Few words before each level dive:

 - "I can win" diffuculity is the easiest and hence fastest for speedrunning. Important thing about it is that while usually player in q3 spawns with 125hp, bots can't get their health above 100hp on that difficulty, so bots spawn with 100, which allows them to be fragged by a rail or a direct/nearly direct rocket. Splash damage on rocket is rather large as well, so a rocket placed under /nearby enemy's feet will most likely deal 100 damage too.
 
 - V1.32 is version of the game that was used in this run (also, the latest version available at the time). Earlier versions of the game probably could be found somewhere on internet and may have certain benefits for speedrun purpose, for example different bot spawn mechanics, but overall consdirered not worth the trouble getting to.
 
 - Q3A speedrun is heavily dependant on RNG of enemy spawn locations as well as bot behaviour which might change based on situation around the map (weather certain item is up, or if bot dropped a weapon on his death before and so on). Behaviour usually consists of bots being attracted to one specific area/item on a map.
 
 - For "I can win" difficulty: Individual Level WR combined time at this moment is more than 6 minutes faster than current SS WR of the game on same difficulty, which tells a lot about nature of the RNG and luck in this game. Perfect execution and aim won't cut it for fastest time, but nevertheless it is important to have good movement and aim mechanics in order to beat the game quickly. "Nightmare!" difficulty can share traits and tactics of easiest difficulty run, but the fact that bots are harder to frag and they deal way more damage back - makes it quite hard to beat fast. IL for "I can win"  was chosen personally because it highlights the game mechanics & engine and pretty much leaves everything up for good RNG without the complexity of hardest difficulty
 
 - "map_restart 0" command is binded to key that is (for most of the levels) triggered on map loading screen (message/announcment of "Fight!" can be seen/heard on first second of map gameplay). Without this restart player would be forced to wait for each bot to respawn on the map in their intended order (first being player, then followed by bots in predefined order), while map_restart 0 will force all arena clients (player + bots) on the map to appear instantly. This command is also used for luck manipulation before some levels in order to get better initial spawn, which allowed for faster level completion. For those rare instances where map_restart is not used - I find slight delay in bot spawn is beneficial.
 
 - "sv_fps 333" / "com_maxfps 333" settings were used in all of those runs. Reason being is that 333 fps gives you jumps of slightly higher altitude and a bit faster movement. More on FPS value and how they affect the game can be found in this video: https://youtu.be/he02vJvKaRs
 
 - "bot_nochat 1" command was used to cut down some time of programmed "trash talk" by bots during match.
 
 - As far as I know Q3 enemy respawn system works in a way that the further you're away from the spawn location - the higher chance for it to be activated, so you'll sometimes see  me stepping away from some spawns that I want enemy to appear in. Also I think it was set to have some priority over the spawns in most distant from frag location places (meaning if you were fragged in one side of the map, chances are that you'll most likely gonna appear in the opposite side of the map, even if this spawn is visible from frag location).

very straight forward execution; map_restart 0 bind was triggered on loading screen to save time for announcer to finish introduction and bot to appear at the end of it. Bot re-spawn locations are predetermined (opposite parts of the map); First frag has to be quick enough for bot not to get Yellow Armor (YA) and has to be precise shotgun (SG) blast in order to get 100 damage on a bot (SG pellet distribution is not all random, but still takes time to get used to). Machine Gun (MG) usage is explained by lowering a chance of having successful finishing SG blast.

instance of not forcing map_restart because programmed initial spawn always will be near Rocket Launcher (RL), and by moving away from your spawn you enable it for bot to possibly appear there too (there are 2 spawns around RL) and use the time for bot to appear to get the RL and setup for a first frag and frag on respawn if you're lucky. got a bit unlucky at the end on 8th frag, where splash from rocket didnt deal 100 dmg, but good RNG on last spawns (ideally you want to get those spawns every time) made up for it. also: was very cautious about spamming rockets cause 10 that you get from start should be enough for all 10 frags (arguably I could've pick up RL on 4th frag because I was so close to its spawn, but from habit decided not to).

