Talk:Showdown Update/@comment-36249842-20190724133330/@comment-34918040-20190727123411

@ Tranzistorua:

I've shown the example with common, rare and legendary because these are the only values where we have official drop rates to compare them with. As you agree that the statistical outcomes will converge to the drop rates in the long run, let's take the next step:

If statistical values have been proven to be reliable for common, rare and legendary, the statistical outcomes for other card types (like engines or C-Class Suspension) will also be reliable, even if there are no official values to compare them with. If the outcome for a card type converges to a value, than this value is the drop rate (=expected value) for this card type. Agree?

That's what the page Pro Kit Box statistics comparison is for. It doesn't show official Gameeloft values, but real outcomes, no matter what Gameloft wants you to believe. Go to the card you need, sort the table, and it will show you the boxes tht currently grant most of your desired card.

But you have a point if you say that Gameloft could simply grant other rare cards instead of the rare card you need and thus keep the drop rates intact while withholding what you need. I already thought of this, too, but didn't mention it so far, waiting for someone clever to find it. There are two arguments against this point of view:
 * Gameloft cannot now what you need or plan. They can if you are participating in an event, presuming that you need cards for the event car. But as for the garage: I have many cars close to pro, but don't care about proeing them. On the other hand, I have several cars which are at a very low pro level, but they are on my definite list to upgrade for future events. Simply didn't do it so far. An algorithm withholding cards for the near max pro cars would thus grant me tons of cards I need, doing the exact opposite of what it was supposed to do.
 * A little bit more sophisticated and not visible without statistics: An algorithm as you suggest would not distort the frequencies of common, rare and legendary cards, that's true. But it would distort the frequencies of other card types the Statistics Project is observing. However, these other frequencies haven't been distorted so far. That's why I've started the V8 test. There's no V8 in my inventory, and there will be the Trevita event, so I'm definitively in need of V8s. I'll publish the results when the test is over.

But all this doesn't help you with your C-Class Suspension problem. I guess you may have had the wrong strategy. Let me suggest something, please tell me how it worked out:


 * 1) Go to the Pro Kit Box statistics comparison page, scroll down to the Parts section and sort first by Suspension, then by Class C. (I did not include values for single part cards when I wrote the code as I didn't want to produce a data overhead, but the double sorting should do the job as well.)
 * 2) Look at the guarantees at the top of the section. The  guarantees 1 of every C-Class part. Every Welcome Box you open will grant 1 C-Class Suspension.
 * 3) Proceed by the table entries from top to bottom. The  has a current frequency of 32.31 % Suspension. That's huge. I've just tested it, opened boxes and got 8 Suspensions.
 * 4) Try . The statistical values for this box have a bias towards Engines and Tech, as part cards were always rejected, so the real outcomes for Suspension should be even higher when you reject all cards but C-Class Suspensions. Don't give up too quickly. It may take a while, but if you get C-Class Suspension, you even get 2 of them becausse the box always grants pairs.
 * 5) Take . The statistics for this box show the frequencies of all four offered cards, so the shown values are the real thing.
 * 6) Go down further in the table.

This strategy always worked for me. I may be completely wrong, and perhaps there is an algorithm that doesn't like you, but personally, I am very sure that this will increase the your frequency of C-Class Suspension significantly.

One last note to give you a picture of the amount of revealed cards you need: If a box has a drop rate of 20 % Suspension and 20 % C-Class, it means that only 20 % of 20 % = 4 % (!) of all cards are C-Class Suspension, and this only if the distribution of classes and part types was uniform—which it's not. Don't expect more than 4 C-Class-Suspensions out of 100 cards!