Defence battles statistics

[OmerB]

I'm wondering how many players attack me when I'm with different defensive teams and boosters, since we don't have this data.

Kinkoid should display statistics, like the amount of challenges that were done against me, and the average of those. Right now, I can do tricks and flip flops, putting interesting teams on defence, but I don't get any feedback... 🤔

[DvDivXXX]

Well, Kinkoid doesn't give you your stats for attack either. The script does that. ^^ So you might have a better shot suggesting this in the Community Script thread. Although I have no idea if they can script that, or how much extra work it would be for either Tom or Zoo (or anyone else into scripting who might contribute), or how much extra loading time the result might come with.

[Horsting]

All raw data the HH++ script has is from the game. It just visualises it and does battle simulations based on it to calculate the chances and average score to expect. This is also the reason why it is (or was) affected by the game bug that opponents crit chance was sometimes shown wrong and girls' gear not correctly applied to stats, leading to wrong shown values.

There is no way one can script stats about attacks against you, as this data is not available to the frontend (the HTML DOM and JavaScript parts). If KK made this data available, or created an API to request it from the server, it could be shown by scripts.

This info would be interesting indeed, but not sure whether it is important or would affect your defence team in league? You (both) have much more experience than me, so I am curious, but here is what I am thinking:

• The team is "cached" for opponents for 6 hours, so whatever changes you do within such a 6 hours window has only an effect for your attacks, not for your opponents attacking you.
• As counter bonuses are pretty strong, you usually always want to have a balanced defensive team. A colour team only makes sense in some blessings weeks, like the massive dark bonus we had recently, so that a balanced team and/or an opponents counter team would just be too weak, despite the counter bonus.
• And you want the strongest boosters up that you can effort, as naturally one has no effect on when an opponent attacks you, but must expect opponents to attack you as fast as you enter with no or lower boosters into a 6h window.

[Master-17]

Theoretically, if all players switched to using the script, then it would be possible to arrange the collection and analysis of statistics somewhere in the cloud, but this is unrealistic))

[DvDivXXX]

13 hours ago, Horsting said:

The team is "cached" for opponents for 6 hours, so whatever changes you do within such a 6 hours window has only an effect for your attacks, not for your opponents attacking you.

That's true but on top of that there are also random partial snapshots in between the global snapshots that happen every 6 hours. They're random in when they occur, how many players they affect (although it's typically around 5-10% of the group). That's unpredictable and I'm guessing this data is just as inaccessible for user-made scripts as the rest.

Makes sense that it wouldn't be feasible at this time. And true, technically the data shown by the script is made "available" by Kinkoid, but without the script it would be nearly impossible and very tedious for a regular player (who doesn't dabble in script and code and API and whatnot) to gather and check.

Honestly, I think the stats of how well you did in defense would be kind of cute to have, but also not necessary at all. My boy Ari/ @Moonlynx, the winner of my latest D3, deliberately used a Season-style 75pts guaranteed "defense" team all week through, to give you a quick and dirty example of how much this impacts your own score. ^^

[OmerB]

They can just give the points made against me (as I'm having the points that I have made against others, of course).

 

[Master-17]

4 minutes ago, OmerB said:

They can just give the points made against me (as I'm having the points that I have made against others, of course).

Without the script in vanilla game you have only Win/Lose "points". Points are ONLY in script and only on one specific device.

[Ravi-Sama]

Meh.  Defense teams don't matter imo.  I'd rather be able to filter and sort players by whether they're boosted or not.

[Horsting]

This would indeed give some sort of offence/defence performance ratio. Currently you know how well you perform offensive (of course) and how well others perform in average against all other players. But you do not know whether e.g. you give your opponents over average points, or whether the they score mostly against other players.

Theoretically the strongest (rainbow) offence team should be also the strongest defence team, but maybe there are exceptions. Especially the AME may have an impact so that one team/girl may be stronger offensive, with first attack and AME bonus, while another team/girl may be stronger defensive, without first attack and AME bonus. So some info about your defence performance could theoretically help to optimise. But given that individual fights can have very different results, and that the differences between teams/girls with and without AME+first attack will be small, I don't think one will be able to derive significant results, but probably more likely draw false conclusions.

