Reach miners trapped miles under the surface of the earth! Use the Prop-pack
to maneuver through a maze of mineshafts! Blast vile vermin with the Microlaser Beam! Dynamite
walls! Negotiate across the lethal lava flow! Rescue all the miners you can before running out of
lives or power!
H.E.R.O. was ported to numerous systems, but many of the ports have unique attributes that create a unique TASing challenge. Details on what make this port unique are further down.
General Game Info
The goal of each stage is to reach the trapped miner at the bottom of each level.
Navigation is done by running left and right, hovering/flying with the prop-pack, and falling.
Lasers can be shot from your helmet to kill enemies.
Bombs must be dropped to destroy walls created from rock cave-ins to progress through the game.
Hitting any enemy is instant death
Hitting the water/lava at the bottom of a stage is instant death
Hitting a hot rock wall is instant death
Your own bombs can kill you instantly if you're too close (but there's ways around this)
TAS Objectives/Notes
Beat the Game as quickly as possible
There are 20 unique levels in the game. After that the level number changes to "PRO" and earlier levels simply repeat until the player loses all their lives.
The game manual states "You've saved the day when the score reaches 1,000,000." But this is not a true endpoint for the game as play can continue beyond that point.
At 1,000,000 points, the score display simply becomes asterisks because the RAM addresses for score are maxed and the programmers had this display change instead of rolling over to 0.
It is possible to continue scoring after this score display is achieved, thus a Max-Score run is not an option for this game.
TAS created over various BizHawk versions and completed on v2.5
To save time, unneeded bombs are dropped prior to ending a stage to limit time lost to bonus coundown.
Occasionally a minor movement delay is necessary to drop an extra bomb, but the bonus time saved at the end of the stage is worth the movement delay.
Attributes Unique to the Atari 2600 Port
Speed is slightly faster than C64 Port
Helicopter movement is a bit easier to control. Response time is faster in taking off compared to other ports
Avoiding being killed by bomb explosions:
On other versions, only 1 frame of elevation is needed to avoid being killed. In this port, it take serval frames to avoid damage. This limits use of a time saving strat more easily used in Coleco and C64 ports.
It is necessary to be completely 'stopped' to lay a bomb; no other inputs can be in the same frame as the down press to drop the bomb.
Uses "Down" to lay bomb; other ports (like Colecovision) use dedicated buttons.
Holding "Left" or "Right", against a wall displaces the main character a pixel or two closer to the wall. This helped cut time, where delays are needed in other ports to avoid creatures.
Creature speed, most notabiy the green snake, is the fastest among all the ports we explored. This makes slightly unique routing over the other consoles.
Map has the same general room layout as that of the Coleco and C64 ports, but rock walls may be in different places.
Of the three ports we worked on, Atari is the only one with bomb persistence.
Explosions from bombs droped in one room that don't explode until after the main character has entered the new room can be used to destroy walls that are on the far right of the new room
The game moves the dropped bomb's horizontal position to a location that will destroy the rock wall.
This allows for more continuous motion through some of the horizontal passageways as there is no delay from having to stop and drop a bomb near the rock wall and then wait for the bomb to explode.
This is the most significant time save of all strats compared to other ports.
How this submission came to be
NYMX was working on the C64 port of the game on his own. This prompted DrD2k9 to investigate the Coleco version. Once each of us completed our own TAS, we swapped files and reviewed each other's work. We each made improvements to the other person's TAS then swapped back again. During all the back and forth, we decided to investigate the published A2600 run. Using knowledge gained from doing the other two ports (and a bit more file swapping), we effectively did a complete reTAS of the A2600 port by testing alternative inputs room-by-room. This resulted in improvements on every level through the run. As we each worked on and made improvements to all three ports, we are both authors on all three submissions.
Comparison to Current Publication
Some of the improvements result from emulation differences between BizHawk versions; most notably at the start of the game. Others come from dropping bombs further from the rock walls; the current publication gets closer than necessary in a number of cases before dropping a bomb which then results in a later passage through the hole blown open. The rest of the improvements come from movement optimization.
This submission is 164 frames (approximately 2.74 seconds) faster than the current publication. The following table breaks down where in the game those frames are saved.
1st Visible Frame
1st Visible Frame
Frames Saved
Cumulative
for Current Publication
for this Submission
this Segment
Frames Saved
Reset to Start game
6
1
5
5
Stage 1
8
2
1
6
Stage 2
546
538
2
8
Stage 3
1249
1237
4
12
Stage 4
2192
2153
27
39
Stage 5
3372
3330
3
42
Stage 6
4489
4441
6
48
Stage 7
5842
5790
4
52
Stage 8
7624
7565
7
59
Stage 9
9491
9421
11
70
Stage 10
11531
11449
12
82
Stage 11
13671
13585
4
86
Stage 12
15976
15881
9
95
Stage 13
18196
18089
12
107
Stage 14
20513
20401
5
112
Stage 15
22661
22545
4
116
Stage 16
24962
24833
13
129
Stage 17
27042
26901
12
141
Stage 18
29210
29061
8
149
Stage 19
31443
31289
5
154
Stage 20
33647
33489
4
158
Final Input
35414
35250
6
164
The following is a video comparison of the current publication (left) vs this submission (right). Due to emulation differences between the versions of BizHawk on which the old and new TASes were made; instead of staring the comparison from power-on, this video cuts the first few frames of the old TAS to allow the first frame of character control to be synced between the runs. This allows for a more accurate representation of improvement from game-play differences.
