New Score Formula?

[Ultimate Newbie]

I was talking to appoco the other day, when i had a brilliant (imo ) thought about how an alteration to the score formula would provide an opportunity for greater race balance, more race specialisation, and more room for different strategy.

Presently, the score formula is similar to:
Score = Value + XP*60

I was thinking of altering this to:
Score = Value + (XP*60)^n

n = the average ratio of the value of the attacker and defender in all past battles.

Now, presently XP is given to players based on the ratio of the attacker to the defender's value in each battle individually. n would effectively be the 'history' of a planet's XP gain giving a more long term approach to score.

ie, if a planet attacked 40 people that were exactly 0.4 of their value, then 1 planet twice his size and picking up some roids, his XP is still likely to be pretty high due to that one instance of a 'daring' raid, whilst really he is nothing more than a n00b basher. This system would look at the past performance of this player and thus determine that his history is poor, and thus assign a negative accelerator to his score - thus reducing his overall score (ie, minimising the bonus of that single good attack).

Similarly, players who consistantly attack well above thier own value recieve a very high accelerator which results in a massive positive impact on their score - and thus rank. This would mean that the instances of n00b bashing would go down alot, as players could potentially loose score by attacking them, and perhaps more importantly, give extra incentive for players to attack the most valuable planets in the universe (providing they can get through to steal roids, ofc) - this would result in alot of score and thus rank.

but hang on, i hear the voices in my head saying; didnt i say that it would lead to more race specilisation and balance? well yes, and this is how:

In this sense, score can be thought of as having three components; value, XP and n (my proposed section).

Each race can have race specific bonuses that emphasise one part of this formula (either bonuses that are like the production times and just hardcoded, or more like the natural progression of the stats, ie ziks = high value). In this sense, using Ziks and Caths as examples, Ziks inherrently have a high value component of their score - this means their XP and n parts of their score (especially n ) will be low, and thus forming an automatic way of reigning in their score.

OTOH, Cathaars who can consistantly attack planets much larger than them (however due to constant incoming among other things) have low value; they would have a low score component but a high XP and a very large n accelerator - thus thier score is automatically adjusted to be more balanced with the other races.

Similarly, Xans could be adjusted to have a high XP component, as they will regularly get through on a large valued planet, however they will need to 'farm' smaller planets more often than caths and thus have a lower n accelerator.

Terrans, being the general race ('lo noah ), miss out on any special bonus (unless testing determines otherwise ofc). I have an inkling that they would have a fairly high n as well, though not as caths, but a higher value generally.

This should lead to a more representative distribution of the races in the top 100 ranks at least - which last round were dominated by Ziks who just relied on their Value to farm their way to the top.

Anyway, this type of scoring should be particularly effective if suggestions like this one (post #6), where stolen ships are given twice the value, as it would increase value but over time reduce n and XP leading to a more balanced outcome.

It would lead to an (beneficial imo) alteration in strategy, as no longer would it be beneficial to XP 'whore'. Huh? you ask? That formula has wh0re written all over it. Well, actually, it doesnt. To wh0re successfully, a player needs to get a large XP gain consistantly - not just one big bang and receive a massive XP boost, as the accelerator on that would be small. It is potentially more beneficial to land multiple smaller attacks (gaining XP wise) in an attempt to maximise the n accelerator effect.

So, Thoughts?

[Tietäjä]

It would not be beneficial to XP whore, if your XP was squared? Think about this. An average higher ranking "XP whore" of last round perhaps had around 130k XP. If we tweak down zikonians, it won't be as easy, but the piece will be there. So let's assume we'll have an XP whore for 100k XP next round. Most of the time, he'll be picking maximum braveries, but we'll give in a bit on that too, let's make it 1,7 overall. His value is a meager 600k at the end ticks.

Score = Value + XP*60 = 6,6mil (6 mil for his XP, 600k for his value)
Score = Value + (XP*60)^n = 600k + (100k*60)^1,7 = 6,651,143,304 (for you a lot who do not like to read long numbers, that's 6,651 billion score; lot for his XP)

Compare it to an "average" player with say 1,2 for his total history (he's quite daring, often attacks people a bit larger than himself); he's pretty hot shot too, he managed to collect 8millions worth value. Maybe he has 60k XP in total (which is more than any player in my gal had this round, so he's quite fine).

