Steve Jackson Games - Site Navigation
Home General Info Follow Us Search Illuminator Store Forums What's New Other Games Ogre GURPS Munchkin Our Games: Home

Go Back   Steve Jackson Games Forums > Roleplaying > GURPS

Reply
 
Thread Tools Display Modes
Old 04-18-2018, 07:27 AM   #11
weby
 
weby's Avatar
 
Join Date: Oct 2008
Default Re: High defenses in 4th edition

Quote:
Originally Posted by Tomsdad View Post
That I'm not sure about though?
I know, feint results are not fully intuitive.

Quote:
1). Any spikes can go in either direction with equal skill. Although this does not negatively affect the person initiating the fient if it goes against them, other than wasting their feint. I.e they can't do worse than get no effect
Defenders really good rolls do not matter. If the defender succeeds by more than attacker the result is 0 not a attack penalty.

Quote:
2). I'm not sure how common it is to get such high relative MoS with equal skills in the lower range.
I dug out my old spreadsheet with the calculations:

Five or more difference in attackers favor comes about 14.5% of the time and when starting really high defensive values even smaller reductions help a lot and on average a feint will give you 1.68 benefit.

But at defense 13 or higher at equal skill feinter will be better off.

At defense 14 he will get through defenses on average of about 13.7% vs the straight attacker 11.1%, so is about 1.24 times as likely to hit.

At 13 the numbers are closer at about 18% for feinter and 16.2% for straight attacker so is about 1.11 times as likely to hit.

I put up a copy of the sheet in https://docs.google.com/spreadsheets...it?usp=sharing.

It is not the clearest(did not polish it for publishing) but the "% at least this much" on the right side is how often you will get at least success by that many more than the defender on feint and the defense scenarios fro defense 14 and 13 are below on left. The numbers are also approximate as the dice probabilities are rounded to 2 digits after decimal point.
__________________
--
GURPS spaceship unofficial errata and thoughts: https://gsuc.roto.nu/

Last edited by weby; 04-18-2018 at 07:32 AM.
weby is offline   Reply With Quote
Old 04-18-2018, 08:10 AM   #12
ericbsmith
 
ericbsmith's Avatar
 
Join Date: Aug 2004
Location: Binghamton, NY, USA. Near the river Styx in the 5th Circle.
Default Re: High defenses in 4th edition

Quote:
Originally Posted by weby View Post
I know, feint results are not fully intuitive.
No, they're not. But, basically, if you are equally skilled you have the breakdown is roughly 45%/45%/10% - 45% of the time you will succeed with a MoS of 1+, 45% the defender will succeed with a MoS of 1+, and 10% of the time there will be a tie. Any time the defender wins or there's a tie the feint was a failure, so really it's just 45% in the attackers favor. And, obviously, the 45% is split between MoS 1, 2, 3, etc....

These odds change dramatically if the attacker has higher skill than the defender; if the Attacker has skill one higher than the defender the split is 55%/36%/9%, if he has skill two higher it's 64%/28%/8%. This means that feinting really works best if you have a higher skill than the opponent, even a margin of only one higher makes a big difference.

However, the one advantage that feinting does have is that you normally know whether or not the feint was a decent success, meaning you can wait until you have a reasonably high MoS before you strike.


Quote:
Originally Posted by weby View Post
Five or more difference in attackers favor comes about 14.5% of the time and when starting really high defensive values even smaller reductions help a lot and on average a feint will give you 1.68 benefit.
14.5% concurs with the numbers I was running. Which is not a great chance of success, but as you say is a better chance than just trying to attack vs his high defenses. You also have to consider what the effects will be on his defenses on the following turn. If you Feint and get a MoS of just 1 the next attack will be 14-skill vs 13-defense (assuming starting with 14-skill and 14-defense), which increases the attack success chance from 10% to 16% (13% vs 20% with deceptive attack). If you get MoS 2 on the feint the roll will be 14 vs 12, which gives a 29% chance of success with a Deceptive Attack of 13 vs 11. So any MoS victory on a Feint will increase your chances of successfully attacking on the following turn, and with a good MoS on the Feint you can dramatically increase the chances of succeeding.



As an aside, another interesting tidbit I found regarding Deceptive Attacks is that it is almost never advantageous to take an attack penalty that reduces the attackers skill skill below 12. This is because the bell curve is so steep from 12-11-10-9 that any margin by which the attacker is reducing their skill almost certainly is worse for for the attacker than the margin by which they are reducing their opponent's defenses. So take a penalty that reduces your skill to 13 or 12, but not any lower than that.




