Sparafucil3

So backing up my last post, here is one final table comparing Standard, LHR, LHR (allowing 1 BLACK), and Pleva Rules. Over long the long haul, the Pleva rules are unbalancing. The tipping point is somewhere between 8 and 15 turns. I do not know for sure if that tipping point is also deck dependent. The effect of adding cards on additional draws is also bad. The LHR do a good job of remaining close to the Standard rules for all ranges but one: 1 FFE fire mission. That is by design. But that design causes some variance relative to the Standard model. It does an EXCELLENT job making sure you get at least one mission (something like 1 in 10 billion chance to not get at least one). But in pushing for at least two missions, it does skew the expected missions to the right. You can see that in the graph below. The LHR +1B (using the LHR but allowing for only 1 Black). Is exceptionally close to the Standard model. You can clearly see how the Pleva model falls apart on long games.

They are close. They are nearly identical when you allow for one BLACK rather than forcing two. IMO, the best OBA system would be one which used the LHR and allowed for a minimum of 1 BLACK card. Any second Red before the FIRST BLACK get's shuffled back into the deck. Any SECOND RED after one or more BLACK cards ends the module. I can now show that empirically. Interestingly, this is effectively the same as saying "First Card is automatically Black" by SSR. -- jim

I know WHY they were created. I haven't talked about that because I agree with the reasoning and I thought it self evident. However, IMO saying the only comparison that matters is comparing the zero fire mission chance isn't fair at all. Take a look at the 8 turn game I showed above and look at the distribution of zero, one, two, three and four plus fire missions. There are nearly 14% more three mission games using the Pleva and Carlsson/Borås variant. The Carlsson/Borås is even worse in the 4+ mission line. Your Leaflet rules how ever show a very similar distribution to the Standard module in the three- and four plus mission lines. I would bet the 2 mission line would be very close too if you allowed the module to expire after 1 Black instead of 2 Black. IMO, for games under 6 turns long, any of these systems are pretty much interchangeable. The Pleva and Carlsson/Borås systems don't eliminate the chance of a RED card--which is their stated intention--but they do GREATLY diminish the likelihood. In that respect, both of these system CAN--but rarely will--fail with their stated intention. But at about 8 turns, the Leaflet system begins to emerge as the better system. I was pretty sure the "add RED cards" strategy would break down over time. Now I know where that is. I was not sure ANY system would be close to the Standard model but I now know that's false. The Leaflet system is a damn good change. I will try to run a comparison allowing elimination of the model after the first BLACK to see if it is even closer.

Here is an 8B / 3R deck. Also, because of the way random numbers are generated in python, I can force the same string of numbers from one test run to the next. As such, the ONLY difference between one run and the next is the changing of the test parameters. The allows me to see the difference the parameter makes and not have to worry about the vagaries of random numbers being different from one run to the next. Also, let me explain the Borås deck just in case. It is similar to the Pleva mod except it starts shuffling 2 RED cards back into the deck on the SECOND RED card, not the first like the original. My group used to calls this "MOD PLEVA". Klas explained it to me as the Borås method so I use that term here. In the 6 turn comparison, all of the methods are fairly close to the Standard deck and they all do a good job of protecting against zero missions. In the 8 turn column, all of the variations are pushing the limits of "fairly close" to the Standard deck. They are close to being no good. This is particularly telling in the Three Missions and Four Plus Mission rows of the data. The Borås method in particular seems to be more deadly. Perhaps later I will run some experiments to see where these models might break down. The longer the game goes, the more the Leaflet method should be preferred though. It is the closest to the Standard rules in the 15 turn game so it's deviation from Standard will be the lowest over the long run. Melvin: would you be open to some errata on your Leaflet rules? The first sentence of rule 11 should read: When attempting Battery Access for an OBA Module and no more than one black and/or red chit(s) have been permanently removed from the Draw Pile and the second permanently-removed red chit is drawn, return it to the Draw Pile instead, and that ends the Observer’s OBA actions for that Player Turn .. The "and/or" there means the deck should end on 1B and 1R which is clearly not your intent.

No simulation is perfect. There are assumptions being made to simplify actual game play. For instance, I don't think there is any way to accurately reflect the decision to convert or not convert. There is no accounting for LOS or extra card draws. I have no mechanism to account for radio malfunction and repair. How often should I cancel and FFE:1 and start a more important fire missions? Those are things you'll have to factor in on your own. Any judgements I made on their game impact would just be more assumptions clouding the results. My biggest desire was to explore the idea of chit draws inside the concept of the game turns, something all the previous literature ignores. When doing that, all of the systems are reasonably close to the Standard rules so as to be considered the same. The Leaflet rules are the best at protecting against zero mission games. The Borås rules due a better job of adhering to the the frequency of the Standard rules. The Borås rules and Leaflet Rules also protect to the composition of the deck when it comes to extra card draw possibilities. I am not trying to change anyone's mind. All of these systems are close enough together so as to be the same within some narrow margin of error. Any of the modifications do an excellent job of preventing ZERO fire missions for a double RED. After that, it all comes down to taste, and there is no accounting for that in any kind of simulation.

Hey guys. I have been doing some thinking on OBA and Klas pointed me to this thread (and Google translate helped too ). I too proofed the article Klas mentioned earlier and I recently re-read the Medrow's article. What strikes me about the upcoming article and Medrow's is they speak in terms of "card pulls" without explaining or considering what that ACTUALLY means. A BLACK card represents THREE player turns of activity (at least). On the first player turn, you draw a card, you place an AR, and resolved to an SR. On the second player turn, you convert the SR to an FFE:1, resolve it, and flip to to an FFE:2. On player turn 3, you resolve the FFE:2 and flip to an FFE:C. Finally, on player turn 4, you can draw another card. To pull 6 BLACK cards takes AT LEAST 18 player turns, a FULL 9 turn game. As such, both of these articles are a little divorced from the game's realities. So I wrote some python code to simulate the games. What you see below is what I see. First, the assumptions: I am not considering situations that require an extra card draw I am not worrying about accuracy, or break down for radios or phones I am assuming the player is smart enough to place his AR in LOS Looking at the image, Game Length is in FULL ASL terms. Yes, 15 is a lot of turns but you need long games to get to the "mathematically expected games" which are determined using "card pulls" Contact Num -- the number to roll <= for Contact. The script does account for the Maintenance DRM The script does account for loss of Contact when conducting actions Conversion Rate: this is a randomizing factor for conversion from SR to FFE:1. Where is 100, assumes you convert EVERY chance you get. Where it is 50, it assumes you would convert it 50% of the time or move it 50% of the time. IMO, 50% is too high but I am not sure what a more realistic number is for it. Black Cards is the number of Black Cards drawn on average. Zero through 4+ is the number of FFE:1's placed. On the Zero mission line, you can see things like .01 (7). This means .01% of 100K games resulted in zero missions. The number (7) is the actual number of games. Only a 5 Black/2 Red deck is being considered here. With those constraints, you can see the Leaflet rules are hands down better for long games. The correlate very closely to the Standard Rules. But those "Long Games" are 15 turns here. When we consider more realistic games (6 and 8 turns), the numbers become much more consistent across all the systems. In the 6 - 8 range, I think the Borås system is actually closer to the Standard rules. Each system is close in terms of Black cards, but the Borås system matches the zero/one/two/three/four plus mission total more closely than the other three. Again, this is just a 5B / 2R deck. I plan on completing 8B / 3R with a Contact Num of 8 (standard German) and then the American 10B / 3R (contact 8 ) tomorrow. After that, I may do all of the card draw possibilities. but only under "game conditions" . -- jim