I was wondering if anyone had given any thought to running the two Savage Frontier games. I know MsKain talked about it as a goal some time for one of his videos.

I was playing around casually with Gateway last night and it could be a fun one to run. Quite different to most of the other games in that there are really not that many magical items. It's probably also second only to Pool of Radiance in the Gold Box series for opportunities to sequence break.

Treasures (which I haven't played for ages) doesn't have as many IIRC, but it's considerably less linear than a typical Gold Box outing.

Nice one.

A few refinements and you’ll be easily under an hour. Here are a couple of strats Ifound useful:

1. There is a potion of speed in the thieves’ treasure room. If you use it before the final battle with the Fire Knives, they will surrender when morale breaks.
2. I got good results with three fighters, a fighter/cleric and two mages. I found the benefits of the paladin didn’t outweigh the extra experience you need. And having two fireballs available is a godsend in the earlier game.
3. Sleep works on a handful of monsters in the first section - royal guards, monkeys, the mages and dogs. It has no saving throw, affects multiple targets and is my go to spell.
4. Once you get the ring of wizardry, you can effectively have Haste permanently active. It makes a huge difference. Age seems to have no effect in this game. That’s not the case for later games though, so be more circumspect in the 3 game continue category.

Good luck!

Addendum: buy banded mail, not plate mail or (1) won’t work, as the potion will only boost your movement to 12. Monsters only surrender if they have an intelligence higher than 3 and if they’re not as fast as the fastest character. This works with Drow, cultists and Zhentarim as well, which is why Haste is your best friend.

I've personally found the trick is picking which battles it's safe to use Quick, which becomes more important at max cycles, because it's too fast to bail out if things go pear shaped. DKK has a lot of battles with high level mages and level drainers where quick combat is dangerous, especially because the route I use gets virtually no armour bonuses, so I'm going through the entire game with my starting armour. Haste still makes a difference.

One thing I really like about DKK is that Quick doesn't carry over between combats so you can pick and choose which ones you use it for.

My party has four elven fighter/clerics (bonus to hit with swords, Kiri-Jolth faith gives an extra THAC0 point and Enlarge can mitigate the lower max strength for the key fights), which helps a lot for the earlier battles, but it does make the many undead fights in level 2 of Dargaard dicey, as the AI will try to turn undead, which usually won't work, and effectively gives the enemy a free round of attacks. I might have to play around a bit with party composition and see if some better combinations can be worked out.

Me neither. It would be infuriating! Although curiously enough, I find DKK the one I lean on Quick the least. Too many battles with multiple level drainers in the final section. 😬

I wondered the same thing, although ultimately, I'm not sure if there's enough of a difference in the way it's played across cycle speeds to justify it. Having played a bit on both speeds, you do need to tweak your strats a bit, but not really all that much.

One thing that would make a difference (and render the computer speed discrepancies moot) would be to ban autocombat, but as that would make all the current runs invalid and probably make me want to rage quit, I'm not particular keen to go down that path! ;)

I agree too about the Steam release; if that plays significantly differently, that will require a new category.

Actually, one more thought on this run (6000 cycles or otherwise) since you mentioned time investments was that I remember once you were wondering about my strategy of fighting the fake Tyranthraxus and picking up his ludicrously powerful items (long sword +5, ring of prot +3, gauntlets of ogre power). On the face of it, it seems like an odd time investment as it's a time consuming trip and a challenging fight at 1st level, but the payoff in my opinion is considerable. It costs about 40 seconds (once you've factored in that you don't have to pick up items elsewhere) and combined with the bracers, it effectively lets you Quick Combat your way through the graveyard, which probably saves the 40 seconds there alone, then after that you only really need the necklace for crowd control. Which saves time there too as you don't have to dupe it as much.

I admittedly haven't tested this out, but it seems the faster way overall to me. Might be worth a shot to see if it makes a difference.

I think that’s fair about waiting to hear other people’s thoughts first. I’d be interested to know if there’s another way.

My sense was that Secret will have the biggest impact because there’s a lot of unavoidable combat and no dust. My use of the necklace was mostly to disrupt the numerous high level mages where you can’t always rely on your own mage. I’ve settled on a strat similar to Champions where you keep two characters off auto combat. Interestingly, the AI actually gets worse after Pool. Characters will cheerfully blow themselves up with wands if you leave them equipped, which Pool manages to avoid. I haven’t officially run Curse yet, but I’ve been doing a lot of work offline on that, and I have some spiffy new strats that make a huge difference - they don’t bust it open like Pool because it’s pretty linear, but they do make it much easier to lean on auto combat, which even at 6000 cycles, is much quicker.

