GDC: Randy Smith Doesn't Save the Day

For every yin there is a yang. Though on the one hand it is comforting to be able to save and load at will, continually loading – thereby undoing events, and making consequences irrelevant – tends to diminish a player’s belief in the game world, making it all the harder for the game to affect the player in a meaningful capacity. The situation is kind of damned if you do, damned if you don’t, in that allowing free saving tends to lead to abuse yet disallowing it leads to player complaints. EALA game designer Randy Smith gave a brief speech on the psychological factors that tend to result in save abuse, and how potentially to avoid or undermine those triggers, such that players are tempted to save and load far less often, thereby allowed to take their in-game experiences at face value. Players base their decision to load on a form of loss analysis – not so much as risk analysis, as virtual loss is different from loss in real life. Risk is, in fact, part of the whole appeal to video games. Instead, players are worried about risks to their play experience. Even if they know there is no quantitative loss to falling from a high platform, players will often be adverse to complex platforming segments simply because of the waste of time when they fail. To contrast, one of the greatest attractions to Resident Evil 4 is the spectacular ways the player can die, and both learn and be entertained by that experience. Instead, Smith dubs what players mull over is “regret analysis” – “How did I play?”, versus “How could I have played?”. If the player feels he has “screwed up”, or was unfairly ambushed, he will load. Ideally, players should judge when to load on the basis of a direct loss comparison. The player knows, when loading, he will lose progress. If the amount of value lost since the most recent save point is less than the value gained, the player should press on; if the player has lost more than he has gained, the logical action should be to load. The problem is, players do not necessarily have the time or inclination to make rational analyses. Instead, what players tend to work by is a shortcut that Smith dubs “prospect theory”: when people experience a small loss, they are more likely to take a gamble. The goal in game design, Smith says, should be to lead players to make more rational choices about when to save and when to load. Part of that is in determining when players are more likely to accept a loss. Typically, when players are unsure they could have done any better, they will continue – as they will if the pending loss has been well telegraphed beforehand. Likewise, if players develop an attachment to the characters or the game systematically recognizes the player’s dynamic choices, the player feels a sort of ownership over his own actions, making him less likely to overwrite them. Another issue is “re-aquirement” – situations where the player is given the opportunity to gain back part or all of his loss. People tend to assume, Smith asserts, that “good things follow bad”. Although this is irrational in real life, there is no reason not to design games to reinforce the impulse. For example, the game might place a full-restore medpack right after a difficult boss battle. Half-Life 2 spaces out recharge points evenly enough that the player is only rarely left scampering for his life, and even when that happens, the player knows a recharge point must not be too far ahead. The inclination is therefore to press forward, instead of jumping back. If the game makes it tough or boring to get things back, though, the player will find it easier and less frustrating to simply try again. An example Smith gave is Thief, in which loss is costly and both time-consuming and frustrating to regain. When the player is caught, therefore, there is little reason not to load rather than allow the situation to play out. Smith wrapped up for lunch with a final few observations. Memorable losses lend to a greater perception of risk, so it is best not to taunt the player for losing and to try to make victories as memorable as possible. If you provide players with a choice of risks to take, and how to take them, they tend to accept ownership of their decisions and roll with them.