"Avoiding the Memory of 'Results' cards" Topic

Combat Patrol uses cards to determine various outcomes. Does it get round the problem of 'Deck Memory' and the card counters?

I'd like to use the mechanism, but have always been put off.

Cut twice when you re-shuffle

Given that each card represents a different outcome depending on what you are testing for card counting is useless.

Have you looked at any of the various videos that have been done on how the game is played?

I've not played, so this will probably be worthless advice. Try shuffling before it's called for (like at the end of a turn). Or add a card that calls for reshuffle when it's drawn.

When the Ace of Hearts comes up in TSATF when resolving shooting results, I reshuffle.

The deck is large enough to be useful but small enough to need to be shuffled/randomized a few times throughout the course of a game.

The same deck of cards is used to obtain a result for different actions/events throughout each turn of the game. There are multiple items on the face of the card. You do not use each pulled card for the same thing. Therefore when you shuffle the deck, you randomize the cards once. They are then self-randomized in that each succeeding card is not used for the same purpose.

The deck of cards replaces the need for pages of result charts and dice rolls. Each card contains result info that is different from the others. You pull a new card any time you need to obtain a result for something. As such, the deck of cards replaces many different charts. With each card being different, you simply pull a new card instead of looking up a chart and rolling a die.

It is a really simple system, just a little bit different than a traditional dice rolling game. Basically, the designer frontloaded all the charts and a randomized dice roll results and spread that info out through the deck of cards.

If you look at the samples on the website and in the videos, you will get a better idea of what I am trying to say.

Last Hussar-
" Does it get round the problem of 'Deck Memory' and the card counters?"

Yes, I saw the game played once or twice at conventions. I then played it at one of the same cons. It works very well.

There is no need to add or delete cards. The deck gives a true random result when needed. And the deck will carry you through each game turn just fine without having to add, subtract or change the make-up of the deck.

In my game of "Oh God! Anything But a Six!" the combat results deck has 120 cards of combat results. I also have twelve decks of cards shuffled together, and placed about the table. There is absolutely NO attempt to restore decks or sort them out, so any pile of cards you pick up literally has a random number and type of each.

To prevent cheating… I

1. All cards played on an enemy unit go into the deck of the guy who has the unit they were used against.

2. Or sometimes I have them tossed into a box on the side and I periodically collect and reshuffle them.

3. Players will be told to reshuffle their decks from time to time.

4. Players will be told to switch decks with opponents across the table. That really pisses off the guys who surreptitiously try and stack decks.

5. Sometimes I will come along with a deck and hand it to a player taking his deck and tossing it into the dead box.

6. So cheating is pretty much impossible, and there is no deck memory.

7. The cards themselves are both good and bad, and some are no effect. You are also often dealing out five and six cards in a specific combat. So you may get a wild assortment of results.

8. The greatest bar to cheating is simply that when placed on an enemy unit, cards are placed face down. Then the enemy unit gets a chance to "roll them off." That is, make a morale check as it were to avoid the face down cards. Only those cards that you don't roll off apply to the unit and the person chooses among the face down cards. Of course everyone wants to see what they missed so the rolled off cards are always turned up to see what the game missed, in addition to what he had to take. If someone has stacked the deck it becomes quite obvious immediately when he has long odds. If you have twelve cards on an enemy unit and when all are revealed you have six eliminated cards, (in the 960 or so cards spread around the table and there are in all those cards 36 "Eliminated cards", if six of them show up in one 12 card draw, it's pretty evident the guy who placed them was probably cheating so he's watched like a hawk from then on.

The "War Stories" board game have poker-sized chips. Pull one, toss it back in the bag, shake. Of course, they have very well-organized icons, so they can get away with less area than a full playing card.

Last Hussar:

In ANY system in which you do not replace the cards and reshuffle after each draw there is some degree of card memory.

In Combat Patrol(TM) each player has his own Action Deck. It is pretty hard to card count four or six decks simultaneously unless you're Rain Man. :) All decks are reshuffled at the end of each turn. As I have observed games using the rules, rarely does a player us half his deck in any given turn. As zoneofcontrol stated, the cards are used for different purposes throughout the turn, and you look at different parts of the cards depending on what you are doing. The results on different parts of the card are independent of each other because of the way you draw cards.

