Ivy Presents: War Data, Part 1

Introduction

I have tracked a rather large amount of war data in my alliance. From very early on it was one of the parts of the game that I enjoyed the most. I’ve been thinking about a number of ways I may analyze it, but wanted to start by sharing a small sliver. The below image shows my points scored from 39 wars, from early November 2018 to present (I missed a few, but most of them are there). I figured this would be a fun (and narcissistic) thing to present first, to start the conversation / gauge interest in this topic.

Methodology

I took my war scores from (almost) every war in this period, then essentially normalized them. The displayed y-axis value is what I’m going to call the number of average teams I killed. An average team is calculated as the points I scored / 1,500 / number of opponents. N.B. this means it is possible to score more than six average teams killed, even though you can only kill, at best, six actual teams. The x-axis shows the number of days I had been playing the game at the time the score was logged.

Notes

  • I used all six flags in every war.
  • I think this overstates my early contributions, as I was mostly doing clean up jobs and therefore would rack up kill bonuses which were disproportionate to my actual contribution.
  • I do not necessarily think this is representative of what a new player could / should do. Like I said, I was heavily invested in war from the beginning, so building war teams was my priority over the map, raids, and events. That’s just how I opted to play.
  • At the same time, I think there’s a chance this actually understates my more recent contributions. Now I only hit complete teams, meaning I sometimes heavily damage a team but do not score the kill bonus. I may try to refine this at some point by awarding partial kill bonus points.

Future Improvements

  • As states above, I may try to kind of pro-rate out the kill bonuses.
  • Right now only points scored in attack are accounted for. I may attempt to calculate something more along the lines of total value contributed, accounting for things like how many points my defense gave up, and how many flags it absorbed from the other side.
  • Taking that a step further, I may even try to account for flags absorbed from different strength teams - that is, it is more valuable to burn up a flag of a player with a near 4,000 team power than a team with a sub 3,000 team power).

6 Likes

The way I approached the game, worked on war teams in stages going for 6 kills.

Team points versus kill points

Separating team points and kill points would probably ( 75% guesstimate) show both effectiveness against whole teams AND clean up since both can be wildly successful or dismal failures. Then two graphs, one for team points and one for kill points.

Example
If you score 30 points for a kill. kill bonus is 25 points. Then 5 team points and 25 kill points.

Defense measuring

Some of our scarier looking teams seem to scare off enemy attacks more than teams that are less scary looking but better at defending.

we track, but do not record, how many attacks a defense team absorbed and how many hours before it is first attacked.

Since our alliance is very lopsided, some of our founding players can go an entire war without being attacked.

Points given up

one way to way to measure defense =

( 1 + ( ( team points / points given up) / team points ) ) * attacker level

Example

( 1 + ( ( 40 team points- 25 given up ) / 40 team points) ) * level 35 account attacking = 48.125

or

( 1 + ( ( team points / points given up) / team points ) ) * attacker’s defense team points.

4 Likes

I’ve been collecting data on our war defense. Like yours we’re pretty lopsided. I keep team power, war points, tank- color, level, how many times hit, killed, 1 shots, 1st half, 2ed half. I’m going to work with your formula and keep collecting data. Thank you.

2 Likes

@Gryphonknight, I am…0% surprised to see you show up in this thread and to learn that you also track something! :smiley: I’m going to think more about your idea to measure defenses. One problem I’m running into is that at some point you’re kind of forced to use some quantification of the attacker. This can be level, team power of their defense, or whatever. But we all know that ultimately that is a proxy and therefore possibly misleading.

For example, a player starting out can pour resources into their primary team and relatively quickly field a defense around…let’s say…3,300. But in that case they would have nothing for their attack teams 2 through 6. I remember when my own defense team was finally at that level; my weakest attack team was a squad of maxed three stars. Again, that’s because I really enjoyed war and prioritized it. Most of the player base is probably between these extremes.

I guess so far I’ve been inclined to use defense team power merely because I’m already tracking that (I do not have player level tracked).

@Rlstorm, please tag me if you ever share any of that; I’ll be interested to see what your data show and how you’re analyzing it!

1 Like

Account level

Especially with Field Aid (WR) team power can change dramatically between wars. Attacker level is a much more stable measure.

It can often be deceptive, especially for alt accounts, but the higher the number, the more player XP the account has accumulated.