another instance of not forcing map_restart as spawn locations are quite distant from each other, and the delay in bot appearing works fine for you to get plazma gun (PG) and reach the enemy. First plazma shot to open the door and check the sound of bot appeared in the room with RL. Otherwise: standard gemaplay involing good RNG on bot respawn locations in the middle and bit sloppy end with SG shots.

forced respawn to get enemies on a map ASAP and always same initial spawn near RL that allows to get to the enemies quicker. Bots initial spawn location on map_restart is almost certainly going to be to the right from RL, towards where I rocket jump (RJ). Upon a rewatch: I noticed that dropped weapon magnet (bot attraction behaviour) might have took place here, where dropped SG by first bot being used to attract the bots to that isde of the map in future. Otherwise: standard gameplay and some good RNG on bots appearing

not the WR time, but still pretty good RNG. map_respawn was forced and ideally you want to have Sarge appear to your left from where you spawn (your initial spawn is always the same) or in Powerup (PU) room at the very least. The good RNG considered if bot after first frag respawns anywhere but RL room (being the most distant one). Missed couple of SG shots unfortunately.

Forced map_restart and step back to get Red Armor (RA) in order to save trouble spending more time on bot frags if anyone appeared on same spawn in the future. Standard strategy involves hoping bots to be attacted to PG on the other side of the map and taking a teleport to the center of the map, and simply camping that tele exit with SG or PG, but if bots get attracted to PU (Quad damage) room for some reason (dropped weapon for most part) - it's even better as you just need to listen for sound cues and move towards bot respawn locations. Unfortunate choice of last frag on bot with RA, but nevertheless very good RNG.

forced map_restart and luck manipulation to get most likely initial spawn on YA and 2 bots to appear near PG. Dropped PG by bots serving as a magnet for them in future, because other unfavourable behaviour involves them being attracted to RL near RA. Somehow got lucky with it cause didn't manage to PG lying after first spawns, and can be seen how bots would take stairs to RL or jumppad to SG at the beginning. Otherwise: pretty standard finish and ok RNG.

forced map_restart and luck manipulation to get the spawn around PG (but not exactly near it), when at least one bot spawns right by PG to pick it up and go towards shards. MG frag on him will result in dropped PG and magnet point for other bots to get attracted to. To keep it lying for some time and let bots get to it, I went for pillars get some frags in a meantime. Otherwise: pretty standard finish of this strat, and nearly WR time.

instance of not forcing map_restart, because the delay allows to get to YA and wait for spawn around it, and then pick up as many spawns as you can going from one side of the map to another. Once it's stopped it's better to pickup Lightning Gun (LG) and luckily meet with bot being attracted to it. For the map there are two major magnets for Hunter: one being LG another one being RL spawn.

Very good RNG all over the map. One behaviour that I consider slow is when bots are gathering around RL in closed room above RA (location not shawn in a run much). However getting them attracted to open room with rocket and leading towards rail / powerup room is rather lucky and lead to quick frags as seen during the run. Getting early Rail Gun (RG) is also good as it allows for some steady frags on off the spawn bots. PG frags still considered quicker, as shown in the end, due to rate of fire.

Forced map_restart and good initial spawn with some quick frags. Apparent attraction towards red serves a good RNG, for bots to gather from all over the map. Pretty standard execution in the middle and recognizition of PU being potential attraction, so had to go for it for quicker finish. Some missed opportunity to save couple of seconds in the end by hitting double rails and connecting with SG shot on first try.

Basic strategy involves guessing right where enemy respawned and saving some time by meeting them half-way from middle point, PG is preferrable weapon of choice, but RL is good too for some cases. Pretty standard and heavily RNG based level.

forced map_restart and few preferrable intitial spawns with player and bots appearing on same level opposite to RG. Quick RG grab and good RNG with spawns ensures quick map compeletion.