I'll also add, as others have said, that the problem here is one of tactics. If the defender is constantly All Out Defending then he's not even trying to win the fight; change the attack tactics to something that will change the results dramatically such as a shield bash. If he is constantly retreating then make him retreat into a corner or rough terrain or a fire pit or into one of your companions. Or back off from him far enough that he has to do a Move and Attack just to try to get to you... then taunt him for being a coward. Or take an All Out Attack (Determined) for the +4 to hit, and use that +4 to Deceptive Attack to give him -2 defenses. Change it up and do something different.
__________________
Eric B. Smith GURPS Data File Coordinator
GURPSLand
I shall pull the pin from this healing grenade and...
Kaboom-baya.

Last edited by ericbsmith; 04-18-2018 at 08:35 AM.
ericbsmith is offline   Reply With Quote
Old 04-18-2018, 09:40 AM   #13
Tomsdad
 
Join Date: Jan 2010
Location: Brighton
Default Re: High defenses in 4th edition

Quote:
Originally Posted by weby View Post
I know, feint results are not fully intuitive.



Defenders really good rolls do not matter. If the defender succeeds by more than attacker the result is 0 not a attack penalty.
No I know that's why I made the point. But they do matter because they negate the attackers really good rolls, and they are just as likely to get them.


Quote:
Originally Posted by weby View Post
I dug out my old spreadsheet with the calculations:

Five or more difference in attackers favor comes about 14.5% of the time and when starting really high defensive values even smaller reductions help a lot and on average a feint will give you 1.68 benefit.

But at defense 13 or higher at equal skill feinter will be better off.

At defense 14 he will get through defenses on average of about 13.7% vs the straight attacker 11.1%, so is about 1.24 times as likely to hit.

At 13 the numbers are closer at about 18% for feinter and 16.2% for straight attacker so is about 1.11 times as likely to hit.

I put up a copy of the sheet in https://docs.google.com/spreadsheets...it?usp=sharing.

It is not the clearest(did not polish it for publishing) but the "% at least this much" on the right side is how often you will get at least success by that many more than the defender on feint and the defense scenarios fro defense 14 and 13 are below on left. The numbers are also approximate as the dice probabilities are rounded to 2 digits after decimal point.
Cheers for the link

OK I'll be honest I'm really not seeing how you get your figures for 5+ @ 14.5%

Have you added up the individual chances of each combination of one winning by 5, so 'attacker rolls 3, defender rolls 8' + 'attacker rolls 4, defender rolls 9'.... etc, etc?

I'm not sure that right, because half those outcomes will be the defender wining by 5?

I.e your probability table is the total of all possible combinations of the QS, but the results don't distinguish between who get's what score.


So the average difference may well be 1.68 MoS between the two sides in the QS, but that's not the same as saying the fienter wins by an average of 1.68 MoS over the defender.

I'm just not seeing how two equal skill QS and net MoS can have net benefit for one over the other. Which your figure of 1.68 suggests. As in all instances the defender is just as likely to get the same range of MoS as the attacker. And while the net MoS position can't improved the defenders result, it does effect the attackers.


(am i just going mad here, it's been a long day ;-), I could well be wrong!)




EDIT, yes, yes I am ignore the above, I see what you've done you have only counted half the results so only the attacker getting net 5+ MOS sorry!

EDIT, EDIT, I still think I'm right about the 1.68 thing though, I just can't see how one side can be net favoured here! gahh I'm having a stats senior moment, but is it that in terms of overall likelihood of end result all these chances of the feinter getting 5+ MoS (or any MoS) get countered by the same number of chances of the defender getting 5+ MoS (or any MoS)

Anyway a couple of points,

At higher defences small reductions aren't actually that effective. Because the chances of successfully defending don't drop as much as they do closer to the midline of the distribution.


e.g 16 dropping to 15: 16 or less is 98.15%, 15 or less is 95.37% that is a 2.78% drop in successful defence chance.*

11 dropping to 10: 11 or less is 62.5%, 10 or less is 50%, that is 12.5% drop in successful defence chance



Then there's the point that while you might see an individual increase in success per attack which is preceded by a feint, you have half the number fo attacks to add this benefit to




*the trick here is although your comparative chances of getting past their defence more than doubles (from 1.82% up to 4.27%) you still have very little chance of actually getting through. Look at 11 dropped to 10, your chance of getting through goes from 37.5% to 50% which is only a 33% proportional increase, but in terms of overall increases in chance of it happening it's a way bigger change.

