be around for 2 or 3. Most of Lion's more lopsided military personalities (especially berserker, Deathseeker and Obstinate recruit) are around for 1 turn unless FGG is played.

In actual games, Lion can compete in politics just fine. They'll usually be at a disadvantage against Crane and Dragon, but there is a lot of variance in the draws. The game comes down to specific attacks and choices and plays, which do not often bear much relation to the skill values that have flipped up in the provinces.

The short version is, the game is much too complex and the situation is too fluid for average skill values to be worth much concern. Differences in clan wide skill to fate efficiency is utterly drowned out by the actual board situation.

That's a spot on analysis.

Eugene is right. You can't just make assumption on the strength of each clan based on basic printed stats of their characters without taking into account all the means they have to raises those stats and their tools to get rid of opposing force.

The learning curve of this game will be through the roof. With each game played, I learned something new. After the first couple of games, you build your deck once again from scratch. Cards you thought were too expensive become must have once you realize that Charge is a staple in a military-oriented deck. Nearly any Conflict character and Banzai mean an attack that looked like a scouting party can take down a province, etc.

You will make more crucial decisions each turn that you had to in the CCG. And that is awesome. I'm really looking forward to having all the clans previewed so that we can have our regular l5r evenings start anew and enjoy this game.

For me, there's no doubt, this is going to be far greater and more enjoyable than the ccg ever was.

Give this a few months and we'll have our opportunities to meet old friends around a game of L5R once again. I won't be able to attend GenCon or the Honored event in London, but I'm confident about being able to attend some tournament out of Belgium as well.

9 hours ago, Eugene Earnshaw said:

I think looking at average totals is not super illuminating except insofar as it gives an idea of the overall clan's conflict orientation. It ignores the fact that only a subset of the characters will be in the deck, and it also ignores that many of the abilities are essentially raw skill bonuses.

For example, consider Lion. 1.53 average political value seems super low.... except, in practice, it is not nearly that bad. Why? Well, Venerable historian has only 1 political value-- but in a majority of games, she will in practice have a value of 4. Honored general has a value of 1, but starts off at 3. Kitsu spiritcaller can bring in the best political value in your graveyard to fight. Ikoma eiji has 3 pol, but also 3 glory, so court games or ring of fire brings him up to 6...

Another thing that the analysis overlooks is that fate expenditure and mono no aware means that each character does not spend an equal time on the board. More expensive characters usually are in play for longer. In practice this means that the statistics on 1 and 2 drops are less influential than they seem, since they are usually around for only a turn, whereas a Spiritcaller or an Eiji might be around for 2 or 3. Most of Lion's more lopsided military personalities (especially berserker, Deathseeker and Obstinate recruit) are around for 1 turn unless FGG is played.

In actual games, Lion can compete in politics just fine. They'll usually be at a disadvantage against Crane and Dragon , but there is a lot of variance in the draws. The game comes down to specific attacks and choices and plays, which do not often bear much relation to the skill values that have flipped up in the provinces.

The short version is, the game is much too complex and the situation is too fluid for average skill values to be worth much concern. Differences in clan wide skill to fate efficiency is utterly drowned out by the actual board situation.

Not exactly. The lens is wider than what you discuss here.

Baseline skill values always matter in a game that is structured around said skill values. This may not be realized turn to turn, or character vs. character, but it will be realized across a large number of games, across a large number of board states, and with varying character types. This is where the averages and modes become significant. Clans win or lose based upon the manipulation of Military and Political skill. And so, the baseline skill each clan is afforded cannot be ignored.

The value placed upon baselines is key. They are enough to mean something, but not enough to mean everything. We can surmise that Crane will likely have to do more with character abilities, conflict selection, conflict cards etc... to shore up their relative weakness in the Military stats when compared to other clans. That's a reasonable (I think) claim based upon the information we have to date. The same probably applies to Lion and politics. Is this to say that neither clan can compete in their respective 'weaker' conflict types? No. You've outlined a few reasons why (character abilities). It's really no different about your comment about the Lion's relative political disadvantage to Crane and Dragon. You're right, there is a disadvantage, and it largely comes from the shortfall in baseline Political skill.

