While I agree that the threat of fighters to a cloaker is overstated, and agree that the cloak effectively negates gatlings and fusions, I would still argue that the cloak is not particularly effective against the Hydran.

At least as important as the cloak itself is the _threat_ of the cloak. If you actually power the thing, you're using a big chunk of power, surrendering initiative, losing speed, etc. But your opponent must always account for the fact that you _might_ cloak. This places limitations on the tactics that can be used against a cloaker. For example, you never cloaked, but the Kzinti never launched his scatterpack, so the cloak had a big impact on the battle even though you never used it.

Against the Hydran, the threat of the cloak is less effective because it is such a reactive ship. Sure, you can't sell out and charge through the plasma for the overrun knowing that he might cloak, but it is generally a bad idea in the Hydran to sell out and charge against anyone. Other than that, cloak, don't cloak, and I can work with it.

Beyond that, actually cloaking against the Hydran is going to be tough. Hellbores are good for sub-hunting. Uncloaking too close to a Hydran is unhealthy. And sure, you can juke the fighters (I've had my fighters juked by a TKE going 16), but if the Hydran ship is nearby, it will not be "trivial".

By Richard B. Eitzen (Rbeitzen) on Wednesday, March 28, 2012 - 06:55 pm: Edit

Paul Scott: I made no statement about reasonableness. As for the zero damage, I was going by your post rather than actually checking the rule. Sorry about that.

Really part of my point is that people often talk of average damage as if they can EXPECT that. You cannot. It can help get an idea of what is going to happen, but it can also blind one to the possible effects of die rolls being off average. Admittedly in this example, even if the fighters did 8 points by astronomically good luck, it would probably make little difference.

By Paul Scott (The_Rock) on Wednesday, March 28, 2012 - 09:51 pm: Edit

Richard,
You are preaching to the Pope on that one. If this forum kept its archives, you search and find me discussing that long long ago. I discussed it on Schirmer's email list before that and no doubt Genie and/or rec.games.sfb before that.

Using the wayback machine, you can see me discussing it on my own site as far back as 2001 (http://tinyurl.com/c32csgy).

If you go over to the "unofficial boards" you will see me discussing it in detail with Peter (Peter doesn't think the distinction between averages and probabilities is significant).

I, personally, think it is the single most significant mistake people make when trying to think about tactics in this game. I have sent articles to ADB twice on the subject and they have decided not to publish them (presumably because they are too boring).

You will even notice, I hope, in none of my posts did I mention or discuss "average" damage.

To the extent, however, that you mean to suggest that any result is or should be accounted for, I do not agree. Use probabilities, not averages, to be sure. But also understand the concept of probabilities and thresholds. Here, for example, the Maximum damage two fighters could do over two turns is 16. The probability of that happening is less than four in one-hundred billion. I do not think anyone needs to concern themselves with that event. Almost more importantly, the difference between minimum damage (0) and the maximum damage limiting your probability to no more than 1 in 100 is 8.

When will 8 damage (which, again, I remind you is a 1 in 100 probability) matter? When will it matter so much that you would decide it is a better idea to stay out of cloak and just do your best without it? (hint - only when circumstances are such that you would not need to cloak anyway).

By Matthew Potter (Neonpico) on Wednesday, March 28, 2012 - 11:16 pm: Edit

I use average damage everywhere because when you're dealing with situations where you get a range of numbers, you have to stop somewhere reasonable. By picking the point where "You'll get this result or a better result half the time" or "You'll get this or a worse result half the time" is the best compromise. We could talk the best case or the worst case situations, but as Paul points out those may be deceptively hard to achieve.

It's very true that "average damage" is deceptive, as someone can roll a string of ones or even a string of fours, and throw your averages out of the ball park. Dice have no memory, and you aren't guaranteed to get some ones in your dice just because you've been putting alot of karma into sixes recently.

But it is the way to bet that you'll get a even sampling across all values of the dice. You can have an occasional good game or bad game, but you get a huge amount of average games.

Don't get me wrong. There is value in knowing the maximums and minimums. It gives you a range that may justify an idea even though the return is so low.

By Terry O'Carroll (Terryoc) on Thursday, March 29, 2012 - 12:58 am: Edit

---

Quote:

Dice have no memory

---

By Andrew J. Koch (Droid) on Thursday, March 29, 2012 - 07:55 am: Edit

There is a link in the header on this page that points us to the archived tactics discussions.
Everything is there back to 2003. Jean was kind enough to point that out to me the last time I complained about the Tactics Thread being wiped.

By Jean Sexton (Jsexton) on Thursday, March 29, 2012 - 09:15 am: Edit

We do try. And with PDFed documents I think you'll find that the search function does better.