Last edited by Tomsdad; 04-18-2018 at 10:22 AM.
Tomsdad is offline   Reply With Quote
Old 04-18-2018, 10:16 AM   #14
Kromm
GURPS Line Editor
 
Kromm's Avatar
 
Join Date: Jul 2004
Location: Montréal, Québec
Default Re: High defenses in 4th edition

Comparing editions, you shouldn't see that big a difference.

In Third Edition, Parry was (skill/2) and low-tech fighters could add up to +4 for armor PD and up to another +4 for shield PD. Thus, active defenses were (skill/2) at +0 to +8, or (skill/2)+4 on average, which reflected the typical skirmisher with a PD 2 small shield and PD 2 leather armor.

In Fourth Edition, Parry is (skill/2)+3 and low-tech fighters can add up to +3 for shield DB. Thus, active defenses are (skill/2) at +3 to +6, or (skill/2)+4.5 on average. Not really a big difference. That fighter with a small shield would get DB 1 and have the same (skill/2)+4.

Either edition allows a single retreat. This gave +3 to all parries in Third Edition. It gives only +1 in most cases in Fourth Edition, save for a subclass of parries that Third Edition would call "fencing parries" . . . those still get +3. But in Third Edition, those parries were at 2/3 skill and not 1/2 skill, so at skill 14, they'd start at 9 instead of 7, and be at an effective +2 anyway. No real change.

Fourth Edition does allow a fighter to use All-Out Defense to get an extra +2. That's new. But it also means that fighter isn't attacking. And Fourth Edition introduces Deceptive Attack, which effectively means that someone with skill 14 could still attack at 12 or 10 to give -1 or -2 to defend, canceling out the bonus.

All told, the two editions are fairly evenly matched. At the high end, Third Edition actually makes it much easier to defend, as you can have a low-tech fighter with total PD 8 retreating for +3 and effectively rolling (skill/2)+11, while in Fourth Edition that's DB 3 and retreating at +1, for (skill/2)+7, or +9 with All-Out Defense.

I think the key difference to note is Deceptive Attack, which in my experience results in more hits in Fourth Edition.
__________________
Sean "Dr. Kromm" Punch <kromm@sjgames.com>
GURPS Line Editor, Steve Jackson Games
My DreamWidth [Just GURPS News]
Kromm is online now   Reply With Quote
Old 04-18-2018, 10:38 AM   #15
ericbsmith
 
ericbsmith's Avatar
 
Join Date: Aug 2004
Location: Binghamton, NY, USA. Near the river Styx in the 5th Circle.
Default Re: High defenses in 4th edition

Quote:
Originally Posted by Kromm View Post
I think the key difference to note is Deceptive Attack, which in my experience results in more hits in Fourth Edition.
Far more. Mathematically Deceptive Attack almost always gives increased chance to hit for the attacker[1]. In my experience players almost always juggle the Deceptive Attack penalty with hit location and other penalties to maximize hit chances (my players really like the Vitals at -3 for the extra damage). Deceptive Attack is the first combat option I generally teach new players now.


[1]Down to attack Skill-12. Below 12 it depends on how high the Active Defenses are.
__________________
Eric B. Smith GURPS Data File Coordinator
GURPSLand
I shall pull the pin from this healing grenade and...
Kaboom-baya.

Last edited by ericbsmith; 04-18-2018 at 10:41 AM.
ericbsmith is offline   Reply With Quote
Old 04-18-2018, 10:51 AM   #16
weby
 
weby's Avatar
 
Join Date: Oct 2008
Default Re: High defenses in 4th edition

Quote:
Originally Posted by Tomsdad View Post
No I know that's why I made the point. But they do matter because they negate the attackers really good rolls, and they are just as likely to get them.
Yes, but defender good rolls do not matter as long as they are better than the attacker, the degree over than number has no effect.


Quote:

Cheers for the link

OK I'll be honest I'm really not seeing how you get your figures for 5+ @ 14.5%

Have you added up the individual chances of each combination of one winning by 5, so 'attacker rolls 3, defender rolls 8' + 'attacker rolls 4, defender rolls 9'.... etc, etc?

I'm not sure that right, because half those outcomes will be the defender wining by 5?
All situations where the defender rolls same of better than the attacker are the same. The attacker gets a bonus of 0. Thus you only count how many attacker rolls are better as a % if the total rolls.

