Battleground Blitz Week Two: Points Are Accelerating
DNA3000
Member, Guardian Guardian › Posts: 19,676 Guardian
Tl;DR: We are on target to hit one billion points without Stark Foundation Donations
So where are we now?
If we project linearly:
About 1.2 billion. But of course, that's cheating a bit, because that includes the Stark Foundation donation of 100,000,000 points, obvious in the chart. This makes it seem like our point earning is accelerating upward faster than it actually is. What if we remove that donation, and just look at the player activity alone?
A bit over 900 million points. But wait, weren't we projected to hit 800 million last week? Yes, we were. Here's last week's projection superimposed on this week's projection:
Our current projection is higher, because our average points per day is increasing significantly. We can see that in our points per day for the first two weeks:
This also removes the points donation. Our average points per day is rising. It bounces up and down, and most of those bounces are aligned with BG milestone reset days. On those days activity is higher, and on alternating days it tends to be lower. But the overall trend is upward. We can also see this, to a similar degree, in the estimated hourly activity:
We can see in particular we used to bounce above and below one million points per hour, and now we're almost always above that rate. The big spike in the middle is not the Stark donation, by the way. It is due to the fact that for a few hours the realm scoring mechanism stalled. That's why it dropped unusually low around hour 192 and then bounced up very high as it caught up. If you want to see what the Stark Foundation donation looks like:
Yeah, that's impossible to mistake for anything else.
Our average point earning rates have increased overall by about 30% over the two weeks of the event. We cannot necessarily assume they will continue to rise at that rate, but there is every reason to believe they will continue to rise as participation generally increases in the second half of the BG season. There are also still Elder Marks in locked milestones that will continue to fuel a lot of activity in the short term, and continued increases in GC population which will increase average point earning per match by some amount. Conservatively, though, if we continue at even reduced rates of point acceleration, we should easily hit one billion points scored by the players by the end of the event. The Stark Foundation donations will just be gravy.
It is probably worth noting here why Kabam even decided to add those donations. They have a mountain of historical data I don't. They almost certainly could predict scoring better than I could, and almost certainly saw what these numbers would most likely look like. They appear to have selected one billion as a target we could and almost certainly would hit, and the additional milestones above as stretch goals for the circumstance where we did significantly more than predicted. So why add tons of points when they were not necessary?
The most logical reason is: sentiment. Left to our own devices, Kabam could predict our participation numbers would push us to one billion if we did what we normally do during boosted seasons. But what if negative sentiment started to take hold, as it did with Crystal Cleanse? Then players could enter a downward spiral where they decide to hold back for fear of wasting effort in a losing cause, causing the numbers to fall short, causing more people to decide the whole effort was doomed to fail. The points injections eliminate that possibility, or at least neutralize it. If players believe that getting to at least one billion is almost inevitable, then they are less likely to fall victim to negative sentiment.
This has some pros and cons. In the short term, it mitigates negative downward spirals. But it can also create a sense of entitlement, that players will think they deserve all the realm rewards no matter what they do, and thus there's no need to actually do anything. That's something to watch out for in future realm events.
Even I originally estimated our point earning potential incorrectly, for a number of reasons: underestimating historical participation rates for one, and not accounting for the large influx of elder marks in the realm milestones for another. But careful analysis of the data shows that the trend is our friend in Battlegrounds Blitz. If we fail to reach one billion points on our own it will be because the game experiences a week of downtime or something. Otherwise, we're on track to get there.
So where are we now?
If we project linearly:
About 1.2 billion. But of course, that's cheating a bit, because that includes the Stark Foundation donation of 100,000,000 points, obvious in the chart. This makes it seem like our point earning is accelerating upward faster than it actually is. What if we remove that donation, and just look at the player activity alone?
A bit over 900 million points. But wait, weren't we projected to hit 800 million last week? Yes, we were. Here's last week's projection superimposed on this week's projection:
Our current projection is higher, because our average points per day is increasing significantly. We can see that in our points per day for the first two weeks:
This also removes the points donation. Our average points per day is rising. It bounces up and down, and most of those bounces are aligned with BG milestone reset days. On those days activity is higher, and on alternating days it tends to be lower. But the overall trend is upward. We can also see this, to a similar degree, in the estimated hourly activity:
We can see in particular we used to bounce above and below one million points per hour, and now we're almost always above that rate. The big spike in the middle is not the Stark donation, by the way. It is due to the fact that for a few hours the realm scoring mechanism stalled. That's why it dropped unusually low around hour 192 and then bounced up very high as it caught up. If you want to see what the Stark Foundation donation looks like:
Yeah, that's impossible to mistake for anything else.
Our average point earning rates have increased overall by about 30% over the two weeks of the event. We cannot necessarily assume they will continue to rise at that rate, but there is every reason to believe they will continue to rise as participation generally increases in the second half of the BG season. There are also still Elder Marks in locked milestones that will continue to fuel a lot of activity in the short term, and continued increases in GC population which will increase average point earning per match by some amount. Conservatively, though, if we continue at even reduced rates of point acceleration, we should easily hit one billion points scored by the players by the end of the event. The Stark Foundation donations will just be gravy.
It is probably worth noting here why Kabam even decided to add those donations. They have a mountain of historical data I don't. They almost certainly could predict scoring better than I could, and almost certainly saw what these numbers would most likely look like. They appear to have selected one billion as a target we could and almost certainly would hit, and the additional milestones above as stretch goals for the circumstance where we did significantly more than predicted. So why add tons of points when they were not necessary?