By Euan Holton (Erjholton) on Thursday, March 29, 2012 - 10:09 am: Edit

From my point of view, it's faster and easier to discuss averaged results than go away and build probability tables for the range of results. It may not be precise or correct from a statistics point of view, but it's close enough for a rapid-fire discussion about a game.

By David Zimdars (Zimdarsdavid) on Thursday, March 29, 2012 - 10:39 am: Edit

Paul:

Thanks for the link on the wayback machine to your site with the tactics article. I found it an interesting read.

You mention two section headers at the top, "The Dangers of Doctrine" and "Advanced Comparative Advantage". Did you complete these and the wayback machine just dropped them? Or is this an unfinished work? (No problem, I certainly start many projects with the best of intentions with a far bigger scope than I have time for).

***

With regards to your classification of firing groups of weapons as resulting in something that can be thought of as an "average" a.k.a. a low variance continuous distribution (e.g. 6 phaser 1s at range 8) vs. a "probablistic" a.k.a. a high variance discrete distribution (e.g. 4 16 point overloaded photon torpedoes at range 8); I heartily agree. With regards to the objection to this thinking, is not so much that it isn't correct as it stands by itself, but I think stems from a lack of insight on what to do next in tactical thought that employs this particular insight. In other words, "I may have a certain probability to do better than the average (which I'll exploit later), I may do worse (which I'll have to deal with), so why not just use 'average' as a quick and dirty placeholder".

***

Reiterating or perhaps expanding upon Paul's descriptions.

An single volley which produces a low variance "average" has only one outcome. For example, 6 phaser 1s at range 8 produce 13 points of damage, plus or minuse a few.

A single volley which produces a high variance "probabilisitic" damage has multiple outcomes; and the game will bifurcate based on those outcomes.

For example: 4 photon torpedoes, fully overloaded at range 8 has (what seems like at first) 5 outcomes

0 hit, 0 points, 6%
1 hit, 16 points, 25%
2 hit, 32 points, 37%
3 hit, 48 points, 25%
4 hit, 64 points, 6%

Adding in the "average" damage doesn't change the number of outcomes or the probabilities:

0 hit, 13 points, 6%
1 hit, 29 points, 25%
2 hit, 45 points, 37%
3 hit, 51 points, 25%
4 hit, 77 points, 6%

If we consider these shots striking a 30 point front shield with 5 battery reinforcement there are arguably only 3 real outcomes (for a SINGLE volley):

A: 0 hit, 0 internals, 6%
A': 1 hit, 0 internals, 25%
B: 2 hit, 10 internals, 37%
B': 3 hit, 26 intenrals, 25%
C: 4 hit, 42 internals, 6%

Where A' is better than A (likewise B' better than B) but the arguement is likely in the short term the effectiveness is likely equivalent.

With regards to outcome A it is easy to see how they are equivalent as no shield damage is scored.

Outcome B is largely equivalent because 26 internals may score a few more weapons and power than 10, but it may not, due to the nature of the DAC (I call this a "shallow volley"). In a single volley, you have to get a really "deep volley" to overcome the DAC (first column weapon hits and hull) to really start hitting batteries, impulse, apr, shuttles, tractors and other stuff that starts to hurt.

Outcome C is starting to be a "deep volley". Again, either F or R hull is probably gone, and power hits (etc. see above) are starting to accumulate.

If the target were to have 10 reinforcement, then there are really only 2 outcomes, A- no damage 68% and B - shallow volley 32%. Really bad odds.

So as Paul said in his article, a shot on the rear shields makes a difference. An of course range 4. But also the Fed should probably make every tradeoff he can to try to catch his opponent when his turnmode is not satified, or when he is not moving every impulse, to try to get a mizia attack.

By Peter D Bakija (Bakija) on Thursday, March 29, 2012 - 12:10 pm: Edit

Paul wrote:
>> (Peter doesn't think the distinction between averages and probabilities is significant).>>

No, Paul, that isn't what I think at all. I think you being a pedantic prig is deeply irritating a lot of the time. But I certainly recognize the distinction between averages and probabilities.

By Paul Scott (The_Rock) on Thursday, March 29, 2012 - 12:25 pm: Edit

David,
I finished Advanced Comparative Advantage but the files are long lost and the Wayback Machine did not capture it prior to me dumping erols and losing my work. I then started work at a law firm and never had the time to remedy it. Even today I think about resurrecting the site but just have not gotten around to it (other than having saved my work locally). dangers of Doctrine never got past some drafts that I had not put up before dumping erols.