Data points

You could track BOTH account level and team power. I often collect data I do not intend to use, that seems useful for future analysis.

If team power seems to be misrepresenting the situation, you can then process the account level data and see how the two analysis diverge. If they do not diverge ( much ) it is just RNG.

Wars tend to be prone to a lot of RNG because limited opponent pool ( no reroll against enemy alliance match and revive timers ), limited attacks and banning battle items.

Sorry real busy there for awhile.
I see that a lot, low level with high team power, you know there’s not much on the shelf. I’ll start putting level and TP in, may even look at some their other attacks.
Start putting some of this together maybe get something useful.
Thanks for the insight, I’ll try to keep up with ya.

1 Like

Found this thread when looking for guidance on how to track my war defense team performance.

Background

My alliance recently switched from blue to purple tanks. Available options are CTiburtus +18 or Clarissa +3. Unfortunately, no costume for my Rigard +18, no Ursena or Kunchen.

I have a deep 4* roster of 30+ heroes with 20 highly emblemed, but only 9 maxed 5* that I only recently started embleming. Unfortunately, many of my 5* are in the same classes, and my red and blue 4* and 5* heroes are not ideal for our coordinated war defense hence I’m looking at optimising emblems.

I’m prioritising my war defense for emblems, so I want to make the best decisions based on my war defense performance.

My current war defense

Please note that my current war defense includes many heroes I use on offence as well, so embleming them will help me in other aspects of the game, and that I don’t seek for input on my war defense as such.

Now back to the topic.

Based on this thread and some others, I’m planning to collect the following data:

Miscellaneous

  • War type (field aid, arrow barrage, attack boost)

Alliance related data

  • My alliance’s war score
  • My alliance’s remaining war flags
  • Opponent alliance’s war score
  • Opponent alliance’s remaining war flags

War defense data

  • My own team composition
  • Defense team power
  • Troop types
  • Troop levels
  • Points available based on defense team power
  • Bonus points for victory

For each attacker

  • Player level
  • War defense team power
  • War defense troop types
  • War defense troop levels
  • Points scored against me

I would highly appreciate input in relation to

  1. What other data might be relevant
  2. What data listed above might be irrelevant
  3. How you approach data collection
  4. How you approach data analysis

Thank you for any input in advance.

@IvyTheTerrible @Gryphonknight @Rlstorm

1 Like

Hey @SING, this is a big question!

Let me start by saying that I did track war data for a while, but stopped a while back. There are a few reasons for this:

  • I decided to track raid data instead, and I was only willing to spend so much time on the game between playing the game, being on the forums, and data collection.
  • Since you can never know the strength of the attackers teams there will always be a substantial gap in the data. Say your defense gets one shot. Does that tell us anything? What if that was a top attack team from a high-end roster? Does knowing the attacking player’s level tell us anything? Not really - they have six flags so we have no idea who they used from their roster. What if it was the fourth flag they used? Again, we have no idea if they use their best heroes first or maybe saved some strong ones for you. I saw no way around these sorts of problems.
  • Let’s say the hardest teams I’m attacking are worth 60 points, with team powers around 4,500. That probably means there will be teams with powers around 4,000 - 4,200 worth 56-57 points. I can probably score six one shots against the latter types of teams often, thus maxing with a score around 340. That seems good, on paper. But the reality is, I think I do more to help my alliance (and have more fun challenging myself) by going up against the best teams each time, win or lose. I mean, someone has to attack them. :slight_smile: We all know total points scored imperfectly reflect a player’s contribution. I became increasingly worried that rigorously tracking my points (or successful one shots) would cause me (perhaps even subconsciously) to select easier teams to boost that specific metric.

Okay, long disclaimer aside, here’s what I would do. Basically, just take a ton of screen shots and then enter all the data into an Excel file I set up.

  • Final points score of each alliance
  • Each alliance’s membership, to get names, player levels, and alliance score
  • The war score log (I used this information to calculate flags used / unused)
  • Each team’s defense power

It took a while to collate all the data. Like I said, ultimately I stopped because of the amount of time, gap in the data, and because I decided I was more interested in other matters.

