Jump to content
Android25

Fame Point and Clan Fame Point Math

Recommended Posts

Yeah we had a short chat about it in BUNEH last night. Still not sure if it's a typo on the main site, but it does say the last night of the campaign is the 17th.

 

It seems very unlikely that it would be a typo.

 

If it isn't a typo, it seems like it would be good news to anybody at ~80,000 FP who's sick of the campaign, and bad news to anybody who started the campaign late and is still trying to make it into the top 3k.

Share this post


Link to post
Share on other sites

Its true, its written on SEA that way as well... welp... Adjusting numbers again

 

This will be crossed out when they're updated :)

 

Logic behind the number I will be posting:

 

Still a 1.93 average expected multiplier, but now .93 times 2/7 instead of 1/3, bringing the new multiplier to 1.265 instead of 1.32.

 

The days past multiplier is calculated differently as well, 28/D instead of 30/D.

 

There, all edited again... I expect the final 3000th value at this point to be at about 75000 (+3000/-1000) (About 1 game of error)

 

Edit:

 

 

So, I made this for Foxey, as we get closer to the end of the campaign, but I figure its somewhat useful for those interested and I should share it.

 

The assumption is that you enter the names and FP values of your clan in the lower grey area, and then the top grey box are values you can edit to see the outcome.

Its locked, so you'll have to download or copy it and edit it yourself in Google docs or Excel.

 

https://docs.google.com/spreadsheets/d/13i1l0lq__lo_TcK0yhh-IAvZLPUoWZ2ZecVODewkN_4/edit?usp=sharing

I edited this spreadsheet. I never filled out the rest of the if statement for if there were 10 or less days, and now its set up to account for only 8 more days of the campaign.
 
I also added a Player ID column. If you enter the 10 digit player ID, the page will automatically load the name and FP of the corresponding player and keep it up to date from that point on.

Share this post


Link to post
Share on other sites

Is it just me or does the third stage give so much more fame points than the rest of the campaign? My clan is a bit concerned @ the minimum amount of fame points required to reach the top 2000 spots here in SEA - we're trying our best to get as many people in the top as possible. Is it normal to get 21k famepoints for just two victories (as I did last night)?

PS: If it's normal - them damn. 100k fame points might not guarantee you getting in the top 2000.

Share this post


Link to post
Share on other sites

By Tanks Accounted for, do you mean that if you are in the position above those number of tanks accounted for that you have the tank secured? 

That's the assumption, but its not set in stone ofc.

 

 

I got 12k in just 1 game, pretty sure that estimated 3000th place needs to go up.

 

same.  this G coefficient might be underestimated.

 

The value doesn't need to change, at least not yet. The G value was accounted for, and so far, it was actually over estimated. I've been saying since the rules came out that you should expect a 1.93 coefficient from G. That means just shy of 2 times the Fame Points you were getting. Since you were getting about 5k FP for a win in the first 2 stages, you should expect around 10k FP in the third stage.

 

But I accounted for this in my estimate. This is why in previous days, the current 3000th place value was multiplied by the fraction of days we have left (28/X) and then 1.265. Because (.265)/(8/28)=.9275 = x1.9275 or about x1.93 coefficient.

 

But I'll let the numbers speak for themselves:

 

4v7mLsH.png

 

Day: Obviously the day of the campaign

Value: the 3000th FP value on that day

Change: How much the FP value increased from the previous day

Historic Average: The Average change of that day and all days before it from the First and Second stages

G Historic Average: The Average change of that day and all days before it (currently only 1 day) for the 3rd stage

True Average Difference: The latest G Historic Average divided by the latest Historic Average

Expected Average Difference: What I expected the True Average Difference to be when the rules first came out

True vs. Expected Average Difference: Expected Average Difference minus True Average Difference

Average expected Place: What the expected 3000th place should be if only taking into account the history. (3000th place value * 28/day)

Expected Place: What I still expect the 3000th value to be based on a 1.93 G coefficient

Adjusted Expected Place: Same equation as Expected place but multiplied by True Average Difference instead of Expected Average Difference.

 

Or you could look at it like this:

 

1OzavS6.png

Share this post


Link to post
Share on other sites

I ran the 3000 position data through a basic econometric forecasting model (ARIMA) for fun, nothing fancy. Seems low but I'll still post it.

 

3000 FP position, Lower 95 percent confidence: 54516

3000 FP position, Upper 95 percent confidence: 56630

 

and

 

3000 FP position, Forecast: 55573

Share this post


Link to post
Share on other sites

Values are updated for night 2 of the 3rd stage. Night 2 being the more real scenario since there aren't landing battles everywhere. The average G value is still lower than 1.93 but the gap is closing. If it gets above, I'll start posting based on the true G multiplier rather than the expected G multiplier.

 

 

Would be interesting to see a graphic of how the projected value has changed day by day

 

 

The 3000th place chart is a bit of a mess, not too much you can take away from it except that it's risen from an all time low over the past few days.

inWQwjC.png

 

The 5000th place one is a bit more interesting, it shows lack of interest in the campaign for people below the 5000th place over time.

