Perfect card odds

[Josquin]

Is this being tracked? How many packs were opened so far, and how many perfect cards were generated? There seem to be too few of them in the AH right now. Are people hanging on to them, or are they more rare than the advertised 1 in 1000 rate?

[vyshka]

I have at least 2 teams in my entry pool with multiple perfects on the active roster.

[Markus Heinsohn]

Originally Posted by Josquin

Is this being tracked? How many packs were opened so far, and how many perfect cards were generated? There seem to be too few of them in the AH right now. Are people hanging on to them, or are they more rare than the advertised 1 in 1000 rate?

They are exactly 1:1000.

[Westheim]

Originally Posted by Josquin

Is this being tracked? How many packs were opened so far, and how many perfect cards were generated? There seem to be too few of them in the AH right now. Are people hanging on to them, or are they more rare than the advertised 1 in 1000 rate?

Does "small sample size" ring a bell with you? Or did you just calculate the total number of cards in the auction house and whether 1/1000 of those were perfect?

[sansterre]

Originally Posted by Westheim

Does "small sample size" ring a bell with you? Or did you just calculate the total number of cards in the auction house and whether 1/1000 of those were perfect?

Of course it's worse than that.

You should expect to see:

Total number of packs opened / 200 * (percentage of perfects that happen to be on the AH at that moment as Perfects, if sold, will likely only be sold once)

All I'm saying is that 1) you have no idea how many people that draw perfects are keeping them and 2) you have no idea how many of the people who decided to sell them have already done so.

[zrog2000]

Originally Posted by Markus Heinsohn

They are exactly 1:1000.

According to my personal calculations, the odds are 0.

[Westheim]

Originally Posted by zrog2000

According to my personal calculations, the odds are 0.

Mine as well, although I will ever only draw one at most, because one would be enough to make me snap and being institutionalized.

[Whoofe]

Originally Posted by zrog2000

According to my personal calculations, the odds are 0.

lol

[Matt Arnold]

Originally Posted by sansterre

Of course it's worse than that.

You should expect to see:

Total number of packs opened / 200 * (percentage of perfects that happen to be on the AH at that moment as Perfects, if sold, will likely only be sold once)

All I'm saying is that 1) you have no idea how many people that draw perfects are keeping them and 2) you have no idea how many of the people who decided to sell them have already done so.

The auction house is also far from an even distribution. I mean, probably 1/2 or 2/3 of the cards that I draw from packs are simply auto-sold because there's no auction market for them, so it's going to skew towards the top. But on the flipside, not everyone is selling their top cards either, as I'm sure a lot of people really like their perfect cards.

[Kushiel]

Originally Posted by sansterre

All I'm saying is that 1) you have no idea how many people that draw perfects are keeping them and 2) you have no idea how many of the people who decided to sell them have already done so.

I agree totally. I was lucky enough to draw the Perfect Lou Gehrig War card. It is not now, nor will it ever be for sale.

[ASURay]

I drew a Trout and two Betts in a total of ~60 packs since the beta ended. I sold the Betts. There is usually at least one of the Betts cards on the AH. The historical perfects seem better and less common. I can't decide whether I'd sell a perfect if I didn't have a duplicate. 55-60K PP (I sold the Betts for 58K - the 10% fee) is nice and can get you a really nice gold team on the AH plus prb 25 packs or so, but that perfect card can also be a guaranteed producer on your roster for the remainder of time. My cousin drew a Betts as well and he is deciding whether to keep it or sell it to get a bunch of low golds plus 25-30 packs. Interestingly, there a quite a few high diamond cards that sell for more than a few of the perfect cards. I want the 1996 Smoltz and it sells for as much as the Betts.

[Josquin]

Originally Posted by Westheim

Does "small sample size" ring a bell with you? Or did you just calculate the total number of cards in the auction house and whether 1/1000 of those were perfect?

