Sticker stats

All the numbers below are subject to change depending on the game balance and player feedback.

Our target is to give the following power increase per sticker:

• 4% as a base power increase to the Pookyball for a Level 0 Common Sticker

• Increased by a 1.25% multiplier to the base number per level

• Max level of a given rarity has the same boost of Level 0 of the next rarity

Rarity & level	Average power increase
Common, lvl0	+4%
Common, lvl40	+6%
Rare, lvl0	+6%
Rare, lvl60	+9%
Epic, lvl0	+9%
Epic, lvl80	+13%
Legendary, lvl0	+13%
Legendary, lvl100	+18%

• To achieve this, we have looked at the probabilities of given events happening (for random stickers) or have looked at how often you can deterministically benefit from the sticker

  • Example 1 : a Club sticker benefits in 1 match out of a matchday of 10 matches. As such, it has a deterministic rate of 10%.

  • Example 2 : a sticker giving extra points if the score ends up being 1-1 has a 11% probability of activating

• However, variance is quite powerful in Pooky. As such, we’ve applied a variance correction formula to the points given. We calculate it by putting the event probability rate at the power (-0.5).

  • $(VCF) VarianceCorrectionFactor = 0.5$

  • $point.multiplier = probability ^{-VCF}$

  • Example : a sticker with a 10% probability will have a variance-corrected factor of (10% ^ (-1/2) = 3,16. For a level 0 common, it means it would grant 4% * 3,16 = 12,65% extra points when it activates.

• The point multiplier has a maximum, currently set at 6. This means any event with less than 2.78% chance of happening will offer the same point boost, regardless of the final probability.

All of the below can be adjusted for balance purposes and depending on how the football games change with time:

• the VCF

• the point multiplier maximum

• the underlying probabilities

Prediction point computation

Prediction point =

• sum of: Base Points multiplied by their respective attributes (from both Pookyball and Attribute Stickers);

• multiplied by: (1 + Sum of Pookyball Boosts)

• multiplied by: (1 + Sum of Sticker Boosts)

• multiplied by: (Pressure Factor)

How are the stickers balanced when they don't award exactly linear, probability-based points?

Pooky is a game where 5% extra points can mean a lot of places in the leaderboards.

Our reward distribution is exponential, meaning that ranking 1st + Last is much better than ranking in the middle twice.

In order to take this into account, we’re giving less points to stickers that have a high variance - because when this high variance happens, you get exponentially more rewards through that matchday. As Pooky looks to reward skill, we have implemented a formula that puts the probability of a given event at a negative power, between 0 and 1.

For instance, with a 0.5 VCF, an event that has 10% of happening, will give 3.1x more points than an event that has 100% of occurring - because a lucky amount of occurrences can boost you significantly. Without this factor, just 2 of the 10% occurrences would amount to 2x the boost of the “guaranteed” sticker… and getting 2/10 correct at 10% probability is 19.3% proba, 3/10 is still 5.7%!

We will heavily monitor the factors, notably the VCF, to ensure that all stickers have a close power level when used optimally..

Detailed sticker stats

For simplicity, this spreadsheet contains the latest data about stickers. It will be updated on a regular basis up until the release of Stickers.

https://docs.google.com/spreadsheets/d/1rPkh9vei0ZPkr8CPC72I0GUw-gSaJV2QUAu0QuptinE/edit#gid=1641744608