Exploring the 'Best Games': Spider and Web by Andrew Plotkin

It’s harder to talk about Spider and Web than the other games, because so much has been said already. One classic discussion can be found here. It won Best Game in the year that Anchorhead, Photopia and Losing Your Grip came out. It has topped many polls about the Best Games of All Time, is high on the IFDB top 100, has more ratings than most other games on ifdb, etc.

My personal feelings on it are positive. I enjoyed the first half when I first encountered the game, and I solved the ‘big puzzle’. But the second half was far too hard for me. Revisiting it today and yesterday, I found the second half much more enjoyable, now that I’m more familiar with the tropes of IF.

So, here are some good design techniques:

==Subversion of expectations==

Spider and Web uses this technique multiple times. It is used in the opening, where the entire genre is subverted. You expect it to be a simple exploration game, when suddenly you are hit with the interrogation scenes. It happens again moments later when You realize that you have a huge inventory of spy tools. Two other big subversions occur later when You reach the ‘big puzzle’ and immediately afterward when you realize that everything you thought you knew about your experience was false.

Subverting expectations is not necessary for a great game, but it certainly helps. Many of the most-rated games on ifdb are games that subvert the players’ expectations. The subversions in this game come in several flavors, some based on IF tropes and other based on standard fiction tropes.


As with the other Best Games I’ve been playing, Spider and Web has great depth. Things that you want to do, you can do, and if you cant, there’s a good reason.

The tools are all intricate, described in great deal, and can match up in a large number of ways. This leads to a sort of sandbox environment.

Perhaps the greatest depth is in the constant comments by the interrogator. These are very well developed, covering every imaginable situation. This must have been very hard to code

==Great NPC==

The Best Games tend to have very memorable characters. Some have great PCs, some have great ensemble casts, but Spider and Web is notable for one excellent NPC, the interrogator.

It’s rare to have a nemesis in IF that is so well developed. The design choice of having just yes/no questions reinforces the PC’s confinement and removes all issues with ‘guess the topic’ which usually plagues conversation in IF. And the yes/no answers are tabulated and commented on by the NPC, which leads back to the ‘depth’ earlier.


I could touch on the puzzles in the game, but there greatness doesn’t stem from each individual puzzle, it stems from their interaction with the topics mentioned above: the subversion of expectations, the depth of the gadget implementation and the guidance/pushing from the interrogator.

People enjoy pointing out things they dislike about this game, and I have done so myself, but it is only a target for negative feedback because it’s so well known. No one can dispute that this is one of the most important IF games of all time.

Thanks for this series of posts, first!

The interrogator was tedious to code but not really difficult. I just had to tabulate the ways each scene could end, either unsuccessfully (player does something stupid) or successfully (player reaches the correct ending move but with some intermediate state not set). This was was a not-too-long list for each scene. Then I just needed first/second/maybe-third responses for each outcome.

As for the mood-tracking over the whole game – as I’m sure I’ve said many times, I think that element completely fails to come across to the player. If I were designing the game today, I might well go with visible mood stats, ChoiceOfGames-style.

That’s really interesting; I don’t think I’ve ever read anything by you about the creation of Spider and Web. A choice of games-esque versions would be/would have been really interesting.