Champions requires you to lean heavily on the wand, so it makes less of a difference there. I haven’t practiced Death Knights for ages, but my strategy there involves mostly grabbing good items early, duping them and going nuts so we should see some solid improvements there too.

Had a look at your max cycles stream on Twitch (couldn't watch live due to the time difference, but made a point to check it out afterwards). I'm more convinced now that this is the way to go, but if other people don't like it (or have a better way of ensuring consistency), all good.

1. I think there's still a discrepancy (your "tour" at the start of the game took about 2.5 seconds, compared to around 6 on my computer), but the difference is obviously going to be much smaller overall and overall time will depend more on speed on the keyboard and route optimiziation - which is good!

2. Higher speed reduces the impact of RNG - unavoidable random encounters in the wilderness with stirges, tigers and giant lizards can be dealt with in a few seconds rather than being minute long record killers.

3. Leaning more on Quick combat will be disproportionately quicker, making finding and duping good items early on to be a good time investment.

4. The main potential hazard is the fact that you can rarely hit the spacebar to end Quick combat in time, which can make complex encounters dangerous if you don't remember to turn it off the battle before.

5. If you do choose to upload your run, congrats on getting the first sub-40 time!

Cool. I'll have to start working on some strats that take advantage of it. :)

I'm not sure it is the best way, but it's the best one I can think of - happy to be corrected by someone more tech savvy. The way I figure it, if it's lightning fast on my plodding old laptop, the difference between a fast and slow machine can't be that much, even over a full run.

Hi all,

After a long absence, I've been starting to get back into running the five Gold Box games currently up and I've been doing some offline practice as well as studying some of the existing runs.

Something occurred to me when looking at LordBreakfast's Champions run (very nice BTW) - even with a constant setting for DosBox cycles, it seems to be strongly affected by the CPU speed - for example, I can't break 37 minutes at the moment at 6000 cycles with an effectively identical route, and while there is a lot of room for improvement, I'm not sure there's 5 minutes' worth. The combats on my runs certainly look a lot more sluggish.

Having looked into DosBox's documentation, it seems that it is affected by the speed of the computer you're using. There are also dozens of internal settings which affect performance. It makes it clear that the cycles speed is there to compensate for the difference between slow and fast computers, and while that's great for casual play, it makes any kind of consistency really difficult.

I don't think there's an easy solution, but having looked over some other older DOS games, there seems to be a convention of using the cycles=max setting. I've had a go at Champions with this and it's playable, although some of the strats need tweaking a bit. It would, however, close the gap between faster and slower computers from several minutes to probably a few seconds if that.

I'm not a massive fan of running at max speed because it's nice to actually be able to follow what's going on in combats rather than having them zip by in a flash, but it does seem to be the best way to take hardware out of the equation.

Thoughts?

Hey, welcome back! :)

I’m still plugging away offline with the Gold Box games - my laptop has just been a bit temperamental with my capture software so it’s just hard to get any videos of runs to upload.

Oh, and nice Pool FAQ BTW.

Hey, congrats on going under an hour! Very nicely done. 😀

I do know what you mean. I was watching one of SummoningSalt’s excellent WR progression videos the other day and he was talking about how a particular WR is contingent on a single enemy in a 10min run moving left or right because one way saves a handful of pixels. Suffice to say, I’m never going to have the patience to put in the thousands of attempts to get that level of play.

The single mage was a calculated risk - I figured the second mage was pretty much just ballast after the fight with the cleric. I’ve found that battle doesn’t go south too often - I watched the rest of your Twitch attempts and you had shocking luck on that time it didn’t work. Sleep affects 4d4 opponents, and a level 5 opponent counts as 10, so you’ve got around a 75% chance of getting him, once you’ve factored in the other two clerics.

You’re right about the party strength algorithm. The exact formula (which I got from Ssjlee’s fantastic FAQ, a document that was my bible for doing this) is:

0.4 for each cleric level 0.8 for each mage level 0.5 for each THAC0 point under 21 0.5 for each AC point under 0 0.1 for each current hit point. Then round down to the nearest integer for each character.

Random encounters in the slums are two thirds of party strength, rounded down to the nearest multiple of two. Leaders are added once they reach 8. I never thought of trying to reduce it to 4, which could potentially be a big time save, so that’s definitely worth playing around with. Oh and by the way, the number of undead in the graveyard is calculated when you enter, which is why in my 100% run, I unequipped everything before entering.

Oh, and I thought I'd add, more as an amusing addendum than anything else, how the game decides whether or not the temple guards let you through. Pool of Radiance has some ridiculously convoluted mechanics, but this one takes the cake.

