Question about how Farsight tracks table sales

[TNT]

so we all know that Farsight tracks how well tables sell. For example The Addams Family has sold very well as was expected.

How does Farsight classify it when someone buys the season packs over individual tables?

 

[dave950lam]

You can get a good idea of how well a table did by looking up the leaderboard standings (though, it may take a while). how many entries will give you the approximate number of downloads.

 

[TNT]

Interesting. Thanks.

 

[invitro]

You can get a good idea of how well a table did by looking up the leaderboard standings (though, it may take a while). how many entries will give you the approximate number of downloads.

Is there any way to find this other than* scrolling through all the ranked scores? There are tens of thousands of them on most tables, and a few are in the hundreds of thousands, IIRC. Scrolling through 10 scores at a time would take a while... even if you could do 10 scores (a page) a second, it would take >2.5 hr to do a table with >100,000 scores. And "only" 17 minutes for a table with 10,000 scores.

So my question for you is: Have you actually done this?

* I suppose if you knew the username of the player at the bottom, you could look them up on the webpage to see their rank. This doesn't seem very effective, though.

P.S. I got the last ~20 scores for every table when I sucked the scores from the webpage until the webpage got locked to 510 scores, which was shortly after May 6 of this year.

 

[MontanaFrank]

Is it correct to add the denominator that some tables will have X amount of non-purchased leader-board additions because of the free table of the month?

 

[invitro]

Here is some raw MySQL output that shows the rank ("max(rank)" below) of the lowest-ranked score for each table on 2015-05-06. (And glitchy or hacky scores can be seen, but I hope there are no errors here, or I'll have some work to do.)

TotAN must have A MILLION scores by now... which means at least a million registered usernames for TPA, excluding XBox platforms. (People can and do have >1 username, but I have no idea of the average # of usernames per person. I would guess it's under 1.5, though.)

mysql> select table_id, t.abbr, min(rank), max(score), max(rank), min(score) from tpa_leaderboard l join tpa_table t on l.table_id=t.id where date='2015-05-06' group by table_id;
+----------+----------+-----------+----------------+-----------+------------+
| table_id | abbr     | min(rank) | max(score)     | max(rank) | min(score) |
+----------+----------+-----------+----------------+-----------+------------+
|        0 | TotAN    |         1 |    76512635720 |    973060 |        922 |
|        1 | RBIoN    |         1 |    37300267340 |    253292 |          0 |
|        2 | ToM      |         1 | 14078474451920 |    257704 |      10050 |
|        3 | BH       |         1 |     1549025135 |     66368 |        309 |
|        4 | BoP      |         1 |    73712835670 |    103479 |      50000 |
|        5 | MM       |         1 |  4540350587400 |    112893 |      15000 |
|        6 | CV       |         1 |     3257841420 |     86943 |      69880 |
|        7 | FH       |         1 |     3099025010 |     57614 |     195000 |
|        8 | MB       |         1 |    41308269070 |     77722 |      65720 |
|        9 | Gorgar   |         1 |     1672055480 |     74602 |         40 |
|       10 | CftBL    |         1 |   312505443610 |     78715 |    1000000 |
|       11 | BK       |         1 |      458512640 |     73791 |        100 |
|       12 | HD3E     |         1 |     2689815940 |     72341 |    1775960 |
|       13 | Taxi     |         1 |      168394480 |     71347 |       1000 |
|       14 | EatPM    |         1 |     2550483880 |     73525 |      13010 |
|       15 | NGG      |         1 |    91354241750 |     75738 |      30000 |
|       16 | BigShot  |         1 |      193278500 |     71357 |         10 |
|       17 | SS       |         1 |    72455767980 |     70239 |       5490 |
|       18 | TZ       |         1 |   214795450970 |     81707 |    1089370 |
|       19 | STTNG    |         1 |  1890199057550 |     87240 |     100000 |
|       20 | AFM      |         1 |   979134836240 |     88586 |      85000 |
|       21 | Genie    |         1 |      739001170 |     63244 |       5200 |
|       22 | DrDude   |         1 |      681963180 |     28461 |      24450 |
|       23 | FirePwr  |         1 |      627774000 |     26973 |       1000 |
|       24 | CC       |         1 |     3203482110 |     31996 |      74280 |
|       25 | CentPark |         1 |         211080 |     27773 |         31 |
|       26 | WWater   |         1 |    99934777050 |     32511 |      55000 |
|       27 | SpSh     |         1 |       76844330 |     28194 |       2000 |
|       28 | Centaur  |         1 |      234567890 |     27211 |        200 |
|       29 | Pin*Bot  |         1 |      202119730 |     27329 |       1020 |
|       30 | CP       |         1 |     5344700720 |     29240 |         50 |
|       31 | WWind    |         1 |      488079650 |     27301 |     130120 |
|       32 | Flight2K |         1 |       15158160 |     20771 |       8000 |
|       33 | GoinNuts |         1 |       25272190 |     20106 |      10000 |
|       34 | T2       |         1 |    45714166080 |     39601 |     128080 |
|       35 | HH       |         1 |       33017260 |     25476 |       5000 |
|       36 | TeedOff  |         1 |    91139885780 |     23649 |     530000 |
|       37 | Co1812   |         1 |    47805292750 |     21412 |      20000 |
|       38 | Victory  |         1 |      825849980 |     20852 |        700 |
|       39 | CBW      |         1 |    26193524970 |     21061 |     500000 |
|       40 | ElDor    |         1 |      360053650 |     18833 |        500 |
|       41 | FT       |         1 |     9999658190 |     20265 |      71230 |
|       42 | BRose    |         1 |    80073912020 |     19453 |     205210 |
|       43 | BK2K     |         1 |    71200687840 |     15413 |       6020 |
|       44 | WD       |         1 |   706269894190 |     17969 |    7144090 |
|       45 | HS       |         1 |      281058310 |     15315 |       7000 |
|       46 | JY       |         1 |    11135004350 |     17140 |     180410 |
|       47 | LCA      |         1 |     2414679870 |     12341 |       5000 |
|       48 | HRC      |         1 |    20424874000 |     14040 |     125000 |
|       49 | Diner    |         1 |    73998483500 |     12708 |      75000 |
|       50 | BSD      |         1 |    63150296660 |     16381 |     917170 |
|       51 | PotO     |         1 |     1399999999 |      8716 |      44630 |
|       52 | PZone    |         1 |     5321064690 |      6940 |     260230 |
|       53 | ES       |         1 |    44600654370 |      6805 |     134000 |
|       54 | SST      |         1 |    19999999990 |      6955 |    2219910 |
|       55 | TAF      |         1 |    69394785960 |     17539 |     550010 |
|       56 | Cyclone  |         1 |       58138650 |      4429 |      32060 |
|       57 | J*B      |         1 |   173189263390 |      3689 |  155692150 |
+----------+----------+-----------+----------------+-----------+------------+
58 rows in set (0.22 sec)