Quote:
So the average difference may well be 1.68 MoS between the two sides in the QS, but that's not the same as saying the fienter wins by an average of 1.68 MoS over the defender.
As said when the defender wins/ties there is no effect. When the attacks wins he gets a bonus.

Quote:
I'm just not seeing how two equal skill QS and net MoS can have net benefit for one over the other. Which your figure of 1.68 suggests. As in all instances the defender is just as likely to get the same range of MoS as the attacker. And while the net MoS position can't improved the defenders result, it does effect the attackers.
Because good defender rolls only count as 0, but good attacker rolls as +1 to +15 or so. It all adds up.


Quote:
EDIT, yes, yes I am ignore the above, I see what you've done you have only counted half the results so only the attacker getting net 5+ MOS sorry!

EDIT, EDIT, I still think I'm right about the 1.68 thing though, I just can't see how one side can be net favoured here! gahh I'm having a stats senior moment, but is it that in terms of overall likelihood of end result all these chances of the feinter getting 5+ MoS (or any MoS) get countered by the same number of chances of the defender getting 5+ MoS (or any MoS)
Because as said only attacker good rolls have a meaning of how good, defender good rolls maximum bonus is cutting attacker bonus to 0.

Quote:
Anyway a couple of points,

At higher defences small reductions aren't actually that effective. Because the chances of successfully defending don't drop as much as they do closer to the midline of the distribution.
It it the same thing for deceptive attacks.

Quote:
Then there's the point that while you might see an individual increase in success per attack which is preceded by a feint, you have half the number fo attacks to add this benefit to
Yes thus in the calculated examples I have things like:
Total hit probability 34.14%
hit probabilit/round 17.07%
and the 1.85% critical chance is there as:
attacker crits 0.93%
thus getting the total/ round of:
total hit probability/round 18.00%

Whereas on the strike the things would obviously be 34.14%+1.85%=37.99%.
__________________
--
GURPS spaceship unofficial errata and thoughts: https://gsuc.roto.nu/
weby is offline   Reply With Quote
Old 04-18-2018, 10:57 AM   #17
Anthony
 
Join Date: Feb 2005
Location: Berkeley, CA
Default Re: High defenses in 4th edition

Quote:
Originally Posted by Kromm View Post
I think the key difference to note is Deceptive Attack, which in my experience results in more hits in Fourth Edition.
Quite a lot more. Third edition at high skill levels usually turned into fishing for criticals because that was the only way you ever hit anything.
__________________
My GURPS site and Blog.
Anthony is online now   Reply With Quote
Old 04-18-2018, 11:12 AM   #18
GodBeastX
 
GodBeastX's Avatar
 
Join Date: Dec 2008
Location: Behind You
Default Re: High defenses in 4th edition

Feints compounded on Deceptive Attack are deadly. It's rare to see defenses hold for very long in play.

Feints have the advantage of success giving you a rough idea of how badly your enemy messed up. When you add this to deceptive attacks and hit locations a successful feint is usually doom for the defender. Vital Hits and attacks of these nature almost always follow feints in our game as enemies struggle to get a critical roll on any defense necessary to stop incoming attacks.

We also started playing with an optional roll to trade skill for damage in one game because we'd end up with some feints and enemies having no defense and high DR or mitigations. Rapid strikes don't always cut it so we just decided it's probably believable someone can telegraph for power or take a wild swing with all their might.
__________________
RPG Jutsu.com - Ninjas Play GURPS
GodBeastX is offline   Reply With Quote
Old 04-18-2018, 11:40 AM   #19
Anthony
 
Join Date: Feb 2005
Location: Berkeley, CA
Default Re: High defenses in 4th edition

Quote:
Originally Posted by GodBeastX View Post
Feints compounded on Deceptive Attack are deadly.
It's actually somewhat rare for feints to be useful. The main use is when you want a called shot against someone with significantly lower skill than you, otherwise two attacks is usually better than one feint and one attack.
__________________
My GURPS site and Blog.
Anthony is online now   Reply With Quote
Old 04-18-2018, 12:51 PM   #20
Tomsdad
 
Join Date: Jan 2010
Location: Brighton
Default Re: High defenses in 4th edition

Quote:
Originally Posted by weby View Post
Yes, but defender good rolls do not matter as long as they are better than the attacker, the degree over than number has no effect.




