Yeah, I looked at it but did not link it directly :).

Thought the "Build in progress, don't close this window." in conjunction with the non-grayed out Close button is a true example of respecting the user's free will. I mean you can click on it if you really want to shoot yourself in the leg, but we told you not to do that :P.

Seems like GOG is pushing the boundaries of their assigned ID range. We have a new entry for the highest value of a registered ID: 2147483059 - belonging to the upcoming The Bard's Tale Trilogy.

The ID is just 588 ids, just a stone's throw, away from the max limit of 2147483647.

did you ever found some sort of pattern of how the IDs are chosen? It feels to me that they are just picked at random.

My working hypothesis is that they are using a combination of trained monkeys as PRNGs and procedurally generated GOGBears as entropy selectors to land the right ID. In short, no, but it doesn't look entirely random :P.

Probably hash code or crc on the title...

... which is then restricted to a 32bit unsigned integer-sized target set. Not that I think it's impossible, but it sounds unlikely.

The approach I've taken is purely statistical, though my predictor is nowhere near to being a vaguely reliable oracle at this point. In a nutshell: I need to collect more data and improve the model, and who knows, at some point I might get some hits.

I will consider posting a frequency distribution graph of the entire ID range with the data I have at this point... that is bound to raise some eyebrows.

An XY graph with release date and product ID as the axes could be useful.

I actually have one of those lying around somewhere, with a derived discrete probability distribution chart on id ranges, both generated in FreeMat, but they are rather dated by now. Generating them automatically with PyChart, using live data sets, is on my "to do" list, so watch this space.

Here you go, but remember that you asked for it :). Filtered out the first 10 ids and some noise from my initial full range scan, so this is just useful incremental data.

Ended up using matplotlib for the plot in the end, felt more comfortable with it.

Very nice chart, thanks for making it! I expected at least some patterns but this indeed looks very random.

You're welcome. There are still a few things I want to port/adapt to matplotlib, so do expect a few more graphs.

It does indeed look random, perhaps a bit too random for it own good (seems a bit too evenly spread, no localized bias etc). Can't say much more at the moment, but I'll keep collecting plot points, slowly getting a bigger picture of it all... and I'll freshen up on statistics while I'm at it - it's been a while since I've done this sort of thing.

Two more graphs some might find interesting: distribution graphs of ids per id ranges (intervals), both with full data, excluding the first 10 ids, and only incremental data - the new stuff I've recorded after my full sweep of the id range.

Edit: added color to the more... interesting areas.

Here are some alternative graphs, expressing pretty much the same data as the above histograms, but considering the discrete probability in an ideal case (equal chances). I'll also include some numbers for the incremental chart (dpn).

Discrete id probability values per interval:
1070000000: 0.005201560468140442
1080000000: 0.013003901170351105
1090000000: 0.010403120936280884
1100000000: 0.007802340702210663
1110000000: 0.007802340702210663
1120000000: 0.014304291287386216
1130000000: 0.007802340702210663
1140000000: 0.016905071521456438
1150000000: 0.0039011703511053317
1160000000: 0.011703511053315995
1170000000: 0.007802340702210663
1180000000: 0.009102730819245773
1190000000: 0.011703511053315995
1200000000: 0.002600780234070221
1210000000: 0.0039011703511053317
1220000000: 0.007802340702210663
1230000000: 0.005201560468140442
1240000000: 0.016905071521456438
1250000000: 0.007802340702210663
1260000000: 0.015604681404421327
1270000000: 0.011703511053315995
1280000000: 0.006501950585175552
1290000000: 0.006501950585175552
1300000000: 0.007802340702210663
1310000000: 0.010403120936280884
1320000000: 0.005201560468140442
1330000000: 0.009102730819245773
1340000000: 0.0039011703511053317
1350000000: 0.011703511053315995
1360000000: 0.009102730819245773
1370000000: 0.013003901170351105
1380000000: 0.007802340702210663
1390000000: 0.013003901170351105
1400000000: 0.009102730819245773
1410000000: 0.009102730819245773
1420000000: 0.013003901170351105
1430000000: 0.009102730819245773
1440000000: 0.009102730819245773
1450000000: 0.014304291287386216
1460000000: 0.0039011703511053317
1470000000: 0.006501950585175552
1480000000: 0.010403120936280884
1490000000: 0.009102730819245773
1500000000: 0.009102730819245773
1510000000: 0.0013003901170351106
1520000000: 0.007802340702210663
1530000000: 0.009102730819245773
1540000000: 0.0039011703511053317
1550000000: 0.007802340702210663
1560000000: 0.010403120936280884
1570000000: 0.011703511053315995
1580000000: 0.007802340702210663
1590000000: 0.015604681404421327
1600000000: 0.01950585175552666
1610000000: 0.010403120936280884
1620000000: 0.011703511053315995
1630000000: 0.013003901170351105
1640000000: 0.006501950585175552
1650000000: 0.006501950585175552
1660000000: 0.006501950585175552
1670000000: 0.009102730819245773
1680000000: 0.007802340702210663
1690000000: 0.002600780234070221
1700000000: 0.009102730819245773
1710000000: 0.006501950585175552
1720000000: 0.002600780234070221
1730000000: 0.015604681404421327
1740000000: 0.014304291287386216
1750000000: 0.007802340702210663
1760000000: 0.013003901170351105
1770000000: 0.007802340702210663
1780000000: 0.009102730819245773
1790000000: 0.005201560468140442
1800000000: 0.010403120936280884
1810000000: 0.013003901170351105
1820000000: 0.010403120936280884
1830000000: 0.01950585175552666
1840000000: 0.005201560468140442
1850000000: 0.007802340702210663
1860000000: 0.007802340702210663
1870000000: 0.005201560468140442
1880000000: 0.009102730819245773
1890000000: 0.006501950585175552
1900000000: 0.007802340702210663
1910000000: 0.005201560468140442
1920000000: 0.013003901170351105
1930000000: 0.009102730819245773
1940000000: 0.011703511053315995
1950000000: 0.009102730819245773
1960000000: 0.010403120936280884
1970000000: 0.006501950585175552
1980000000: 0.010403120936280884
1990000000: 0.01950585175552666
2000000000: 0.006501950585175552
2010000000: 0.009102730819245773
2020000000: 0.010403120936280884
2030000000: 0.007802340702210663
2040000000: 0.007802340702210663
2050000000: 0.006501950585175552
2060000000: 0.007802340702210663
2070000000: 0.007802340702210663
2080000000: 0.010403120936280884
2090000000: 0.011703511053315995
2100000000: 0.0039011703511053317
2110000000: 0.007802340702210663
2120000000: 0.010403120936280884
2130000000: 0.013003901170351105
2140000000: 0.015604681404421327

Seems GOG has preformed some changes to the data stored in the famous gogData variable, more precisely bothersome to me, they've added:

"video":{"id":"8-mf0cX_6EY","provider":"youtube"}

video type entries for some games, which contain id tags that look awfully like regular game ids, at least they do to my very simple non-DOM-aware text parser.

That has briefly broken my new ids detection script, but I've added some code to ignore these bogus entries for now. It's high time I switched to a regular JSON parser for gogData though, so this will be the next item on my dev list.

A different point to consider is: could this be a sign of an upcoming site refurbishment?

Found what seems to be a test entry for "UE4 Galaxy Test Game".

Does anyone have an idea on what "UE4" stands for?