The most logical reason is: sentiment. Left to our own devices, Kabam could predict our participation numbers would push us to one billion if we did what we normally do during boosted seasons. But what if negative sentiment started to take hold, as it did with Crystal Cleanse? Then players could enter a downward spiral where they decide to hold back for fear of wasting effort in a losing cause, causing the numbers to fall short, causing more people to decide the whole effort was doomed to fail. The points injections eliminate that possibility, or at least neutralize it. If players believe that getting to at least one billion is almost inevitable, then they are less likely to fall victim to negative sentiment.
This has some pros and cons. In the short term, it mitigates negative downward spirals. But it can also create a sense of entitlement, that players will think they deserve all the realm rewards no matter what they do, and thus there's no need to actually do anything. That's something to watch out for in future realm events.
Even I originally estimated our point earning potential incorrectly, for a number of reasons: underestimating historical participation rates for one, and not accounting for the large influx of elder marks in the realm milestones for another. But careful analysis of the data shows that the trend is our friend in Battlegrounds Blitz. If we fail to reach one billion points on our own it will be because the game experiences a week of downtime or something. Otherwise, we're on track to get there.
31
Comments
or some visualisation software
I think former makes more sense
On that point, you still need to do 5000 points, which that should be the level of participation that Kabam is looking for an average player.
I don’t think they intended the average player to max out all the milestones: that would be asking for a lot. Also, if the design intent was for the average player to score 5k points, that would imply an expected participation level of 200k players, which is much lower than the number we know participate in even an average season where nothing is going on.
One source of points I originally didn’t account for before the event started, because it didn’t exist before the event started, was the 50 point bonus for completing the “win one” objective that resets every two days. If a player tries to do at least that (which makes sense as it also has the bonus banquet stuff) that’s 700 points above the normal points for doing matches. That makes 5k much easier to reach, and also makes the 3500 target very easy to reach for a player just trying to burn off all the milestone marks.
If 300k players ultimately participate and half of them do those objectives as they reset, that’s a hundred million points just from those 50 point bonuses.
If you're only matching against very strong players, it is because you've advanced faster than all the rest of the level 40s. That means you're better than all of them and should be matching against stronger players. But more likely, as has been demonstrated time and time again by players willing to post their match history, you aren't getting matched against only super strong players. You're getting matched against a variety of players that are commensurate with your current VT tier.
Stronger players move up. That leaves weaker players behind. That means every VT tier gets weaker as stronger players leave. Early in the season it is possible to *sometimes* run into very strong players because they just haven't played enough to promote yet. But as the season advances, those players leave, and lower progress players have a better opportunity to advance behind them.
Over 250k players play battlegrounds. Do you think this game has 250 thousands Valiant players? Most of those players are Uncollected, Cavalier, and Thronebreaker players. And they play season after season because they are realistic about their expectations, and because most of the BG rewards are not progression sensitive, so while it might be harder for a UC to progress in BG, the rewards in BG are nose-bleed high for them when they do.
If you are a fairly new player, your biggest stumbling block isn't going to be roster strength, although that will ultimately put a ceiling on how high you can expect to go before you advance further in the game. Rather, it is your lack of knowledge and experience with how the game works and how the champions match up that will cause you to lose not just to Valiant players with tons of 7* champs, but also Uncollected players with low 6* champs in their deck that know how to play them well.
It just takes time and experience. Experience you won't get if you don't try. This game is super friendly to new players. There's no penalty for losing. If you win, you earn points. If you lose, you earn less points. What's the downside to trying?
BUT
I have a LOT more Eldermarks to claim and that will push me well beyond 5k.
There are a lotta marks ngl.
PUSH Y'ALL
Main point is the linear projection of the actual data is not a projection of the data, but the projection of what looks to be the curve of best fit for the data.
same issue with the projections from the plots from week 1 and 2.
and the average earing rates just does not feel right as a measure for the event
However, more directly, there's no such thing as a "projection" (read: extrapolation) of the data that is not based on a best fit curve of some kind. When the data points do not lie on a directly extensible curve, there's no way to directly extrapolate the data. It is generally understood in basic analysis what this means ordinarily, since there is no other relevant alternative. No idea what this means.
I’m waiting over 5K individual points *hoping* the Realm (and the Stark Foundation) comes through for me (and them too).
Dr. Zola
When I say "linear projection of the data set" I am colloquially referring to the linear extrapolation of the linear best fit graph of the data. Technically, "projection" has a completely different definition mathematically, but its a word most people understand.
However, more directly, there's no such thing as a "projection" (read: extrapolation) of the data that is not based on a best fit curve of some kind. When the data points do not lie on a directly extensible curve, there's no way to directly extrapolate the data. It is generally understood in basic analysis what this means ordinarily, since there is no other relevant alternative.
even with the explanation, the graph doesn't hold up to what you want to find and show, and methodologies that can be easily be critiqued will always produce questionable results.
now linear best fit/curve fitting with a linear curve is a well known method used especially in linear regression, but to use that new curve for extrapolation can only be confidently done if :
1. the curve is able to accurately (usually 90%+ accuracy that is validated) represent the given data,
2. the data is of a linear nature
3. it follows a linear structure everywhere.
This statement in particular is wrong and you can project and extrapolate from a given point or a set of points, based on what curve you want. The best method is to have a function that describes the plot, but even without that there are many ways you can extrapolate assuming particular conditions.
definitions of extrapolation and linear extrapolation from wikipedia.
My question with this set of plots are,
1. are you taking the points gained every hour and then dividing by the total number of players in the event?
2. or is it the average taking the amount of players that gain those point during that hour period? (what should be the actual value).
What should I be doing to get to 5k points, do I tank vib2 for 2 weeks or get into GC and suck hard as long as I win one match every objective reset?
In return for your helpful advice I will burn 9k EM on my bronze 3 alt to help the over all points.