Score = value + XP*60 = 8mil + 60k*60 = 11,6mil (3,6mil worth XP)
Score = value + (XP*60)^n = 8mil + (60k*60)^1,2 = 81,716,210 (that's 81million score, and a lot of XP in it).

Are we sure we want to play with ^n factors and XP? I mean, it will HEAVILY affect scores. Run a few calcs on your own on it, sovereign. Do tell me if I completely fked up these maths, but hey, maker r13 zikonian value round, and r14 xp whore score round. The accelerator is HUGE in any case. A very small difference in it results in massive score differences. Why? Because the root of the piece you are accelerating tends to be rather big.

Also, I expected you'll cap the n at 2. I can only imagine what 6mil^2 is. I know there are a lot of zeros involved.

[Ultimate Newbie]

Originally Posted by Tietäjä

Also, I expected you'll cap the n at 2. I can only imagine what 6mil^2 is. I know there are a lot of zeros involved.

3 600 000 000 000

well, assume n = (1 + m/100) then

This would give a score for your examples as;
XP wh0re of: 8 423 116 , and
Normal: 12 314 973

Somewhat better, i believe .

[Lοki]

Score = Value + XP*60

Wouldn't it be just better if you would adjust 60 to the race? That would simplify calculations and when you make 60 -> 80, a guy'd just gain 133% score (by xp). (100% is the current situation ofc) In this case n is 20: 60+20=80

The idea itself is nice. Though I think this change should not affect the calculation TOO much. Because every player with a high score would be an xp whore... depending on the race how much ofc (especially cath). With your calculation, the ziks would especially attack to steal ships, thus gain value=gain score. If they attack for xp, they couldn't keep up with the other planets. Noob bashing race...
I think we should just add a rather small n to 60, and change the ship stats as well.

[Gerbie]

Score=Value+Xp*60
XP=Sigma(1 to n))f(roids_gained, target_value/attacker_value)+g(cov_ops)+roids_initiated+h(scans)

If you change Score to depend on Xp*Sigma(1 to n))f(roids_gained, target_value/attacker_value) you'd get a 2nd degree relationship. (Score ~ (Sigma(1 to n))f(roids_gained, target_value/attacker_value))^2)

Obviously Xp would have to start depending solely on roids_gained (without bravery factor).

(soz about my not so clear way of noting this)

This change will most likely lead to PaX kind of results. And it's just not an easy thing to sell to your customers. It's too complicated.

[Tietäjä]

Originally Posted by Gerbie

If you change Score to depend on Xp*Sigma(1 to n))f(roids_gained, target_value/attacker_value) you'd get a 2nd degree relationship. (Score ~ (Sigma(1 to n))f(roids_gained, target_value/attacker_value))^2)

Obviously Xp would have to start depending solely on roids_gained (without bravery factor).

No, it's not too complicated. Not it does not depend solely on roids gained, without bravery factor. If you make it (the XP for attacks factor) (XP*60)^(1+n/100) where n is the average bravery factor, your score will instead depend more heavier on the bravery factor. Do a little simple mathematics and find out how.

[Heartless]

If you want to include a history of a planets "braveness", why not make it a simple factor?

score = value + (xp * mod * n)

mod = 60 at the moment.
n = the average ratio of the value of the attacker and defender in all past battles.

It would still be easy to calculate (considering your average value ratio is displayed somewhere) and easy to adjust.

[Nadar]

While we're on the discussion of score-formula: What about getting more score for roids/ships/xp? It got absolutely no other use than pleasing my eyes. At end of rounds the biggest planets is 30-40 million in score when they were several hundred millions in score before. But as I said, there's no need/use in it, just looks nicer

[furball]

Originally Posted by Nadar

While we're on the discussion of score-formula: What about getting more score for roids/ships/xp? It got absolutely no other use than pleasing my eyes. At end of rounds the biggest planets is 30-40 million in score when they were several hundred millions in score before. But as I said, there's no need/use in it, just looks nicer

Been suggested before. It'd be popular but make the db dump a fair bit bigger so was declined.