I think the issues you fear have been sufficiently mitigated in Combat Patrol(TM). I provided three demonstration videos on the rules' Web page: link I'd suggest you watch these and form your own opinion. The slight risk of a card counter I think is greatly outweighed by the streamlined gameplay.

Thanks for your interest in Combat Patrol(TM): World War II.

Buck Surdu

What is the problem of deck memory?

As surdu2005 points out, deck memory is an inherent part of a mechanism using a deck. How, when, and what goes in to deck management should be an intentional part of the mechanism.

The problem of deck memory is that certain results become more likely as you work through it- you know that I'm assuming. The problem with this is that's not how real life works.

Tank type A has a 75% chance of killing Tank Type B.

If using dice, it doesn't matter how many times you fire with different tanks, that stays at 75%. With a deck if Tanks 1, 2 and 3 hit, tank 4 is more likely to miss, even if in a situation that is not affected by tanks 1-3.

This is separate to the card counter problem that can also happen. I should have changed the grammar slightly to make it clear that counters are a separate problem, but one is caused by memory.

My question was had CP solved this. And No, it hasn't. While it does make counting more difficult, with multiple resolution types per card, someone who is interested in just 1 or 2 would still be able to track them. Also given the amount of paraphernalia around a game- counters, dice etc, it would be easy enough for someone determined to count to do so without being detected.

Even without counters, the fact that the deck skews if not shuffled does present a problem I don't wish to start a Sim v Fun argument, but surely we all agree that a historical based game should at least root in reality.

Even without counters, the fact that the deck skews if not shuffled does present a problem

It is not a big problem. If it is too big for your taste, then shuffle more.

Duncan

Even without counters, the fact that the deck skews if not shuffled does present a problem. I don't wish to start a Sim v Fun argument, but surely we all agree that a historical based game should at least root in reality.

Last Hussar:

To assert that a card-based mechanic has no root in reality is arrogant, without context, and insulting -- the kind of behavior one expects on TMP. Just because you prefer one mechanic over another does not mean that it is more or less "rooted in reality." Many charts, dozens of modifiers that provide the illusion of precision, and biased dice are no more "rooted in reality" than is a card based mechanic.

Now let's actually get what I think was the original point of this thread: does a card based mechanic impose undue bias on the combat results one is simulating?

As I said at the beginning of my initial response, ANY card mechanism that doesn't involve reshuffling the deck THOROUGHLY after EACH card draw has SOME memory. The question is whether that memory matters. Almost no dice used in our hobby are truly random. Nor are pseudo random number generators used in professional simulations (I do that for a living too, and I know this to be true). That is why they are properly called PSEUDO random number generators. A small failure in the random number generator has little bearing on the fun versus simulation argument, nor whether a set of rules has a "root in reality." I can make the argument that while fun and streamlined, Combat Patrol has at least as much -- and perhaps more -- root in reality than other games.

To get directly to your point, many games use lots of charts and modifiers to give themselves the illusion of precision that doesn't exist. My philosophy is that any phenomenon that has less than a 5% impact on the outcome should NOT be reflected as a modifier. It is in the noise level and will be addressed through pseudo random numbers and the law of large numbers. Anyone who thinks they know enough about battle phenomenology to have a +2% modifier here and a -3% modifier there is deluding himself. (Again, remember I do this for a living, not just as a hobby.) Those minor variations are accounted for in the die roll.

Tank type A has a 75% chance of killing Tank Type B. If using dice, it doesn't matter how many times you fire with different tanks, that stays at 75%. With a deck if Tanks 1, 2 and 3 hit, tank 4 is more likely to miss, even if in a situation that is not affected by tanks 1-3.

First the chance of hitting a tank being 75% is an illusion of precision, not an unassailable fact. A game designer uses what information is available to predict as accurately as he can the chance of hitting a target at different ranges. So in your example, the designer decided that 75% was about right. Another designer can look at the same data and determine that the probability of a hit should the 80%, or 70%, or whatever. It is a FACT that different games have different chances of hitting the same vehicle at the same ranges, because each designer's analysis is based on different assumptions, determined by which elements he is choosing to represent and abstract. I argue that that number is really 75% +/- 5% (and perhaps much wider variance than that) no matter how accurately the designer THINKS he has analyzed available source data. So if the card-based (or any other mechanic) stays within that bound, I assert that it has LITTLE or NO impact on simulation, "rooted in reality," or fun.