 

xqZ5tAF.png

Share this post


Link to post
Share on other sites

I ran the 3000 position data through a basic econometric forecasting model (ARIMA) for fun, nothing fancy. Seems low relative to Android's but I'll still post it.

 

3000 FP position, Lower 95 percent confidence: 54516

3000 FP position, Upper 95 percent confidence: 56630

 

and

 

3000 FP position, Forecast: 55573

 

Missed a few datapoints, I misunderstood the provided spreadsheet (and thanks for keeping track of it).

 

updated with missed points and last night actual value...

 

  • FORECAST: 59286
  • L95: 56102
  • U95: 62471
  • STD: 1625

(rounded to whole famepoint)

 

Do we get a cookie for whoever is closer?

Share this post


Link to post
Share on other sites

There is no way in hell it will be that low.

 

ARIMA models are quite powerful, don't underestimate them. I'm not a time series economist so I'd guess they'd do some other fancy things but I'm throwing it out there.

Share this post


Link to post
Share on other sites

ARIMA models are quite powerful, don't underestimate them. I'm not a time series economist so I'd guess they'd do some other fancy things but I'm throwing it out there.

I just don't believe that its powerful enough to take into account that the averages for the last 8 days are guaranteed to be at least 1.5 times larger than the averages for the first 20 days. Even if you take the lowest of the historic average numbers and multiply it by 20, then add it multiplied by 1.5, 8 times, you still get 61376. That's like, the absolute lowest possible estimate based on the data and I'm telling you its going to be too low. If you do the same thing to the highest average and multiply by 2 (a more reasonable but possibly a bit too high G value) you get 80928, which at this point I would start regarding as the highest possible value.

 

If you take the true average of the first 20 days and do the same math a third time, you come up with 74034... My prediction still stands from the 8th

 

 

I expect the final 3000th value at this point to be at about 75000 (+3000/-1000) (About 1 game of error)

 

 

Edit: One other thing that has to be remembered is that it really doesn't matter at all how many fame points the top 2999 people are getting. As long as the 3000th place and above continue to behave as they have been, nothing should change. The chances that people outside of the 2999th position are going to be consistently getting fights where their clan owns 3 or more provinces and the other clan owns 3 or more provinces, or that they'll be reliably winning fights with no provinces against clans with 3 or more (the first 2 situations that that bring the G multiplier above my expected average of 1.93) is very slim.

Share this post


Link to post
Share on other sites

ARIMA models are quite powerful, don't underestimate them. I'm not a time series economist so I'd guess they'd do some other fancy things but I'm throwing it out there.

Are you sure you have your inputs correct? The last two days should be weighed much heavier than 1-20 as they will be more representative of the g inflation.

You also aren't accounting for the inflation that will come with more activity from those at the edge. Many of the people over 100k fame will most likely be sitting out battles in favor of those who need points. im not sure any software will account for that without heavily tweaking the models.

Share this post


Link to post
Share on other sites

My guess, based on the fact that everyone thinks it will be around 75k, is around 85k. I expect a last minute flurry of battles for the last couple of days because of good clans trying to get folks the tank.

Share this post


Link to post
Share on other sites

My guess, based on the fact that everyone thinks it will be around 75k, is around 85k. I expect a last minute flurry of battles for the last couple of days because of good clans trying to get folks the tank.

I've considered this, in an impossible scenario a single player could potentially win in excess of 52500 FP in a single day, assuming all chips are available, all fights are won, you manage to win on a 10 stack, and somehow none of the 7 fights are at the same time. But I just think there are too many mouths to feed and not enough chips to push out for it to change that much.

Share this post


Link to post
Share on other sites

I've considered this, in an impossible scenario a single player could potentially win in excess of 52500 FP in a single day, assuming all chips are available, all fights are won, you manage to win on a 10 stack, and somehow none of the 7 fights are at the same time. But I just think there are too many mouths to feed and not enough chips to push out for it to change that much.

Some of our players earned 35000 points in one night on Stage 1 and 2.

 

it is entirely possible for 52500 Fp to be gotten in one night given enough effort.

Share this post


Link to post
Share on other sites

Some of our players earned 35000 points in one night on Stage 1 and 2.

 

it is entirely possible for 52500 Fp to be gotten in one night given enough effort.

 

 

 

I've considered this, in an impossible scenario a single player could potentially win in excess of 52500 FP in a single day, assuming all chips are available, all fights are won, you manage to win on a 10 stack, and somehow none of the 7 fights are at the same time. But I just think there are too many mouths to feed and not enough chips to push out for it to change that much.

 

 

Correct if I'm wrong, but assuming -G- sits on 8 provinces (entirely probable), pushing out 3 or more stacks in different time zones and winning all those fights, the person who joined all those 3 fights would get quite close to 52.5K.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...