I like your categorization of shallow and deep volleys. I have just read this and thus not run any numbers, but intuitively I think your numbers are right. A quick look tells me that 10 internals has about a 60% chance of hitting a drone or torp and 26 has a 90% chance. So there is some significance between the two wrt heavy weapons. 26 internals still has a low probability of hitting a heavy phaser (except, of course, on the Hydran and LDR) but it is better than the almost 0 chance that 10 has. Probably not enough there to worry about. In short, I think there are some circumstances where you would treat that as 4 outcomes rather than three, but I otherwise agree.

"I think stems from a lack of insight on what to do next in tactical thought that employs this particular insight."

100% agree. Averages vs. Probabilities matters for nothing if you don't know where to go next. Thinking of outcomes as a probability matrix is only relevant if you also properly understand "thresholds" and objectives. That is the part that completes the sentence and allows deeper tactical thinking. If you do not make that step, then averages and probabilities are equally useful and averages are a lot easier.

By Paul Scott (The_Rock) on Thursday, March 29, 2012 - 12:48 pm: Edit

"I think you being a pedantic prig..."

Peter,
I don't know if the irony was intended, but I think that, in part, makes my point.

Both "pedantic" and "prig" indicate formalism for the sake of formalism. I am only passionate about this topic because I know that when I first starting thinking about tournament tactics in SFB, this shift in my own thinking was what allowed me to progress from an ok player to whatever I am now. There is nothing formalistic about my passion for this thinking. I am certain that someone who has not gone through this analysis and who starts thinking this way *will* become a better player.

I think it is the single most important thing a player can do to move from OK to good. The only other option I see is years and years of a lot of play against a lot of opponents so that the conclusions you would reach from analysis come from intuition instead.

By Peter D Bakija (Bakija) on Thursday, March 29, 2012 - 04:09 pm: Edit

Sigh. Oh, Paul.

By Andrew J. Koch (Droid) on Thursday, March 29, 2012 - 04:17 pm: Edit

Peter and Paul stop fighting. You're upsetting Mary.

By Andrew J. Koch (Droid) on Thursday, March 29, 2012 - 04:26 pm: Edit

Peter,
I don't mind Paul's analysis a bit. Some players are analytic like that, I would say Dave Z is another. Maybe Ken Lin also.
Then you have folks (like yourself, I think) that have gained Paul's insights through raw experience.
I don't analyze anything really, I am on the path to power through repetition, but I am interested in analysis from others. Stop throwing Paul off your lawn.

By Jean Sexton (Jsexton) on Thursday, March 29, 2012 - 04:36 pm: Edit

* Looks at the topic *

* Checks nametag to see if it says Mary *

"Prig" and "pedantic" are perfectly good words. Unless Peter is trying to imply that Paul is a fop which is a tad archaic, the words are in current usage. I would hate to equate a broad vocabulary with formalism.

Please do carry on discussing concepts. I think all of you have respect enough for each other to avoid personal attacks.

Jean
WebMom

By Paul Scott (The_Rock) on Thursday, March 29, 2012 - 04:51 pm: Edit

They are perfectly good words, though I am not sure "pedantic" adds much, if any, meaning to "prig." ;)

"I would hate to equate a broad vocabulary with formalism."

I would not either. What I meant by "Both "pedantic" and "prig" indicate formalism for the sake of formalism." was not that Peter was being a formalist for its own sake, but that "pedantic" and "prig" as descriptors suggest that the "pedantic prig" is a formalist for formalism's sake.

Anyway, enough (from me anyway) on vocabulary.

By David Zimdars (Zimdarsdavid) on Thursday, March 29, 2012 - 06:05 pm: Edit

Since I've returned to SFB, I've tried to ask a lot of questions in hopes of becoming a better player. I've always been very analytical and its an interesting phenomena to observe that some have obtained truly outstanding records (and I humbly observe, far better than myself). It is an interesting question to figure out how to play SFB to transform oneself from a a very good to a great player. This, I think is a very similar question that Paul Scott references asking himself. While there are many historically great players who do not seem to participate online (here or on SFBOL) anymore; I've been fortunate enough that many sill do; and so many of those have been forthcoming answering questions. I truly appreciate participation in conversations like these.

***

One impression I do have as a result of my analytical observations, is that while we have been fortunate to have so much opportunity to read term papers, "victory at articles" and discussions such as these; there is a relative paucity of truly analytic and thorough descriptions of "putting it all together". Much, but with a few exceptions, serves as a "preamble" or "bookends" to such a discussion.