Forced map start and preferrable intitial spawn across the PG, the point of first attraction for bots on a good RNG. Dropped on death PG only ensures follow-up attraction and helps fast completion in the end. This map gives 50 armor from the spawn, so bots take longer to frag. Checked scoreboard in the middle, because wanted to make sure that bot didn't frag himself from RL splash and there wasn't -1 score to any name (otherwise would result in restart of the map / re-try).

Rather large map so quickest time possible probably very RNG based again, where standard strategy involves camping the teleport exit with SG and hoping that bots are being attracted to go through it (to pickup dropped weapon or something) on another side of the map. Forced start; standard initial spawn near LG, in hope of some bots appearing nearby. Angle on teleport exit allows to rush towards the corridor if bots appear from there.

Some attraction points on this map involving middle RL / PU room and high PG spawn. Good RNG probably considered middle RL spawn attraction, which is shown on the map. Rail initial spawn is so-so (good time can be achieved without it), but can lead to quick first frags and overall good to have for steady one shot frags if bots simply going your way, so worked out pretty fine. Misjudgement on last bot's stack, thinking that non-quad rail would finish him, costed me additional rail reload time.

not forced start as it allows to get RG and see where bot is spawned. If spawned accross the map - then only acceptable spawn would be high level, leading to SG and easy snipe with RG. Otherwise: pretty standard back and forth, hoping for good respawn location.

attraction towards red is most preferrable, and if you're able to snatch LG from some bot - and keep on high ground with RL and LG for quick pick offs - then it's pretty quick and easy map. Can go wrong if bots start to gather in some random spots like upper YA or LG.

rather large map, so heavily depends on good RNG and bot behaviour. One of the strategies that I've used involving rushing to BFG of the spawn and picking off frags on bots that are attracted to it. Following up with little back and forth between middle part of the level. Got a bit scared on low hp in the end, so decided to play it safe with rockets.

Starts off with luck manipulation on map_restart where you spawn near Battle Suit spawn in the corner and one bot spawning near RG across you. Getting RG from the start is probably a must for fastest time, so need to time your first frag a little, to not let RG drop in the lava. Then it's hopeing for bots to be attracted to middle RL point and picking up frags while they're coming.

players spawn with 50 armor, so direct rocket is not enough for a frag unfortunately. Otherwise: pretty standard frags on guessed right repsawn locations.

basic strategy: getting rail ASAP and getting lucky with bot behavior (ideally you won't be needing to wait for them to appear on your screen at all).

some good RNG in the beginning involves bots clubbing up together and you managing to snipe 3 or 4 frags w/ MG or RL before you reach RG on 9th or so second (ideally ASAP). Then it's no miss rail shots with potential double rail frags if you're lucky.

has a trick when forcing map start (map_restart) and openning the map from the menu, where you'll always have predetermined spawn, and since bot clients load faster, one bot will always share your spawn location, resulting you appearing second on a location with another bot, and "telefragging" the bot from the very start. This is not achievable on restarting the map when map is played in sequence right after DM17 completion however, so on SS WR you can see player disconnecting from DM18 from start, to force map_restart when starting the map from the menu, to get 1 point advantage from the get go. Other then that: best RNG for bots to be attracted towards RG where you camp and pick up frags from all sides.

pretty standard run where ideally you start with SG and PG frags and then in the middle pick up RG from one of the bots. Beware of some bots spawning below RG platform and picking up both YA and teleport item. Bot will use teleport if his health gets past certain lower threshold, so to not trigger that threshold I had to use MG on bot first for his health to be low enough for rail finish but not low enough to trigger his teleport usage. Not entirely sure if it was justifiable in my case, but if bot drops from that platform you must keep that in mind, because first rail will result in wasted time and bot appearing somewhere else on a map.

is straight forward with still some potential save of 1 second on quick railing and better luck with spawns.