 

[invitro]

Is it correct to add the denominator that some tables will have X amount of non-purchased leader-board additions because of the free table of the month?

Excellent question! I think it is indeed correct, unless scores on the monthly free table aren't eligible for the leaderboard, which seems very unlikely.

I don't have any idea of how to estimate this number though. I would love to, if anyone has some spare tuits!

EDIT: Comparing max(rank) before and after the monthly free table period might lead to a good estimate. But scores on a demo table (when a table is unpurchased, and has the small time limit) probably go on the leaderboard too. Anyone know?

 

[Worf]

Well, the other way is online tracking. Like every time you complete a game, your score is posted to Farsight PinballAnalytics.com site. It's how those high score lists are generated.

There's always been a complaint that if you get a high score and you go offline, you lose it because it's not posted online.

I'm sure Farsight uses that information to determine how popular tables are by how many unique machines post scores, and how often they post scores online.

 

[EldarOfSuburbia]

There is a way! I happen to think it's a bug with the website, but right now it works very well.

Go to the website leaderboards for any table, and click on anyone's name (so in my example I've clicked on Tarek's name listed as the #1 score in TAF - but it doesn't matter).

You then get your score comparisons across all tables - owned or not - for the player whose name you clicked on.

For each table you get the following output (sorry for the formatting, bad c&p):
Addams Family
Tarek Oberdieck
Rank: #1 (pc)
Score: 84,543,403,430
EldarOfSuburbia
Rank: #26871 (pc)
Score: 5,342,810,560

So where it gives your "rank" it's actually giving the number of scores in the database for that table. (The "bug" is, it should list my rank as #25.)

Okay, so I've started so I may as well extract them all. A few minutes of c&p.

 

[EldarOfSuburbia]

Here's the results (again apologies for any bad formatting - sorted high to low):
Tales of the Arabian Nights 1071733
Theatre Of Magic 267050
Ripleys Believe It Or Not 261872
Medieval Madness 122601
Bride of Pin-Bot 112191
Attack From Mars 99575
Star Trek, Next Generation 96778
Cirqus Voltaire 94143
Twilight Zone 90419
Creature of the Black Lagoon 86459
Monster Bash 85613
No Good Gofers 83094
Gorgar 81324
Elvira and the Party Monsters 80675
Black Knight 80650
Harley Davidson (3rd Edition) 79685
Big Shot 79480
Taxi 78662
Elvira Scared Stiff 77865
Black Hole 74937
Genie 70023
FunHouse 65128
Terminator 2 : Judgement Day 49356
White Water 38737
Cactus Canyon 37918
Champion Pub 35058
Dr. Dude 33314
Space Shuttle 33136
Whirlwind 32715
Central Park 32605
Centaur 32393
Pin-Bot 32376
Firepower 31424
Haunted House 31337
Teed Off 28662
Addams Family 26871
Class Of 1812 26214
Cue Ball Wizard 26165
Victory 25875
Flight 2000 25076
Goin Nuts 24366
Fish Tales 23651
El Dorado : City Of Gold 23370
Black Rose 23038
WHO dunnit? 21485
Bram Stokers Dracula 20723
JunkYard 20382
Black Knight 2000 18641
High-Speed 18398
High Roller Casino 17291
Diner 15613
Lights... Camera... Action! 15140
Starship Troopers 12624
Phantom Of The Opera 12534
Party Zone 11208
EarthShaker 11165
Cyclone 9055
Xenon 8882
Jack-Bot 8474
Red & Teds Roadshow 5987
SafeCracker 5515
Addams Family : Gold Edition 4576
The Getaway : High Speed 2 3575
SafeCracker Assault on the Vault 1300