For example, most term papers (not all for sure) describe fairly short term exploits (the mizia concept, or even the dips doodle, or fractional accounting some such). On the other side of the bookend we have Paul Scott (and surely others) discussion of "Comparative Advantage" and probabilities/averages which serve as a good preamble or frame for the synthesis what you want to do over the next 8 impulses, 1 turn, 3 turns, etc - what I might call "the action plan" and "the execution of the action plan".

One might argue that is just so simple that executing an action plan is just stringing together all the short term paper exploits in some sequence. But doing so and seeing how to do so requires experience, prior experimentation, and creativity. And it is in this arena of "between the bookends" that I think truly great SFB players excel, and for several reasons, I think it is also an area least communicated (not the least of which is you need a lot preamble, and to know all those term paper basics). Not being proactive with an action plan and seeing it quickly means either being reactive, or missing opportunities - and I think the best don't often do that.

As Paul said, concepts like "Comparative Advantage" let him take the next steps in becoming a better player. This was a good tool in building a better player - I'm trying to learn from those who forged ahead and see how they used those tools. It certainly would be a shame to have all this knowledge be lost to history. On the other hand, we are fortunate that so much of the basics (and advanced material) is captured in the Captains Logs.

By Peter D Bakija (Bakija) on Thursday, March 29, 2012 - 07:32 pm: Edit

Paul wrote:
>>I am only passionate about this topic because I know that when I first starting thinking about tournament tactics in SFB, this shift in my own thinking was what allowed me to progress from an ok player to whatever I am now.>>

That is all well and good. Just for the sake of clarity, 'cause clearly, you missed out on this, the only reason this digression is going on currently is 'cause you wrote this:

"Peter doesn't think the distinction between averages and probabilities is significant"

You could have *not* written this in this discussion, and this whole digression would have not ever happened. I wasn't involved in this particular discussion. I would not have questioned anything you wrote in this thread, and would have both read and likely left this whole thing uncommented on. But you randomly decided to make an irrelevant and, franky, kind of dicky comment at me for no good reason. So here we are. Please consider that next time you decide to do something similar.

Please continue with the discussion.

By Richard B. Eitzen (Rbeitzen) on Thursday, March 29, 2012 - 07:54 pm: Edit

I dropped out of this discussion because of Paul's inability to stay focused on what I actually said in lieu of putting words in my mouth. It's pretty sad.

By Andrew J. Koch (Droid) on Thursday, March 29, 2012 - 08:50 pm: Edit

Richard, dude, Paul did no such thing to you, unless the post was deleted. People are allowed to disagree with you, or even talk past you. Sometimes what you post does not get focused on.

By Jean Sexton (Jsexton) on Friday, March 30, 2012 - 09:16 am: Edit

Thin Ice Warning!

Please turn this topic back into a discussion of tactics. I don't want to hear commentary from Paul Scott about other people and I don't want to hear about Paul Scott's posting habits. Contrary to recent posts, this topic has nothing at all to do with people and everything to do with tactics. Kindly keep future posts focused on tactics.

Do not make me use the #8 Griswold tactic as applied to heads.

Jean
WebMom

By Brendan Lally (Brendan_Lally) on Friday, March 30, 2012 - 11:57 pm: Edit

Perhaps someone can provide a concrete example of where probabilities knowledge and their use has benefited their game vs their previous thought processes.

Also, since my high school math is rusty, would you mind showing your work on the probabilities calculations? (or direct me to a website)

I think I can do two variable probabilities (say with two bolts at R5):

Probability of both missing: 11.1%
Probability of one or more:88.9%
Probability of both :44.4%

Cheers

By Chris Proper (Duke) on Saturday, March 31, 2012 - 09:57 am: Edit

Combinations and Permutations. Yeah, those rusted out long ago.
Knowing the odds can be calculated is one thing. Spending more time running the numbers than flying your ship is another.
In the example you give:
The chance of 2 bolts hitting: 2/3x2/3=4/9=44.4%.
SFB is about asking the right question: What happens if I empty my plasma tubes against an enemy who is this close in front of me?
If you'll hit his only down shield for a bunch of internals and then get away you might want to do that. If he's going to anchor you with his remaining tractor and feed you 70 you need a better plan.
Knowing the probabilities is good, knowing the game is invaluable.

By Paul Scott (The_Rock) on Saturday, March 31, 2012 - 04:26 pm: Edit

Brendan,
Since I discussed this at length on the unofficial boards, I'll use an example from a RAT.

Two Feds meet at R3, 1 hex off perfect oblique. What should each fire (if they should fire at all) and why?

Here, btw, is a good link for doing Binomial Probabilities: http://faculty.vassar.edu/lowry/binomialX.html