In a card-based mechanic, depending on which cards are drawn in which order and how thoroughly the deck was shuffled at the start, the chance of hitting Tank 4 might just as likely go up as down if you don't reshuffle the deck after each draw -- and it may not change at all, depending on which cards have been drawn from the deck previously. Any changes in probability are very likely to be less than 5% -- which I assert is insignificant. It is true that dice are memoryless, but it is also true that almost all dice are biased because of flaws in manufacturing. Go ahead and float your dice in a cup of water and see. While casinos have gotten pretty good at un-biasing six-sided dice, no such quality control is typically done for your dice.

One thing I really like about the card-based mechanic is that people don't study the charts to "game" the modifiers. They just employ good tactics and play the game. Over several years of development, while not everyone's cup of tea, Combat Patrol has proven to be fun and a good simulation, where "good simulation" is defined as good tactics are rewarded and bad tactics are punished.

At the end of the day, if you have a visceral or philosophical objection to using cards to resolve combat results, then Combat Patrol is not the set of rules for you.

Buck Surdu

My apologies to all. I mistook the question in the OP as an honest request for information. I did not realize that it was nothing of the sort.

My experience with Combat Patrol, as stated above, is limited to a handful of games. However, I have gotten a reasonable view of the mechanics and where they take you through the course of the game.

In actual combat, it is the package of personnel, weapons, tactics, etc. that determine the outcome. The talent of an individual combatant or the capabilities of his rifle will have an influence on the end result. However, it is the PACKAGE of all of these things used together that lead to the outcome.

My observation is that the Combat Patrol rules system is similarly a PACKAGE. The card draw is simply a mechanic within the system. The sum total of all the other mechanics (i.e. activation, etc.) allows the PACKAGE to simulate combat. As the old saying goes; The sum total is greater than the combination of the individual pieces.

I'm very sorry if my comments come across as anything but honest enquiry. I am not pushing an agenda; I am genuinely interested in making a card resolved game work – I have played a very enjoyable 'pulp' game, Diamond Geezers at Salute, which first started making me look at the wider use. I have written a Carriage racing game (think F1 in 18th Century Paris) with cards. I have other games that are card driven.

To try and explain my concerns at their use allow me to use Black Powder as an example.

In BP it is usually 4+ to hit, 4+ to save. A Hit on 6 disorders, even if saved.

I have a deck, 1-6 10 times, 60 cards.

I fire, and draw 6, 6, 6.

This does two things.

1) Reduce the chance of future successes, as there are now only 27 out of 57 successes left- a 47% chance, rather than 50% chance.

2) reduce the chance of Disorder from 1 in 6 (10 in 60) -16% to 7 in 57 = 12%.

In fact drawing just 1 card from the deck reduces the next result being the same (succeed/fail) by 0.85%

Of course, bigger decks reduce this effect, but are in themselves unwieldy.

Reshuffle cards obviously allow the deck to be reset, but until they come up, the odds are still skewed.

The multiple uses per card doesn't change odds, except in the fact you go through the pack quicker, so reshuffle happens quicker, it just means you only need one deck, rather than multiple- again something that my Carriage racing does, so I am familiar with the concept. In fact it could mean that because all the Good shooting cards happen to come out, a unit will fail in shooting, for no other reason than units on the other side of the table took actions which, in reality, would not affect the firers ability.

In last nights game I needed to roll 6 or less on 2d6 – which I believe is 15/36. Any one success would have really hurt my opponent. I rolled 8+ 4 times in a row. BUT I know that was random- the odds never changed, they were not subject to actions unrelated.

I would like to use card resolution in a game- it can allow a subtlety in results not always available in a dice/table combo. But, and for my psyche it will always be a sticking point in historical games, the fact the deck has a memory, which is a pity when it comes to CP, because as soon as I saw it I downloaded the free rules.

To reduce card memory, take a deck, say 52 cards, and randomly only use 30 in a smaller deck. When you reshuffle,
shuffle in all cards again, and randomly pick out 22 and set them aside and use the deck of 30.

Last Hussar,

We all understand the phenomenon that you describe. The point is that the variation is small in most cases. While it is likely that every single card drawn after the first will have a slightly higher or lower chance of success (0.85% in your example) than the deck was designed for it won't vary much. And to paraphrase Mr Surdu's post, 50% hit chance is no more an exact model of the world than 49.15% hit chance is.