So the idea behind it is basically this: Mace knows there are hidden treasures in the temple, but he can't find them. So instead of looking harder, he figures the council will probably send some adventurers to fetch them and so he decides to let the party find the treasure and then ambush them afterwards. Thing is, he doesn't know exactly when they're coming, so he sends his men off patrolling the area with the instructions to only attack adventurer looking sorts if they absolutely have to, and then to report back.

So: if you encounter groups of four orcs with a symbol of Bane on their shields and you talk to them or you run away, they'll report back to Mace, the temple guards will be told that you're coming and they'll let you in. If you don't encounter them or you kill them, they don't know you're coming and they'll check for a holy symbol. You can get one either by fighting the Orc Leader in Podol Plaza or by fighting any of Mace's patrols. If you have one, they'll let you in; if not, you'll have to fight them. In any case, the holy symbols seem to be incredibly heavy, so it's virtually impossible to have one anyway.

And the net result of avoiding alerting Mace? Zero. You have a trivial battle with eight orcs and you get attacked by Mace when you find all the treasures, leave the temple or destroy the altar regardless. It's a very complicated mechanic for something with almost no consequences (I suspect it was supposed to be more then got written out), but if you were wondering why you sometimes get stopped at the door and sometimes not, that's why.

Oh, and by the way: why do they rely on the blind orc to look for the holy symbol when there's eight perfectly sighted orcs who could do it just as well?

Hey, nice run, MSKain. And thanks for the shout out. :) I think a few run-throughs of that and some RNG luck and you'll be comfortably getting sub-13.

I've been practicing the Any% run over the last couple of weeks, and I did one which I thought was frankly a bit ragged, and was astonished to see it come in at 10:46 (hopefully the video works this time round - not sure what happened before). While a big part of that improvement is that I'm much more comfortable with the duping than I used to be, it got me thinking with this route at least, how much further it's possible to take it. The sum of my best segments is a 10:31, and with a cleaner run through the Slums, it could go below that. Sub-10? Don't know if that's viable, at least not without an alternate route.

The 100% route I'm less certain about because I've only run it through twice. My best is about 48 minutes or so, but that was more ironing out some of the more obvious mistakes in the earlier 52 minute one. Under 45 is definitely viable. How far under 45 is an open question at this point.

Some more observations from trying a few dry runs on this:

1. Buying the boy from the buccaneer base isn’t such a good idea after all. I thought they just take money, but they take a lot of magic items as well, which can throw a spanner in the works. The time you save you’ll probably lose again recuperating the losses, even if it’s just grabbing the swords from the fighters before fighting T.
2. This run is surprisingly forgiving. Unless you have a character killed outright, you can have people level drained or knocked out in key battles and miss out on experience and there’s enough so that you can get by comfortably.
3. On the same note, doing the graveyard as early as possible will minimise the amount of impact if you do get drained, the size of the encounters, and the anguish of losing a run late in the game. One third of the random encounters in the eight areas have mummies, which can quickly result in a run becoming unstuck.
4. The AI is better than I thought. You can safely use Quick combat with the necklaces of missiles because characters won’t use fireballs if there’s a risk of taking out their own side. Wands of lightning are dangerous though because the game utterly fails to take account of rebounding.

I’ll have to wait till I can sit down and try this all in one go now, but I’m going to go out on a limb and say that under 50 minutes might even be possible.

I only found it a few years ago, but great minds! :)

Using that and Quick combat (which I never really experimented with before - thanks MSKain!), and doing them en route to the hobgoblins and the treasure room in the NW corner, I've managed to go from starting to being rewarded by the clerk in under 10 minutes. It's definitely a massive time save.

But the slums is basically the game's training grounds, so I can see why it would require a lot of combats, but in terms of speedrunning, the slums actually has as many unavoidable fights as the rest of the game combined - although in practice, 8-10 random encounters are unavoidable too, and there are always a few battles to get to key treasures.

I've just watched your run (hooray for lengthy commutes to work by train!) Nice job. I think execution alone can shave a big chunk off that one, and I'd say the route optimizations you've picked are on the ball. My game plan when I had a casual go was to do the any% run, clearing the slums on the way, doing Sokal, the wealthy area and the temple, then doing the necessarily duplications and clumping together the missions by area. I don't know if sub-1 hour is possible. I'd say I was on track for about 1:10 before it went south when I realised I'd forgotten to free the lizardmen in the pyramid, so there could well be 10 minutes worth of optimizations that I can find just through execution. Of course, doing it in bits and pieces is a lot easier than doing it in one sitting.