But it would be actually quite possible to have HH++ do an additional simulation without AME bonus and with the opponent doing the first attack. So you could have an additional defence win chance + expected/average score (for the opponent) just beside the ones shown now. @OmerB would this serve your need or are you aiming for something different with the actual defensive league statistics of a particular week?

Edited June 27, 2023 by Horsting

[bolitho76]

1 hour ago, Horsting said:

how well others perform in average against all other players

Not really. You don't know how much fights they have done at a specific point, this information isn't available (or I hadn't found it until now were you can see it)

1 hour ago, Horsting said:

Theoretically the strongest (rainbow) offence team should be also the strongest defence team

The strongest offence team isn't rainbow. If the opponent have an elemental team in most cases the counterteam is best. A blue or black team can also be the best offence team and outnumber a counterteam if blessings are enable this. Best offense Team against balanced Teams is often a blue team or the team of your 7 strongest girls and this is (when you have a solid amount of 5*/6* girls) often an elemantel team. A black team can also be the best team against balanced. Depends extremly on the situation of you and your opponent.

Rainbow is normaly the best defence team because there is no counter against.

 

More than to have the information how my opponents have performed against me it would be intresting to have the exact amount of done fights from the other in the League.

[Horsting]

3 hours ago, bolitho76 said:

You don't know how much fights they have done at a specific point

After the league has finished, I mean. But yeah, in any case it is a hard or impossible to derive helpful info from this.

3 hours ago, bolitho76 said:

The strongest offence team isn't rainbow.

Of course, what I mean is that you should usually use the rainbow team which appears strongest in HH++ stats (hence offensive) as defence team, i.e. that a team which is stronger offensive, usually is stronger defensive as well. No doubt that a counter team, or blue/black (if opponent team does not counter it) is better for offence, depending on the girls you have. But usually does not mean always, as I tried to point out, especially regarding AME. So an additional HH++ simulation how the selected opponent would attack you could be interesting.

Edited June 27, 2023 by Horsting

[Moonlynx]

23 hours ago, DvDivXXX said:

Honestly, I think the stats of how well you did in defense would be kind of cute to have, but also not necessary at all. My boy Ari/ @Moonlynx, the winner of my latest D3, deliberately used a Season-style 75pts guaranteed "defense" team all week through, to give you a quick and dirty example of how much this impacts your own score. ^^

Haha, I'm a bad example - two weeks ago I've lost by less than 10 points against one of the top players. Could have been a win if I had a strong defense team. ^^

But, yes, that's the magnitude we're talking about when it comes to defensive stats. That is, if you give away 75 points to your opponents. If you just don't care about optimizing and don't deliberately choose an extremely weak team, we're in low single digit territory. I don't know if statistics that revolve around your defense are that important to your performance.

[Horsting]

28 minutes ago, Moonlynx said:

I don't know if statistics that revolve around your defense are that important to your performance.

There are 300 fights against you in one week. If you would want to derive anything from those, you could:

• Have one defence team in one week and another defence team in another week and compare the average score against you. Of course this only works if you do not get stronger in any way between those weeks, ah and blessings break comparing it completely. Also the opponents change between two weeks. Given that we are talking about nuances, I'd say impossible to derive anything from stats between two weeks, and longer time ranges won't work because you progress.
• Have the defence team changing back and forth within one week. Back and forth, e.g. between the 6 hour windows, is important to eliminate the effect of opponent progress, boosters, probably contests, day of week/weekend, and all such which might have an effect when opponents are attacking during the week. Blessings reset also break tests, so probably it works only for phase 1 or only for phase 2. And of course it means that statistics need to be provided by team or by opponent. What breaks this completely is the much too low number of fights, which do not allow to derive any significant results.

So I think it really is impossible to test defence teams based on league stats. Too much is disturbing consistent conditions for a too low number of samples.

So if it is about tuning defence teams, not about pure curiosity, then the only way I see is doing simulations like HH++ does for offence.

Edited June 28, 2023 by Horsting

[DvDivXXX]

You do you, Horsting. But to me the main takeaway from this should be: it doesn't matter anyway. Not: let's dig really, really deep into simulations and stats to find out whether a strong defense team earns you +5, +10 or +20 point margin on your 6k+ weekly score compared to no defense team at all. ^^

[Horsting]

You are probably right. Not sure how many players would really make use of this for the last few points, and in how many cases the defence simulation really leads to a different defence team than what you chose already based on the current offensive simulation.