If, as you say, you would like to use card resolution, but deck memory is a sticking point for you then that's something that you have to come to grips with. I suggest that you take the +/- %5 guideline to heart. Most of the effect of deck memory falls withing that range. Worrying about +/- 0.85% is like putting a +1 modifier in a percentile dice system – it is far below a plausible level of resolution.

Duncan

The problem of deck memory is that certain results become more likely as you work through it- you know that I'm assuming. The problem with this is that's not how real life works.

I do know what deck memory is, but until you identified the mechanism (unit A killing unit B) and your referent model (uniform independent odds), I didn't know what the problem was.

I needed the clarification because parts of real life do have memory. I write a lot of scenarios that use cards to model those phenomena.

If the number of responses within each significant category is large with respect to the total number of draws, the difference is marginal. That's what all the reshuffle recommendations are about, keeping the difference in probability between uniform independent and convergent categorical close. You don't have to reshuffle every time, just frequently enough to the number of draws less than half the options within one category.

Last Hussar:

I inferred no nefarious intent in your comments. As you say, card-based resolution has a finesse to it that is difficult to achieve with dice and charts; however, it does display the memory property you seek to avoid. I think we are in agreement about that. The issue is one of degree and importance. I don't place as much weight on changing the probability of a hit a a couple of percent during the turn as you do. I think the law of large numbers will bring the results toward a central tendency over the course of the game, but you are right that the probability of success on some action is impacted by the cards that have bee drawn previously during the turn. Sometimes that will increase the probability of success, but sometimes it will decrease it.

There is a lot of personal preference in the selection of a set of wargaming rules, and it is possible for rational people to disagree.

I believe there is no solution to your dilemma. Without reshuffling after each draw (which defeats the speed advantage of card-based resolution) or having an infinitely large deck (literally), I think that any card-based resolution scheme will have the memory property you seek to avoid. If you discover a solution to this this problem, please share it with the group as I, for one, would be very interested.

Buck Surdu

P.S. I also very much appreciate the unusually civil tone of this thread. Thanks, all.

the probability of success on some action is impacted by the cards that have bee drawn previously during the turn.

This isn't true for all card based mechanics. I assume the cards are used to resolve a series of player optional actions?

Reshuffle cards obviously allow the deck to be reset, but until they come up, the odds are still skewed.

The probability of success on some action is impacted by the cards that have bee drawn previously during the turn.

This isn't true for all card based mechanics. I assume the cards are used to resolve a series of player optional actions?

I thought the question was avoiding card memory? If that is the only issue, then there are ways to obscure a player's ability to memorize the cards. I gave an example above. Obviously, another is just to have lots of cards, like the GMT card games like For the People with 110.

To try and explain my concerns at their use allow me to use Black Powder as an example.

In BP it is usually 4+ to hit, 4+ to save. A Hit on 6 disorders, even if saved.

I have a deck, 1-6 10 times, 60 cards.

I fire, and draw 6, 6, 6.

This does two things.

1) Reduce the chance of future successes, as there are now only 27 out of 57 successes left- a 47% chance, rather than 50% chance.

2) reduce the chance of Disorder from 1 in 6 (10 in 60) -16% to 7 in 57 = 12%.

In fact drawing just 1 card from the deck reduces the next result being the same (succeed/fail) by 0.85%

No argument there. Perhaps the question should be: Why is that a bad thing mechanically?

In some ways, how is that different from 'memorizing' the odds of rolling a six with two modifiers… other than the odds of success being far more variable, unlike dice, changing with each card drawn-- in other words, more difficult to calculate with continual changes in the odds?

The first question before deciding on game mechanisms, is what experience you want the players to have. That certainly includes what they can and can't do to calculate their chances of success in the game processes.

If you don't want the players to count cards, then there are ways to frustrate that. If you want the odds of something happening to remain constant, never changing, is that a significantly different experience from counting cards?

I thought the question was avoiding card memory? If that is the only issue, then there are ways to obscure a player's ability to memorize the cards.

I believe the OP is talking about the "memory" in the deck, not in the players. Last Hussar identifies player "card counting" as a separate issue later on.

