[Cards suggestion] Change MV formula to exclude outliers

[Bawkie]

Personally I don't think you need to discount trades other than self-trades from the market value calculations. I believe MV needs changes though.

How about extending the averages set to 25 trades and where the number of valid trades is below 25, junk value is used in its place.

Example 1:
S2 CrunchyBread has trades 5,000.00, 5,000.00, 5,000.00, 0.02, 1.00 (5/25 trades) so 20 are junk (0.01). The average is 625.05. This doesn't eliminate inflation, but decreases its speed.

Example 2:
S1 Farrakhan has trades 514.00, 1,000.00, 507.00, 2,499.66, 2,499.97, 500.41, 500.01, 1,250.00, 1,253.67, 1,200.00, 500.00, 500.00, 500.00, 500.00, 1,257.34, 775.00, 775.00, 775.00, 775.00, 775.00, 775.00, 775.00, 775.00, 775.00 (25/25 trades) so 0 are junk. The average is 914.88. This is not too far off its current inflated value which consists of regular trades as opposed to two or three large trades.

Example 3:
S2 Kylarnatia has trades 15.00, 22.00, 22.50, 21.25, 9.00, 9.00, 9.07, 10.00, 12.00, 10.00, 9.55, 13.98, 9.18, 9.25, 13.99, 12.00, 14.00, 14.00, 14.00, 14.00, 13.00, 13.50, 13.95, 12.02. The average is 13.18. This is close to its current value and about right for a non-CTE legendary right now.

Hopefully this would have the following effects:
◼ Values of cards cannot change as drastically from large trades due to an increase in the averages taken, and the use of junk values where few trades have taken place.
◼ Trading a card more times up to 25+ creates an average over time, which should become more accurate. Think of junk value as a placeholder value ready to replaced with a more representative value.
◼ Inflation is still possible for those wish to partake, but there should be less cards at 1000+ value due to inflation. It would take far more higher valued trades to bring a card to 1000+ value.

Trades information accurate as of 28 Apr 2021 18:11 UTC

[Almonaster Nuevo]

At the end of the day market value is what someone is prepared to pay.

Part of the problem is the never-ending suppy of free cards, so the available bank keeps rising (and mostly goes to those who invest their time in collecting it). None of the proposed solutions addresses that, and I think it would be quite hard to do at this stage.

On the inflation side, I see more of a problem with the 10 transaction window. That means that following big trade, you get a lot of sharks (like me) grabbing a couple of cheap copies and aiming to sell them on at an inflateed price before the big trade drops out of the window.

One way to remove the window effect is for the MV to be calculated directly from the previous MV (set to JV initially) and the latest trade, keeping the current exclusions. The exact proportion is debatable, but assuming 10% for consistency with the current setup, the newMV would be ((9*old MV)+trade)/10. If you like Bawkie's approach of reducing the effect of each sale, then the proportion could be set to 4% instead.

The effect would be that the impact of each trade fades slowly with time, but withut the sudden jump when an outlier goes out of scope. I think there is also a case for also including a (much smaller) component of JV in the calculation, as an anti-inflation measure. Again, the exact proportions are a separate debate from the underlying mechanism.

You would need to record the MV behind the scenes with more precision, and only round it for display, otherwise commons would tend to stay glued to 0.01 MV

The calculation is simpler, and gives a meaningful MV right from the the get-go. It also means that you do not need to check the history before deciding whether to offer. Inflation still works, but people are less likely to get caught out by sudden value drops.

[HumanSanity]

Personally I don't think you need to discount trades other than self-trades from the market value calculations. I believe MV needs changes though.

How about extending the averages set to 25 trades and where the number of valid trades is below 25, junk value is used in its place.

Example 1:
S2 CrunchyBread has trades 5,000.00, 5,000.00, 5,000.00, 0.02, 1.00 (5/25 trades) so 20 are junk (0.01). The average is 625.05. This doesn't eliminate inflation, but decreases its speed.

I feel like this is still more ease of inflation than I'd like to see, although it is something.

Is there a reason something like this couldn't be combined with 9003's proposal? A large part of the problem is allowing the outlier numbers to be factored in at all because they inevitably shew things upwards, especially where we take averages rather than medians. As an aside, Luc/Fauzhija's idea about taking the median definitely has potential problems but is not a terrible angle to consider in discussions.

[Almonaster Nuevo]

I'm not convinced that it is consistently possible to differentiate an outlier from a "genuine" price, especially when there have been relatively few trades.

[Refuge Isle]

At the end of the day market value is what someone is prepared to pay.

If only this were the case.

[Almonaster Nuevo]

At the end of the day market value is what someone is prepared to pay.

If only this were the case.

I ment actual market value as opposed to MV or asking price.

[Comfed]

If only this were the case.

I ment actual market value as opposed to MV or asking price.

You’re right, but the deck value stat that gets you a shiny badge is determined by game-calculated MV.

[Racoda]

Another solution to consider (somewhat similar to Almonaster's) is to make MV be influenced most by the most recent trades, ie. replace the current 10 trades average with a weighted average (or a similar system).
This doesn't stop inflation, but it makes inflated cards' MV much more volatile – if an inflate card sells at 0.01, this most recent trade can tank it pretty significantly. On the other hand, a recent inflation will also increase the MV much more.

This system doesn't hurt the MV of cards that are consistently traded at around the same price, but it quite correctly reflects the volatility of inflate-and-forget cards' MVs.

[HumanSanity]

Another solution to consider (somewhat similar to Almonaster's) is to make MV be influenced most by the most recent trades, ie. replace the current 10 trades average with a weighted average (or a similar system).
This doesn't stop inflation, but it makes inflated cards' MV much more volatile – if an inflate card sells at 0.01, this most recent trade can tank it pretty significantly. On the other hand, a recent inflation will also increase the MV much more.

This system doesn't hurt the MV of cards that are consistently traded at around the same price, but it quite correctly reflects the volatility of inflate-and-forget cards' MVs.

I don't see this as solving the problem unless you also increase the spawn rate of CTE cards. Otherwise, you make it easier to inflate massively. Sure, deflating gets harder, but it's never really been that adding sales doesn't adequately deflate but rather that there aren't enough opportunities to add sales to deflate