I had used the averages to illustrate how clans might play, generally. Crane and Lion would probably favour speed and their use of character abilities to make up for any shortcomings in Military or Political conflicts. Dragon may not favour one play style over another because their baseline stats are strong and balanced. Crab would probably do best to focus on politics via conflict cards and just rely on their excellent military stats to see them through. etc... (Nothing to drastic in that interpretation, IMO)

I have the most experience right now with the Crab. The neutrals bolstered the Crab's lower end political need. At the higher end there were a few characters with mid range political scores. That combine with the stronghold boost and something like Ornate Fan was good enough because all I was trying to do was prevent a break on a province if I could help it. I rarely was trying to actually compete politically with the Crab.

40 minutes ago, Anemura said:

Baseline skill values always matter in a game that is structured around said skill values. This may not be realized turn to turn, or character vs. character, but it will be realized across a large number of games, across a large number of board states, and with varying character types. This is where the averages and modes become significant. Clans win or lose based upon the manipulation of Military and Political skill. And so, the baseline skill each clan is afforded cannot be ignored.

The value placed upon baselines is key. They are enough to mean something, but not enough to mean everything. We can surmise that Crane will likely have to do more with character abilities, conflict selection, conflict cards etc... to shore up their relative weakness in the Military stats when compared to other clans. That's a reasonable (I think) claim based upon the information we have to date. The same probably applies to Lion and politics. Is this to say that neither clan can compete in their respective 'weaker' conflict types? No. You've outlined a few reasons why (character abilities). It's really no different about your comment about the Lion's relative political disadvantage to Crane and Dragon. You're right, there is a disadvantage, and it largely comes from the shortfall in baseline Political skill.

I had used the averages to illustrate how clans might play, generally. Crane and Lion would probably favour speed and their use of character abilities to make up for any shortcomings in Military or Political conflicts. Dragon may not favour one play style over another because their baseline stats are strong and balanced. Crab would probably do best to focus on politics via conflict cards and just rely on their excellent military stats to see them through. etc... (Nothing to drastic in that interpretation, IMO)

You are missing the point. Yes, the average values of military and political skill give us a bit of information. My point is that the information they give is almost useless. About the only thing they tell us is that certain clans will tend to have a slight advantage in certain kinds of conflicts against certain opponents. This is not nothing, but it is not much. For example, the disparity between Lion and Crane in political and military conflicts, respectively, has as much to do with their strongholds as with their personalities.

These skill values are NOT "realized across a large number of games, across a large number of board states". What is realised are the actual skill values of characters that are brought into play on a turn-to-turn basis. This is a subset of the characters included in decks, which is itself a subset of all the characters printed for the clan. Averaging up the skill values for each clan does not tell you much about the skill values that are realised in game, which is what matters.

Not only that, as I explained, they are often actively misleading. It is misleading to say that, for example, Honored General's political skill is 1, because IN PRACTICE it is 3. Yes, this is due to an action on the character, but since the action is triggered every time the character is brought into play, to fail to account for it in an evaluation of the baseline political skill of the clan gives you an incorrect understanding of how the clan plays.

Analysis based on these values that goes much beyond 'in the crab-crane matchup, crab will have some advantage in Military and Crane in political' is not likely to be very helpful.

Well, I've updated the numbers anyway

43 minutes ago, Tonbo Karasu said:

Well, I've updated the numbers anyway

Despite protestations that the numbers aren't useful, I enjoy just being able to look at them.

30 minutes ago, wolfien8 said:

Despite protestations that the numbers aren't useful, I enjoy just being able to look at them.

Kind of helps me a touch when working on my Lion spreadsheet. I'm ahead on avg mil and pol for Lion in general. Even if pol still sucks.

51 minutes ago, RandomJC said:

Kind of helps me a touch when working on my Lion spreadsheet. I'm ahead on avg mil and pol for Lion in general. Even if pol still sucks.

I just like numbers, I used to maintain a spreadsheet of details about my friend groups collection of Xwing ships. We don't play as much now so I don't maintain it, but it had stats on how much each person had for squad totals in each faction with min, max w/o attachments and max w/ attachments. It was beautiful.

1 minute ago, wolfien8 said:

I just like numbers, I used to maintain a spreadsheet of details about my friend groups collection of Xwing ships. We don't play as much now so I don't maintain it, but it had stats on how much each person had for squad totals in each faction with min, max w/o attachments and max w/ attachments. It was beautiful.

Sounds good. Building a deck in excel feeds that need of pointless spreadsheeting that makes my nerd heart happy.

Then using that to compare my #s to other numbers makes me happy as well.