[ChubbyChecker]

The only problem I see with making bravery have a bigger impact on XP is that it will encourage XP whoring since the XP whores are comparatively much braver than planets who's value increases steadily throughout the Round. So if an idea like this is implemented then something should also be done about the XP whores.

Quite a few ideas have been suggested for this, such as introducing diminishing returns into how XP affects Score. Personally I think a simple idea like making Value have to be a certain minimum percentage of your total Score would work well. So eg, if Value has to be at least 20% of your total score then a planet with 500k value will not be able to have a total Score higher than 2.5mil until their Value increases.

[Nadar]

Originally Posted by furball

Been suggested before. It'd be popular but make the db dump a fair bit bigger so was declined.

Big? So PA cant handle that big dumps now, but they could in r4 with 180k planets?

[Ultimate Newbie]

Originally Posted by Chubbychecker

The only problem I see with making bravery have a bigger impact on XP is that it will encourage XP whoring since the XP whores are comparatively much braver than planets who's value increases steadily throughout the Round. So if an idea like this is implemented then something should also be done about the XP whores.

Originally Posted by Ultimate Newbie

It would lead to an (beneficial imo) alteration in strategy, as no longer would it be beneficial to XP 'whore'. Huh? you ask? That formula has wh0re written all over it. Well, actually, it doesnt. To wh0re successfully, a player needs to get a large XP gain consistantly - not just one big bang and receive a massive XP boost, as the accelerator on that would be small. It is potentially more beneficial to land multiple smaller attacks (gaining XP wise) in an attempt to maximise the n accelerator effect.

Quite a few ideas have been suggested for this, such as introducing diminishing returns into how XP affects Score. Personally I think a simple idea like making Value have to be a certain minimum percentage of your total Score would work well. So eg, if Value has to be at least 20% of your total score then a planet with 500k value will not be able to have a total Score higher than 2.5mil until their Value increases.

XP wh0ring should neccessarily be considered a bad thing. Indeed, in my first post, i suggested it as a means to achieving an alternative path to victory - as opposed for going R3 styles and going solely for 'Value'.

My stated formula is the opposite of diminishing returns. Sort of.

[Tietäjä]

Indeed. I agree with Sovereign about the XP whoring -- what makes it so bad, except for the evil eyes to conservatives give? It's not "old skool" PA, and it's after all even with the new style formulae suggested not really a way to win a round (you can go calc with the "best" XP whores of last round).

[coffee-]

I really would not agree that XP whoring is a bad thing. It is just simply something different from what is well known in PA. There will always be supporters, and critics of anything new, especially in the elitest views of many old school players.

[Mr Person]

I think the problem many people (including myself) have with XP whoring is that it is a fairly selfish way to play, as by definition the player is making himself of little or no help to others, and this idea makes me worry that the amount of people doing it could proliferate.

Perhaps that idea requiring value to be a minimum percentage of a player's score would help keep it in check a little (even if it was only 15% or so)

[Ultimate Newbie]

Originally Posted by Mr Person

I think the problem many people (including myself) have with XP whoring is that it is a fairly selfish way to play, as by definition the player is making himself of little or no help to others, and this idea makes me worry that the amount of people doing it could proliferate.

Whilst this is true, there is still definately a role for XP wh0res to play in a team based fashion. Whilst not being able to contribute with defence fleets, it still takes roughly 2-3 fleets to successfully (ie, making it too costly to land as opposed to not loosing roids) defend against every attacking fleet.

Thus, XP wh0res are helping to reduce the likelihood of defence being sent against other attackers, and other XP wh0res. Further, last round i tended to use highly offensive orientated planets (lo Ham ) to retal attackers. At one stage, i wasnt really defending against Terran BA incoming - instead, i was just sending two or three Cath CR fleets right back at them. The terran would then face the choice of either recalling his attack and thus making the caths recall, or call on his alliance for def (drawing anti CR away from other areas - which is important as that terran could have normally defended himself but now must draw on others), or draw defence from his galaxy (and with terrans lacking Frigates to flak with unlike Xans, it was more difficult for in-gal Bombers to assist and/or it would take more and thus more strain on defence etc etc).

Sending those fleets was little cost for my alliance as CR tended not to be very good at defending much, and most targets couldnt be attacked by them anyway - so many CR caths jumped at the opportunity to retal.