I've got a few ideas for tightening up some areas off the bat, some major, some just minor tweaks:

Slums: One of the biggest time saves would be to use an exploit that I discovered a little while ago that’s a bit cheap, and certainly would fly in the face of your preference for doing things properly. You can actually skip most of the planned encounters. This being an eighties game, where any way to tighten coding is vital, it keeps track of the number of fights you’ve had in the area, and marks it as complete when you’ve done 24 and resolved the Ohlo mission. Since random encounters (counted separately) stop when you've done 15, that would normally force you to do the 9 set encounters, but if you try to rest in an unsafe area and you’re interrupted, you’ll have a fight which will also add to the total fight counter. Since most of the random encounters are much shorter than the set encounters and you don't have to walk to each of them, it can easily chop 4-5 minutes off, even if you return with the necklace. You can also reset the random encounter count by attacking the gypsy, but I don't recommend this because it makes the combats harder.

On a less cheap path, a four person party, suitably tweaked, will result in 6 monster random encounters, which will be nicely arranged to knock out 5 with a sleep spell and you can zip through them pretty quickly.

Sokal Keep: Target your fireball at the bottom left orc leader. That should take out enough. There are 50 monsters in this fight. Once you’ve killed 25, morale will break and they’ll surrender the next round.

Kovel Mansion: You have to kill a set number of thieves to clear the mansion (which is complicated by the fact that the number the game adds to the counter does not always correspond to the number in each encounter). When there are still thieves around, someone will try to attack you at most of the doors. Taking out one of these thieves plus the two ambushes is enough to clear the mansion. The fight with the thieves beating up the old man doesn’t count – they were sent by Tyranthraxus and aren’t connected to the guild. You can check to see if the mansion is cleared by trying to rest – you’ll be interrupted by thieves with slings if it’s not.

Nomad Camp: The quickest (and least heroic) way of completing this quest is to attend the feast, go “meh” when they ask you for help and abandon them to their fate. And the council is cool with that, rewarding you for completing the quest. Not sure how well that would sit with you, but it’s legit as far as the game’s concerned.

Kobold Caves: Just a little optimisation here: there’s a secret door at the end of the cavern where you fight the wave fight which lets you enter the king’s room through the north end. You start the battle much better positioned.

Buccaneer Camp: Fighting the entire camp is enormous fun, but if you’re doing this quickly, best way is to buy a pass from the merchant, go and see the captain and then buy the boy.

By the way, it’s pretty difficult to actually fail this mission. About the only way to do it is to surrender to the captain after fighting off most of the guards. That said, it is the only mission you can actually fail – you’ll get told off by the clerk when you return to town – unless you count Cadorna threatening you for stealing his treasure.

You also really don't need to go out of your way to get any treasures at all. The any% run, which has virtually no diversions to get non-essential treasures, gives you enough experience to take out Tyranthraxus, so especially with infinite dust and fireballs, you'll do fine in any area in the game without doing anything you don't actually have to do. Most areas can be accounted for with a minimum of combat. You can also avoid combat with virtually any encounter with intelligent creatures in the wilderness by speaking and choosing Abusive.

You also mentioned early on that you weren't sure what exactly I was doing in some places because of the speed of the inputs (and probably poor buffering in places to be fair), so when I get some time, I might do a post explaining the strats and the rationale behind certain decisions, which could also help people who have potential improvements and optimizations from what I've done.

Actually, one question about conditions for this category - is taking out Norris the Grey mandatory? It doesn't come up in the roster, but it is one that gets a "clerk speaks" reward. Also, I think the game doesn't mark the library as complete until you've cleared it too, even though they're two separate rewards.

Not that either of them are particularly lengthy interludes, but my initial experimentation with this category has suggested that sub-1 hour might (just) be possible and every time save counts! :)

Cool. I look forward to watching that. Hope you’re feeling better soon.

I’ve got a route mapped out in my head, but we’ll see how that plays out in real life. I’m not overly concerned with level up points, given that I’m planning on duping the necklace and dust and I effectively managed to account for Tyranthraxus with two 4th level fighters and a wand. One advantage of a smaller party, I might add, is that it makes most of the random encounters smaller too, and that’s no small consideration when you have plough through 15 of them. A quick experiment with my any% party took about 13 minutes from game start to clearing the slums. My inclination is to do 3 fighters and a mage and take the hits in the mansion and the kobold caves from the lack of a thief. I don’t think a cleric adds much in a speedrun.

I’m not sure I’ll have the patience or the time to refine this one (like Ozzy - small child in the house that cares not for 30 year old games), but I’m genuinely curious to give it a go and see how I manage it.