9 hours ago, Eugene Earnshaw said:

You are missing the point. Yes, the average values of military and political skill give us a bit of information. My point is that the information they give is almost useless. About the only thing they tell us is that certain clans will tend to have a slight advantage in certain kinds of conflicts against certain opponents. This is not nothing, but it is not much. For example, the disparity between Lion and Crane in political and military conflicts, respectively, has as much to do with their strongholds as with their personalities.

These skill values are NOT "realized across a large number of games, across a large number of board states". What is realised are the actual skill values of characters that are brought into play on a turn-to-turn basis. This is a subset of the characters included in decks, which is itself a subset of all the characters printed for the clan. Averaging up the skill values for each clan does not tell you much about the skill values that are realised in game, which is what matters.

Not only that, as I explained, they are often actively misleading. It is misleading to say that, for example, Honored General's political skill is 1, because IN PRACTICE it is 3. Yes, this is due to an action on the character , but since the action is triggered every time the character is brought into play, to fail to account for it in an evaluation of the baseline political skill of the clan gives you an incorrect understanding of how the clan plays.

Analysis based on these values that goes much beyond 'in the crab-crane matchup, crab will have some advantage in Military and Crane in political' is not likely to be very helpful.

Respectfully, I understand your point, I just disagree with it.

This is a conversation about probability. If you acknowledge that probability is inherent to every card game, then you are acknowledging that baseline stats are relevant across the majority of games and the majority of board states. Even if those baseline numbers are significantly affected by random variables, game to game. If you don't believe that probability is inherent to card games, then you will tunnel your vision to turn by turn happenings in each game (which you have). The first interpretation has random elements that affect a baseline. The second interpretation has only random elements. For a game that is largely built around numbers, I find the second interpretation to be fallacious.

The interesting point of note here is that you can make the inference that "a certain clan has a slight advantage against a certain opponent" based upon your statistical knowledge of each clan, but are also contending that those stats "don't tell you much about the skill values realized in a game". If they don't tell you much, how can you make that inference? It clearly tells us something. We just disagree to the degree of what it tells us.

In practice, baseline statistics affect each game whether or not variance within each game appears to be significant. There's a starting point which that variance has to affect in order to be considered significant itself. The starting point is the baseline. Once that's established, then we have to quantify the variance. Once quantified, If we deem that variance to be so statistically significant so as to render the baseline insignificant, then the baseline is not the baseline. So far, nothing we have discussed has led me to believe this, but if you want to begin to quantify that variance I'm all eyes/ears.

Lastly, a subset of characters for a given clan should still follow the average or mode for that clan 'most often'. You are still pulling from the card pool that comprised that average or mode. If you want to contend that the subset of characters that make it into decks are _not_ representative of the greater trends in clan skill value, then please describe why you think this using actual cards. All that said, since most of the characters we have seen to date will make it into decks, this is not a significant point in the discussion at present. Also, HG's case is rather unique. It may be that there is a point to giving him an honour bump instead of a fixed 6/3 stat line. Too soon to tell. But sure, give him the bump. Do you think that changes Lion's overall Pol average in a significant sense?

The data are useful. Of course they are. You just have to place them in the proper dimension.

They give us some guide on where each clan is going. They don´t mean that a lion clan deck is always going to be weak on politics, because as has been well stated by eugene, you are not going to place every character x3 on the deck.

But the numbers tell us that if we are going to make a lion deck and do not want it to be extra vulnerable in politics, we better consider placing some extra attention on any advice the Ikoma can give us. Also, they tell us that if we want to make a political oriented deck, we would probably do well to look to another clan than the lions... unless we are completely lion-hearted, or we are aesthetic artists...

The example I chose is not the best, as the conclusion is so clear at first sight... but other clues can be more subtle, and the numbers help.

You can always ignore this data, of course. I wont.

Karasu, keep going.

Speaking from a small bit of personal experience, actually sitting down with your cards in a spreadsheet and working cost v skill is very useful baseline. Knowing what the clan average to work off of when tuning my latest Lion Deck, helped me quite a bit on what to put 2 of or 3 ofs into.

I was able to get an average of 2 cost per card with a roughly 2/2 avg skill. In play testing this worked great with my low draw, honor run.

Proper use of statistical ideas requires that we understand their limitations. So for example, you say that a subset of characters should still follow the average or mode of the population. That would only be true if they were a random sample of the population. But they are not. Pretending they are a random sample is an idealization that is likely to be misleading.

Why is it likely to be misleading? Because the characters put into play, and the characters put into decks, are consciously chosen in order to maximize the chances of winning a game. Therefore any systematic disadvantages in the card pool are likely to be minimized in the actual sample characters that are used.