XP wh0res would also be ideal to use as retal, as it tend to be obvious to defenders whether the player is an XP wh0re and whether they were likely to land (ie, low value, high score = highly likely to land, thus low likelihood to recall and thus tying up a slot for the whole duration) etc.

Perhaps that idea requiring value to be a minimum percentage of a player's score would help keep it in check a little (even if it was only 15% or so)

Proportional scoring is an interesting idea as a limiting factor - but presently the max XP you can get from any roid is 20 - so that is an inherrent limiting factor anyway.

Its an interesting suggestion nonetheless .

[noah02]

Originally Posted by Nadar

Big? So PA cant handle that big dumps now, but they could in r4 with 180k planets?

DIdnt we get fed up of server error back then and constant stop starts etc...?
I know it use to do my head in when trying to log on and it was server error all the damn time lol

[Lοki]

Score = Value + XP*(60+n)

Wouldn't it be just better if you would adjust N to the race? That would simplify calculations and when you make 60-->20, a guy'd just gain 133% score (by xp). (100% is the current situation ofc) In this case n is 20: 60+20=80

The n-idea itself is nice. Though I think this change should not affect the calculation TOO much. Because every player with a high score would be an xp whore... depending on the race how much ofc (especially cath). With your calculation, the ziks would especially attack to steal ships, thus gain value=gain score. If they attack for xp, they couldn't keep up with the other planets. Noob bashing race...
I think we should just add a rather small n to 60, and change the ship stats as well.

I still think this is the easiest way to get things done

[Jaro]

It seems like many ppl do not like XP wh0res, especially those that do suicide missions. How about changing the XP calculation so that you get only a portion of the total XP depending on the ship value you lose in battle. For example, if you lose no ship value, then you get 100% of the total XP calculated. If you lose 50% of your ship value, then you get 50% of the XP. Lose 90% of your ship value, then you get 10% of the XP calculated. This should reduce the suicide missions and XP wh0ring.

But it may cause ppl to attack noobs again, and get near 100% XP due to low ship lose. So, to keep that from happening, we apply the n factor.

Just a thought.

[sale]

If you lose 50% of your ship value, then you get 50% of the XP. Lose 90% of your ship value, then you get 10% of the XP calculated. This should reduce the suicide missions and XP wh0ring.

i agree as some said this before loosing the fleet inst really a success,
with this addition, u can make sure this wont happen...

so y not??

[Ultimate Newbie]

Originally Posted by sale

i agree as some said this before loosing the fleet inst really a success, with this addition, u can make sure this wont happen...
so y not??

It goes back to the whole point about XP. XP was designed to be a reward for capturing roids which has been deemed to be the point of playing PA (as more roids --> more ships --> more roids etc).

By including an XP component seperate to value, you encourage more battles as its possible to loose value, but due to XP have a net gain in score. More battles leads to more roid transfers. more roid transfers reduces stagantion (which virtually everyone agrees is not fun).

R13 colourfully demonstrated the short-run benefits of XP whoring but the long term results given by pursing Value turned out to be superior (as most people realised fairly early on). This is why Caths were no-where to be seen, and Ziks dominated the top 100. Further, i think it must be said that XP whoring is not a bad thing. In my opinion, providing alternative paths to victory is a good idea as it introduces much greater depth of strategy into the game.

By limiting the amount of XP gained by the amount of value that survives a battle, you effectively negate the reason for XP existing. You might as well remove it and return to sim planet.

[Supersander]

If you lose 50% of your ship value, then you get 50% of the XP. Lose 90% of your ship value, then you get 10% of the XP calculated. This should reduce the suicide missions and XP wh0ring.

Originally Posted by sale

i agree as some said this before loosing the fleet inst really a success,
with this addition, u can make sure this wont happen...

so y not??

Now people land on a target even if they have the chance of losing a part of their fleet cause they awarded xp for it. If you remove this everyone will recall it's fleet with the slightest bit of def and nothing would be happening in the universe. Player should be encouraged to land even if there is def not disencouraged.
XP-whores are just a part of the game, and tho they never win in the long run they do make it more interesting and a challenge to def against.