feos: The ending check the other 2 ports went through works the same here, the difficulty isn't increasing. The game actually halts once you reach 1M score, but that's not a meaningful goal for fastest completion, because there's no new content in the PRO levels. For a max score run? Maybe. Anyway, accepting over [3351] A2600 H.E.R.O. by StarvinStruthers in 09:51.01 inheriting the tier.
Joined: 1/24/2018
Posts: 304
Location: Stafford, NY
I previously tried to improve the current submission for this game in BizHawk 2.3, but got stymied on the first "U-Turn" hover (on Level 8). For some reason, I couldn't get the player to "snap" to the ground at the end and had to manually land him instead, wiping out the previous gains I had accomplished. I gave up at that point because I assumed that the previous submission took advantage of an emulation bug (and that "proper" emulation would result in a slower submission as a result), so I would like to know if my initial conclusion was wrong or if there really WAS an emulation problem in that version...
c-square wrote:
Yes, standard runs are needed and very appreciated here too
Dylon Stejakoski wrote:
Me and the boys starting over our games of choice for the infinityieth time in a row because of just-found optimizations
^ Why I don't have any submissions despite being on the forums for years now...
I previously tried to improve the current submission for this game in BizHawk 2.3, but got stymied on the first "U-Turn" hover (on Level 8). For some reason, I couldn't get the player to "snap" to the ground at the end and had to manually land him instead, wiping out the previous gains I had accomplished. I gave up at that point because I assumed that the previous submission took advantage of an emulation bug (and that "proper" emulation would result in a slower submission as a result), so I would like to know if my initial conclusion was wrong or if there really WAS an emulation problem in that version...
I'm not sure exactly all the ways the emulation differences affect the run. What I do know for sure is only at the beginning of the run. There are 5 lag frames at the beginning in the old run. There is only 1 lag frame in the new one. This allows for a faster start in the new version.
As far as improving the old run. I copied the inputs from the old TAS and put them into the newer BizHawk version. I then went through every single screen making improvements everywhere I could (NYMX then found further improvements). So I don't even know if our improvements would carry over and work in the older versions of BizHawk.
I suppose it could be theoretically argued that all the frames we saved from minor movement optimizations could be attributed to emulation differences; but I strongly doubt that's the case.
In the comparison video, I dumped the current publication video from BizHawk 1.11.9.1 and the video for this submission from BizHawk 2.5. Even still, if you look close, it's possible to see some instances where the current publication gets closer than necessary to the rock walls before dropping bombs. So there was some optimization done beyond emulation differences.
All that said: Without seeing your inputs, I can't explain why you had trouble with the hovering "U-turn" in stage 8 on your attempts.
NYMX understands the flying mech a bit better than I do, so maybe he'd have some insight.
Joined: 11/14/2014
Posts: 932
Location: South Pole, True Land Down Under
What we discovered was a frame rule, that we believe is 8 frame in length...although, we don't understand it enough to carry a good conversation.
BUT! this frame rule was responsible for "Eating" up gains that we made. This was evident when I first found a flaw in Level 3. At first, my gains were almost all gone, by the end of the level. So I went back and tried to figure out how to clean this up. What I found was changing some of the actions to different screens. For one, the old TAS laid a bomb down near the spider, above a hole. This was inefficient. So I killed the spider first and laid the bomb in another area...to prevent having to slow down. Well, this got us past the frame rule and helped to maintain our cuts.
This is evident, at the start of each level...where we would randomly find a frame or 2 (sometimes up to 4 frames) of "No movement" where inputs were applied to take advantage of the dead waste.
When I did my review, this discovery aligned the remainder of the TAS to fall in a different "Beat", helping to maintain our cuts.
In regards to your U-turn, I suspect that your helicopter spin up was not done as late as possible to allow early forward movement and upward climbing at the optimal time to land. This is a tricky manuerver and can burn you on effort. On Coleco, we didn't have this problem as we could brush the edge while dropping a bomb....a technique that we started calling "Toe Pick". Basically, find the top of the curve for the best situation...if that makes sense.
I recently discovered that if you haven't reached a level of frustration with TASing any game, then you haven't done your due diligence.
----
SOYZA: Are you playing a game?
NYMX: I'm not playing a game, I'm TASing.
SOYZA: Oh...so its not a game...Its for real?
----
Anybody got a Quantum computer I can borrow for 20 minutes?
Nevermind...eien's 64 core machine will do. :)
----
BOTing will be the end of all games. --NYMX
This movie has been published.
The posts before this message apply to the submission, and posts after this message apply to the published movie.
----
[4304] A2600 H.E.R.O. by DrD2k9 & nymx in 09:48.27