Let me illustrate with an example. Suppose that it is advantageous to winning to have balanced political and military skill. This may or may not be true, but it's just an example, so bear with me. If this is true, it would be crazy to use the average stats of the card pool to predict anything. Why? Because when people build decks, and when they bring out characters, they will try to balance the political and military skill of their characters, because they want to win, meaning that those values in play will be BIASED samples of the personality base. Making inference about gameplay from the personality averages will be useless, because there is a biasing factor that tends to smooth out the skill differences between clans in actual play. The main constraint the card pool provides in this scenario is any limits it puts on the ability of players to achieve the optimal equal military-political ratio.

I think we had a similar disagreement about statistics in another thread about duelling. I'm not objecting with the use of statistics in appropriate contexts. It is just that there are various circumstances that make it inappropriate to draw the standard inferences from data. One of those is where the sample is not actually random. If the sample is chosen from the population in a way that causally depends on individual values of the represented variable, then the sample will not be predictable based on the statistical properties of the population. So, for example, predicting the heights of professional basketball players based on the distribution of heights in the population is doomed to failure, because above average height is a positive causal influence on being a professional basketball player. That is the idea I have been getting at here.

@Anemura, @Eugene

I think you both have good points. The way I see it, Eugene Earnshaw is correct in pointing out that not all of those characters are going to be played, so the statistics are relevant only if they take into account those cards which are good enough to make the cut; Anemura however is right in pointing out that averages are useful and should not be ignored (something Eugene too appears to recognise in the above post).

In addition to that, right now, most of the Dynasty cards are going to make it into decks simply because most of the neutral options are not good enough (pending future revealed cards, of course - not many slots left for dynasty though, IIRC), so those statistics are actually quite meaningful. They do not go much into details, so it would be unreasonable to expect from them to give us a lot of information.

Bottom line: I wouldn't try to read too much into them, but they are still useful to give us a broad idea of what a Clan is capable of.

I also think that the information above would become a lot more interesting if integrated with the Stronghold ability and the conflict cards a given clan has at their disposal; Lion is a military powerhouse also because of those two factors, not just because of the amount of Mil skill on their characters.

On ‎27‎.‎07‎.‎2017 at 3:12 PM, Eugene Earnshaw said:

Proper use of statistical ideas requires that we understand their limitations. So for example, you say that a subset of characters should still follow the average or mode of the population. That would only be true if they were a random sample of the population. But they are not. Pretending they are a random sample is an idealization that is likely to be misleading.

Why is it likely to be misleading? Because the characters put into play, and the characters put into decks, are consciously chosen in order to maximize the chances of winning a game. Therefore any systematic disadvantages in the card pool are likely to be minimized in the actual sample characters that are used.

Let me illustrate with an example. Suppose that it is advantageous to winning to have balanced political and military skill. This may or may not be true, but it's just an example, so bear with me. If this is true, it would be crazy to use the average stats of the card pool to predict anything. Why? Because when people build decks, and when they bring out characters, they will try to balance the political and military skill of their characters, because they want to win, meaning that those values in play will be BIASED samples of the personality base. Making inference about gameplay from the personality averages will be useless, because there is a biasing factor that tends to smooth out the skill differences between clans in actual play. The main constraint the card pool provides in this scenario is any limits it puts on the ability of players to achieve the optimal equal military-political ratio.

I think we had a similar disagreement about statistics in another thread about duelling. I'm not objecting with the use of statistics in appropriate contexts. It is just that there are various circumstances that make it inappropriate to draw the standard inferences from data. One of those is where the sample is not actually random. If the sample is chosen from the population in a way that causally depends on individual values of the represented variable, then the sample will not be predictable based on the statistical properties of the population. So, for example, predicting the heights of professional basketball players based on the distribution of heights in the population is doomed to failure, because above average height is a positive causal influence on being a professional basketball player. That is the idea I have been getting at here.

With only the core set we will probably play and play against almost every character, so knowing where their overall strengh and weaknesses are is important for judging probable situations throughout a game.

Coming from AGoT Lcg your example for balancing mil and pol is not correct. While we try to balance mil and pol, there will also be incentives to tip the scale, because some cards are simply stronger than others and some effects (especially from Lion) favor one conflict or the other. So while most board states won't represent exactly the board situations you'll rncounter ingame, it gives a good look at tendencies.