Thanks for the statistics lesson. In my own experience, dating back all the way to the beta, I opened over 600 packs without drawing a single perfect. Perhaps you can enlighten me on the odds of that?

[sansterre]

4.94% for 600 packs. Not at all probable but definitely possible.

[Josquin]

Originally Posted by sansterre

4.94% for 600 packs. Not at all probable but definitely possible.

I get 2.7%. 0.999 to the 3600th power? Not taking into account the mandatory bronze that sometimes happens with crappy packs.

[Westheim]

Originally Posted by Josquin

Thanks for the statistics lesson. In my own experience, dating back all the way to the beta, I opened over 600 packs without drawing a single perfect. Perhaps you can enlighten me on the odds of that?

Sure! It's an easy math exercise, actually.

600 packs are 3,000 cards (since one card can never be a perfect). The odds that a card is not perfect is 999/1000, so the odds for none of 3,000 cards to be perfect are (999/1000)^3000

Which is 4.97%.

[Josquin]

Originally Posted by Westheim

Sure! It's an easy math exercise, actually.

600 packs are 3,000 cards (since one card can never be a perfect). The odds that a card is not perfect is 999/1000, so the odds for none of 3,000 cards to be perfect are (999/1000)^3000

Which is 4.97%.

Why only 5 chances per pack? The mandatory bronze only happens when you draw a common on the first 5 tries.

[Galeg]

Originally Posted by Josquin

Why only 5 chances per pack? The mandatory bronze only happens when you draw a common on the first 5 tries.

I'm pretty sure you are guaranteed a bronze, not at least a bronze.


Also, even if you bought x amount of packs and got your percentage down to a tiny number, that still only makes you a statistical anomaly.

[Josquin]

Originally Posted by Galeg

IAlso, even if you bought x amount of packs and got your percentage down to a tiny number, that still only makes you a statistical anomaly.

There are two competing hypotheses: One is that I'm a statistical anomaly, and one is that the odds are not 1 in 1000. As additional evidence accumulates, one of the hypotheses becomes increasingly more likely than the other.

[zrog2000]

Originally Posted by Josquin

There are two competing hypotheses: One is that I'm a statistical anomaly, and one is that the odds are not 1 in 1000. As additional evidence accumulates, one of the hypotheses becomes increasingly more likely than the other.

One hypothesis is that it's a statistical anomaly. The other that you're suggesting is that the devs are liars.

[sansterre]

Originally Posted by Josquin

There are two competing hypotheses: One is that I'm a statistical anomaly, and one is that the odds are not 1 in 1000. As additional evidence accumulates, one of the hypotheses becomes increasingly more likely than the other.

The odds are 1 in 1000 per card, but 1 in 200 for a pack, as a pack contains 1 bronze and 5 random cards.

4.94% is about the lifetime home run rate of Jason Bay, which is to say, it definitely happens, but you wouldn't exactly want to set your watch to it.

Please remember that the 'average' Perfect card draw from 600 packs should be 3. But only about 22% or so actually get 3. 39% or so get less than three, while the remaining 39% will receive more than 3. Which means that 78% of all users will be able to point to the fact that they opened 600 packs (a considerable amount) and did not receive the amount they should have, they received too many/too few and therefore the odds are actually higher/lower than is listed.

tl;dr Drawing no Perfects in 600 packs is unlucky but proves/disproves nothing. It's like randomly going up to someone, finding out that she has a 125 or better IQ and saying "obviously this wasn't a random sample because the odds that I would just randomly grab someone with such a high IQ is quite low" (in fact around 5%). Just because it's unlikely doesn't mean that it can't happen, which seems to be your implication.

Markus and the Devs have much, much, much better data on the effective probability of drawing a Perfect than we do (inasmuch as we can only look at 600-700 packs, where they're looking at 100k packs easy) and they equally have little incentive to lie about it.

For me, I've probably opened 700 packs and received 4 Perfects; 1 Trout, 1 Betts, 1 deGrom and 1 Mariano. Luckier than I deserve.