Fix position card avg entry preview calculation so it matches position chart

[0xwalde]

The modifiedAverage functions in PositionCard.tsx and PositionChart.tsx are supposed to be calculating the same thing, but they get different results. It looks like the calculation in PositionCard.tsx is incorrect because it is adding the size property of the potential trade to get the new total size when it should be using sizeDelta instead.

In addition, the data being fed into the PositionCard version of the calculation can be positive or negative depending on the side of the trade, which isn't being handled quite right either.

Related issue

https://github.com/Kwenta/kwenta-private/issues/261

How Has This Been Tested?

Tested on Goerli Optimism in the 4 following scenarios:

• Adding to an existing long position
• Subtracting from an existing long position
• Adding to an existing short position
• Subtracting from an existing short position

Video:

Kwenta-avg-entry-short-fixed.mov

Kwenta-avg-entry-long-fixed.mov

[vercel]

@0xwalde is attempting to deploy a commit to the Kwenta Team on Vercel.

A member of the Team first needs to authorize it.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated
kwenta	✅ Ready (Inspect)	Visit Preview	Dec 1, 2022 at 5:28PM (UTC)

[Tburm]

I still think this presents an incorrect average price. I tested reducing a long position, and it decreases my average price. This is incorrect, if I buy 1 unit at $1000 then sell .5 units my average price is still $1000/unit.

The expected behavior to match the subgraph calculation would be:

• If increasing position size calculate a new average
• If decreasing position size the average stays the same
• If switching sides from long -> short or short -> long, the new average price is the current oracle price

If you want to take another shot at fixing this let me know, otherwise I can pick up where you left off. Thanks for pointing this out!

[0xwalde]

Ah yeah, thanks for the feedback Troy! I was mainly concerned with making the card preview match the chart preview, but you're right that they're both wrong in several of those scenarios.

I pushed up another commit with some new calculation logic that should handle all of the cases you pointed out-- let me know if that seems to behave better. Thanks again!