Detect artificial price hikes for deal rating

[Former user]

I hope it's ok to post this here. I've noticed that sometimes stores will inflate the price of a game prior to it going on sale, so that they can claim a very large discount. Cheapshark currently seems to fall for this tactic and assign a high deal rating.

For example "Painkiller: Recurring Evil" is currently on sale for $6.99, which is listed as a 90% discount, and Cheapshark has given the deal a rating of 10. However, a look at the price history (link below) shows that the regular price has been $9.99 for months, other than a few days just before the sale.

https://www.steamprices.com/us/app/206760/painkiller-recurring-evil

[ArmoredCavalry]

This is actually a side effect of how Steam lists retail prices on their search results (which CheapShark uses to gather prices).

Essentially if an individual game is cheaper in a pack, then Steam will list that pack's sale price and retail price for the individual game. So, the Painkiller Pack was $6.99 (usually $69.99), and this price info was listed for the individual game.

I've seen this problem come up a few times. One possible fix I could look into is to never increase a retail price by more than X%. However, in some cases the retail price could have really been set too low as an error, and wouldn't get set back correctly...

So, not sure there is a great solution to this particular issue.

[Folksong]

How important is the "discount %" for your deal rating algorithm? Do you have your own database of previous sale prices that it could be compared against instead?

I would say personally that when I look at a deal the "% off" has no importance, it's just marketing. Just because it was listed at a high price at some point doesn't mean anyone in their right mind would ever pay that much. What matters is the absolute price and how it compares to previous sale prices for the title.

[ArmoredCavalry]

Discount % is one part of the deal rating, but can be outweighed by other factors.

I do have all previous sales prices, the problem there is performance though. If I'm gathering every single previous sale price, for every single price shown on the page, that is a pretty big hit. Also, the thing I worry about is that this would still be open to abuse. For example a brand new unknown developer could set their game to a $60 price tag, then set it down to $5 (the real price). Less likely, but still very possible...

I see what you mean about % off being less important that current price compared to previous prices though. I might be able to at least do a sort of "sanity" check, so that if a game has only dropped x% vs the previous price, it can't jump up to a 10.0 rating.

[Folksong]

That sounds like a step in the right direction, thanks for considering this issue.