Lastly, I want to offer a few thoughts about defense. It’s important to realize that war and tower defenses are different. A few things I would think about with war defense:

  • The goal should be to reduce the likelihood of a one shot. Usually the best way to do this is, if you have a reviver, get them on the field.
  • Your alliance is attempting to stress the attacker’s supply of a certain category of heroes. You are already doing this to their holy heroes by fronting all dark tanks. Other things to think about include cleansers and dispellers. For example, if your alliance can all put forth a Kunchen, Boss Wolf, or Clarissa, great - that’s a huge challenge to cleanse across all the flags. But some will have to put out Seshat or maybe Cyprian or whatever - those do less to push the attacker’s stores. Similarly I would maybe rather face a blue tanked alliance with a mix of Aegir and Richard than all Aegir’s (even though I think he’s probably a lesser tank), because I doubt my alliance has enough Melendors and Caedmons to counter (whereas against Richard the Kirils and Boldtusks of the team come into play).
2 Likes

Thanks for taking the time to reply in so much detail. I understand that collecting this much data might involve a lot of time, and that is the most daunting aspect for me. And who knows if it even pays off!

This is exactly what I’m trying to achieve. I don’t have any revivers, so I’m trying to build a war defense team that absorbs as many flags as possible. With a limited pool of 5*, I’m in the bottom 10 of war defenses in my alliance in terms of team power. I get one-shot by stronger opponents at least twice every war - I don’t have data to back it up, but I think almost three times on average.

The more I reflect on your response, the more I wonder whether collecting data of my war defense team performance is the best way to achieve my goal.

Really appreciate your input!

1 Like

I’m happy to help!

You have a strong defense; if you’re getting farmed then you’re probably in a strong alliance. In which case I wouldn’t worry about it too much - improvements will come.

Looking at your defense, here are my comments:

  • The biggest and easiest opportunity for improvement is probably your tank. I believe the tank’s sturdiness has an outsize impact on the team’s viability. Once it falls the entire team is much more susceptible. So look at Clarissa as a great home for more emblems and strip another Paladin for her if you can.
  • Your flanks could be improved to fast snipers. Overall, with both being healers your front line (and therefore entire defense) is fairly passive. More importantly, I don’t think they synergize well with Clarissa (I think - it’s still early in her release, so I’m not 100% confident I understand how to best use her). I think the idea should be that Clarissa, given her speed, puts a DoT on your opponents. That then brings them into sniper range for a kill.
  • I actually like Musashi and think he’s a stronger defender than many give him credit for. But I think I’d flip him and Lianna. That maxes the chances that he hits three (since him going first maximized the likelihood that all attackers are alive) AND gives Lianna a better shot at a kill. I’m thinking in particular of a scenarios where Musashi’s minor damage hits someone and then Lianna finishes them.
1 Like

Thanks for your feedback! I agree with your points above. Given Clarissa is my only viable 5* option for tank, she’ll be getting my paladin emblems. I don’t have enough to spare at the moment, but she’ll be first choice in the future until she hits +6 or +7. It’s going to be a long grind, unless next PoV will have Paladin emblems.

As for your other suggestions, I’m locked in by our coordinated alliance war set-up: green-red-purple-blue-yellow, and so Lianna and Musashi will have to stay where they are, at least for war. Raid defense is a different matter entirely, as you pointed out above.

Your reasoning in regards to Clarrisa’s role makes sense to me. Unfortunately, I don’t have any viable red or blue 5* or highly emblemed 4* snipers. However, I have been thinking that a centre of BT, Clarissa, Isarnia might be more effective than my current centre. I will probably give that a shot.

I think there are two interrelated issues at play here:

  1. A thin 5* roster
  2. Our current war set-up is not ideal for my roster

There’s not much I can do about either but grow my roster and work with my alliance to find overall better set-ups.

I might still collect data, but maybe scale it down.

1 Like

Ah, I didn’t know that, of course. Not sure I see the logic for it, at an entire-alliance level, but that’s not my call! I am intrigued by your Boldtusk and Isarnia idea, especially if they bear emblems. But regardless, dont’ sweat it. Seems like you have a good and developing roster, a sound understanding of the game, and an organized alliance. You’re on the right track and everything else will develop just fine.

1 Like

Probably should have mentioned that in the background section above - sorry about that. BT is +18, unfortunately no costume, and Isarnia is next in line for wizard emblems and will be at +3 for war, so same as my other 5* in this defense.

Thanks again for your help!

1 Like

Cookie Settings