I have been trying to figure out what the mechanic of concern is. Deck memory (categorical convergent probability in a Poisson process without replacement) has different effects on the gaming outcomes depending on the nature of the randomization artifact (a deck of custom(?) cards) and how it is implemented in the mechanism.

I think this is a deck of custom cards that are drawn for resolution of different classes (uniform independent probability?) combat actions that are directed by the players (i.e., the players choose when (in the deck) to implement what type of combat action).

In that case, I agree with all the above recommendations about big deck (buy an extra one or two, shuffle them together) and reshuffle occasionally (the more frequent, the closer to independent probability but also, as pointed out above, the more overhead), possibly when there is a lull.

I play a lot of card based boardgames and have found that players get an instinctive feel for times to reshuffle, split, and breed discard decks for recycling. I have also found this to be the case for card based mechanics in my wargames, though I do not have any mechanics like the ones that I think are the subject of this discussion.

etotheipi:

Thanks for the clarification. I was confused by
Does it get round the problem of 'Deck Memory' and the card counters? I assumed the card counters being the players.

Okay, if we are talking about a Poisson process, then the question would have to be what rate of variable occurrences we are shooting for? That would dictate what effects in gaming outcomes we are shooting for and could include on top of that. That would include having cards of different classes. Large decks do reduce the need for exactness in any of the above.

It looks like my search continues for the ultimate resolution continues [cue Hussar wandering into the sunset as credits roll].

To be clear, my concern with 'Memory' is that actions on one side of the board affect unconnected actions on the other side of the board: If the card turned for Squad A is a success, then the odds for Squad B gaining a success decreases. With the multi-role nature of CP decks, even unrelated actions will affect unconnected situations – "Because the Fire resolution card for Squad A had a 'Pass Morale" result on, Squad B are less likely to make their test.

I know the 0.85% is miniscule- but I was showing that 1 card in 60 has a larger effect that you may suppose. Once you multiply that by a few cards before reshuffle, you start getting into the realms of 5%+

(For clarification for those not having seem the terms before- Memory is the effect of the fact that a deck is finite, and drawing cards alters the proportions in the remaining pack. If the first 5 cards off the top of a poker deck of 52 are black, the proportion of cards left are 26-21 red-black, increasing the chance of a red being the 6th card. Dice do not do this- if you roll 10 6s on a d6, the chance of a 6 coming up on the 11th roll remains one in 6. A card counter is a person you remembers what cards have come out, thus allowing them to play the odds on the next result- in the above example if the prize is the same no matter what you guess, you'd bet on Red.)

While larger decks do reduce the problem (ie 10 decks in the above example gives a 260/255 split, the chance of a red is closer to .5 than 25/21) there is a level of practicality- larger decks are harder to handle.

Now I accept on a practical level from those who have played the game that they don't feel this matters. The problem is I would know. To me its like playing D&D where the DM says if a roll is identical to the one taken before, you must reroll – you can't crit, because I just did.

In a beer and pretzels game that wouldn't worry me, but I believe the author is trying to get a level of reality, and the mismatch between the maths and the aim makes my brain itch.

It strikes me that an app would solve the problem (yes, I know pseudorandom, but there are ways to make that immaterial), as you could have each results series in there as a seperate deck, and a random pick from the complete set each time you hit go. However gamers militate against the non-physical.

As to IABSM, Ditto, the cards are used to determine activation only. Each platoon and leader has a card, and activates when that is turned. there is also a reshuffle card, (tea break). this means a unit has a 50/50 chance of being activated in a turn. Despite what my usual opponent claims, the deck makeup doesn't affect this. If you have 1 of your cards in, and the enemy has 20, you still have a 50/50 chance of activating that one unit. Its just the opponent has a higher probability of getting at least one. Memory isn't a problem here – you don't want the same card to come out repeatedly.

To be clear, my concern with 'Memory' is that actions on one side of the board affect unconnected actions on the other side of the board: If the card turned for Squad A is a success, then the odds for Squad B gaining a success decreases.

Something that you have not caught onto yet is that in CP each player has his own deck. So, my draws do not have any influence on yours.

Duncan

I wasn't sure about that, I admit I haven't managed to read the rules fully yet, which is why I was asking how it solved the Memory effect. Does that mean I need two decks?

I tried to word what I said because I was unsure on the point, but it still stands- Actions of a squad affect the resolution of other squads