Sitting like a Log - Manipulation of the PvP Point System

[Sparton_LOTB]

I wrote an excel spreadsheet, where you can run your own calculations. I'm not a excel pro but i tested it with spartons numbers, and the spreadsheet calculates correctly.

I took a quick look at your math; it looks like the same calculations I made, so that's a good basis to work off of.

i just played around a little and i like this approach: a flat one with -10 on lost attacks, and 15 points on a defense win, because the hyper good-o is better then the good with 500 attacks. so quantity matters here too. but that are only first testings so far. in this approach everyone profits aside the hyper-grabo noob :P

Hyper-Good-o also edges out a better good by a fair margin, and the lazy great... maybe that's OK, but I'd prefer that win rate mattered more.

Also, R1ck and BillLion basically touched upon this already, but I'm leery of any system that flatlines points earned from attacks, because it means that if you presume lower-ranked targets are easier, it encourages picking on lower-ranked (ie weaker) players, which just means it creates a divide between people who get to the top first and then everyone pushing everyone else out of the competition. Being at the top should have it's own advantage of being easier to place at the end of the week, and not create a feedback loop that makes it harder for others to compete.

i'm positive that we will find something that will work for all. it would be great sparton, if you could tell us exactly what you want to reach.

In general, I want all of the following to be true:

• Participation is valuable, without being a substitute to beat people with much better win rates
• Participation should not be discouraged
• You should want both a higher attack win rate and defense win rate
• You should be able to choose between risk of fighting a tougher opponent for better rewards, or fighting people closer to your rank for less rewards
• You should not be encouraged to fight people below you to optimize your climbing speed (assuming you're not already at the absolute top)

The only exception to the above is for people at the lower to lower-mid rankings; at that point, participation is assumed to be more of a factor than win rate (as the lower divisions are less competitive, and more about your own progress).

Also, the Saturday rush is pretty much meant to highly encourage participation while still having a good win rate; basically, participation becomes a tie-breaker at the end between players with similar win rates.

Let’s do some ShaolinMath-O

[...]

What happened in reality? He told me earlier he was up to 593 attacks – from 500 at last check – and yet again dropped below 3000. He’s working his way back up now – but the process will repeat. You can’t fight math assuming the 1:1 ratio holds up. His best bet is to sit tight and go all out Saturday before the attacks on him can catch up to the attacks he’s made.

I think it's important to keep in mind that the theoretical of 12 points for wins is not completely accurate unless he's so far above everyone else in rank that he never sees anyone within 480 points of him (in which case... if no one is ever within 500 points of him, that means he has 1st place in the bag, and all the theoreticals are moot).

I'll still agree that what we have now is making climbing... weird right now, at least for people jockeying for the top divisions. Curiously, most people competing in the middle have reported it much more sensible for them, so while changes need to happen, I hope we can maintain some semblance of what we have for mid- and lower-ranked players.

Yes – defense matters – but we’re forgetting something – THE DEFENSE IS CONTROLLED BY AI. It simply can’t compete. So to ding someone 20 points everytime they’re hit just because they were motivated enough to climb the ranks – while giving them 12 on wins - makes no sense.

The defense is controlled by the AI, but the influence of strong/anti-meta builds and strategies is heavily influenced by the player, which is why we can see a typical variance of 5% to 30% for players across the player base (assuming an appropriate sample size of defenses... but also after accounting for different quality of opponents based on ranking). Having a good defense relative to a poor defense means holding 6x more than some competitors (or 50-100% better if we assume many top competitors are around a 15-20% "good" hold rate).

I would think losses could be something like -7, -5 or -3 points depending how high above or below you the attacker is while attack points could remain close to what they are now. You'll still feel it but it won't be a freewill. And to Rick's pt keep offensive points on a scale prevents people from picking on lower ranked teams all the time.

This is pretty close (although more extreme) to the idea I'm thinking of of reducing defensive point losses by a fraction (such as 30% less losses regardless of division). This offsets the participation issue while still maintaining the risk/reward relative to how much higher or lower your opponent is ranked relative to you.