All situations where the defender rolls same of better than the attacker are the same. The attacker gets a bonus of 0. Thus you only count how many attacker rolls are better as a % if the total rolls.
Yes but there's also the situation when the defender doesn't roll better ('better' = get more MoS) than the attacker but does get successes, those successes reduce the attacker's success total and thus the feint's effect

This is why neither side is favoured in terms of probability of net result here (although ties go to the defender). Because if the skill levels are the same there is an exact same probability of MoS for each. In terms of probable net outcome they cancel each other out. The attacker has to get more MoS than the defender, now there is a chance of that happening (ie combinations of results as you table shows) but there are an equal frequency of the opposite combinations occurring. And because the skills are equal and thus chances of MoS are equal those frequencies are equal.

I think you are hung up on there being no equal and opposite effect in terms of an end result in regards to penalties inflicted on a failed attacker. I.e the attacker winning by net 3 MoS and the defender winning by Net 3 MoS are not the same in terms of who get's what penalty. But that doesn't matter it's that each one has the same frequency or distribution of MoS that matters, because it's the net margin of victory of both of their roles that determines the effect of the feint.


Quote:
Originally Posted by weby View Post
As said when the defender wins/ties there is no effect. When the attacks wins he gets a bonus.
but that bonus can still be reduced by the defender's successes even if the defender didn't win overall

Quote:
Originally Posted by weby View Post
Because good defender rolls only count as 0, but good attacker rolls as +1 to +15 or so. It all adds up.
Hang that's not right, a defender's good roll negates an attacker's good roll.

Because it's the net MoS of both that gives the final result, not just the attackers MoS. So say the attacker gets 5 MoS on their roll, and the defender gets 4 MoS on their roll, end result isn't a -5 feint, but a -1 feint.

If the Defender got 5 MoS the feint would be negated. Now if the Defender got 6+ at that point there's no further positive benefit, but there doesn't need to be one because the feint has already reached it's maximum failure state. I.e it can't get any less good than no effect and a wasted action.


If you and your foe both succeed,
but you succeed by more, subtract
your margin of victory from the foe’s
defense. For instance, if your skill is 15
and you roll a 10 (success by 5), and
your foe’s skill is 14 and he rolls 12
(success by 2), you win by 3, so he will
defend at -3 if your next maneuver is
to attack him
Campaigns pg365


Quote:
Originally Posted by weby View Post
Because as said only attacker good rolls have a meaning of how good, defender good rolls maximum bonus is cutting attacker bonus to 0.
Right but how good it is at doing that is function of how good it is (and how much it needs to be is a function of how good the attackers was).




Quote:
Originally Posted by weby View Post
It it the same thing for deceptive attacks.
In terms of reducing the chance of successful defence by penalising by 1 yes, but the effect of the DA is different than the Feint because it doesn't have the QS that evens the effect out.

And unless you have really high defences most defence scores are grouped closer to the bell curve peak, and generally speaking most attacking scores are higher than them so you tend to get more 'bang for your buck' trading 2 off you attack for 1 of your target's defence.

But yes the closer your target's defence score is to your attack score the less benefit you see from DA's (unless you are over 17 on the attack in which case it's a free penalty even if it's not a big one)


Quote:
Originally Posted by weby View Post
Yes thus in the calculated examples I have things like:
Total hit probability 34.14%
hit probabilit/round 17.07%
and the 1.85% critical chance is there as:
attacker crits 0.93%
thus getting the total/ round of:
total hit probability/round 18.00%

Whereas on the strike the things would obviously be 34.14%+1.85%=37.99%.
sorry I may have missed that example, but I'm not sure that works, 34% every 2nd round is not the same as 17% each round for two rounds? You're thinking in terms of an abstract average hit per round, but that's not the same.

even leaving aside the effects of multiple chances, you have two chances of critical hit and that chance is likely static for each event.

Either-way in terms of comparing it to two attacks it really depends on how much you increase your chances of getting through with a feint. That depends on where the defence scores are on the bell curve and the chances of net MoS on the feint.

That's before we get into questions of what happens if person A hits person B and inflicts a shock penalty after person B feints but before that feint is capitalised on (what you gain in penalising the opponent's defence you might well lose in penalised attacks, and that's leaving aside the other issues of being hit!)

Last edited by Tomsdad; 04-19-2018 at 12:11 AM.
Tomsdad is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Fnords are Off
[IMG] code is Off
HTML code is Off

Forum Jump


All times are GMT -6. The time now is 06:22 PM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2024, vBulletin Solutions, Inc.