 

[shutyertrap]

That's a pretty steep drop off for season 4. I really find it hard to believe that 17% of TAF owners have TAF Gold though. And with a minimum buy in of $100 to get it, that means FarSight would have at least cleared $457,600 in the kickstarter! So these numbers don't tell the whole story.

 

[invitro]

Here's the results (again apologies for any bad formatting - sorted high to low):

Fantastic! Stupendous!

 

[EldarOfSuburbia]

That's a pretty steep drop off for season 4. I really find it hard to believe that 17% of TAF owners have TAF Gold though. And with a minimum buy in of $100 to get it, that means FarSight would have at least cleared $457,600 in the kickstarter! So these numbers don't tell the whole story.

There was a brief period where TAF Gold was accidentally available to a lot more folks than just the KS backers. So those scores may have come from then.

Also I think it was possible to gift TAF Gold codes to other users. The KS page lists 359 backers at $100+.

 

[invitro]

That's a pretty steep drop off for season 4. I really find it hard to believe that 17% of TAF owners have TAF Gold though. And with a minimum buy in of $100 to get it, that means FarSight would have at least cleared $457,600 in the kickstarter! So these numbers don't tell the whole story.

The season 4 drop does look pretty steep. I wonder if the (real-life) S4 tables are on average less-notable than the S3 ones... they seem about the same to my eye.

TAFG quietly replaced TAF for a few days, a few months ago, on at least PC. I think I posted about it in the TAF forum, and have seen a few other mentions of it. I played it a few times and took many screenshots. I don't know if that's the whole story, but it might be.

 

[EldarOfSuburbia]

Part of the S4 drop-off is likely to have a couple of reasons:
1. Not all of S4 is available on PS3/4/Vita yet, so there may well be users of those systems holding off on getting S4 passes until then. (Of course the same goes for X360/XONE as well, but those scores don't contribute to the numbers I listed.)
2. iOS S4 passes weren't available until Safe Cracker was released, and there may be iOS users who are still holding off getting the pass. iOS is a pretty big chunk of the user base.

 

[invitro]

So these numbers don't tell the whole story.

Also, there's that bug with the TotAN scores, where we believe that scores from other tables get marked as TotAN scores. I forgot this above ... Also, I can't remember if TotAN is the only free table, or if RBIoN and/or ToM is too.

There is a bit of a definition problem here... what does "whole story" mean? I think the numbers are accurately equal to the number of people who have played a table while being logged in, modulo the TotAN bug (I don't think any similar bugs are specifically known, though there is always evidence for strange things going on).

 

[invitro]

Part of the S4 drop-off is likely to have a couple of reasons:
1. Not all of S4 is available on PS3/4/Vita yet, so there may well be users of those systems holding off on getting S4 passes until then. (Of course the same goes for X360/XONE as well, but those scores don't contribute to the numbers I listed.)
2. iOS S4 passes weren't available until Safe Cracker was released, and there may be iOS users who are still holding off getting the pass. iOS is a pretty big chunk of the user base.

Awesome! You're on a roll today Eldar. I think reason 2 is more likely to have a huge effect... just being available for a (much) lesser time will dampen the numbers. Was the S3 pass available on iOS at the same time as the other platforms?

I sure hope I can remember all this stuff. I think there is a need for some kind of FAQ or other document with all of this technical information gathered into one place, if there isn't already.

 

[EldarOfSuburbia]

Was the S3 pass available on iOS at the same time as the other platforms?

S3 was the first season Apple prevented season passes from being available in advance.

 

[shutyertrap]

There was a brief period where TAF Gold was accidentally available to a lot more folks than just the KS backers. So those scores may have come from then.

Also I think it was possible to gift TAF Gold codes to other users. The KS page lists 359 backers at $100+.

But 359 backers is a lot different from 4500 unique users. And no way did FarSight give away a couple of thousand codes. Look, I'm not saying there isn't truth in the numbers you've pulled, but clearly there are other factors in play too.

 

[Gorgar]

But 359 backers is a lot different from 4500 unique users. And no way did FarSight give away a couple of thousand codes. Look, I'm not saying there isn't truth in the numbers you've pulled, but clearly there are other factors in play too.

Another factor (that still won't be enough to explain the discrepancy) is that some people got TAF Gold on multiple platforms and you can't use their Farsight IDs on PSN. So some users might have two scores posted.