It had a number of interesting design ideas, but there's no denying it was a raw experience.
Problems with IV were fundamental, especially AI, the extremely linear campaign with constricted path and a megaton of scripted events. And that's ironic because the main designer wanted it to be "back to strategic roots". Potion of immortality is something not easily solved with a balance patch where you just tweak a few numbers. The 'heroes as units' system is fundamentally hard to balance in a game where heroes grow on logarithmic scale and armies on linear scale. In hindsight it might seem silly to expect hero power progression that remains distinct for hero types and scales perfectly to plus infinity, but that's more less what people expected after huge maps of Heroes3, with caverns which morphed from an extra place to put loot (as initially designed) to a whole second world. It frustrates me to this day, because there are many very interesting design ideas in there.
Of course, All the initial games (Heroes I,II, III) were very incremental in design. All creatures from H1 made it to H2, all from H2 made it to H3. It had to change, it had to be re-imagined. If for no other reason that extra permutation of stats and abilities start feeling very same-y when the system doesn't have enough depth. It's easy to spot that, for example, level7 creatures used the 250 life, 30-40ish damage and growth 3/week template with little deviation. Heroes III felt the most template'y and elsewhere I made some very good points that Heroes II had much better design, especially spell balance. The most powerful spells in Heroes III are the first level ones - stuff like haste, slow scales extremely well with army size and you get mass version through an Expert magic skill.
H3 spells are weird bunch, because 1st level ones are really really damn strong(haste, bless and slow especially, H5 fixed that by making initiative and speed two different parameters), middle is mediocre, and latter levels have few literally broken spells(town portal, dimension doors, resurrection, animate dead).
Heroes V has an issue in that initiative is exponential. Units with high initiative can act several times more often than low initiative ones. This reduces rule clarity, and no even the extra track doesn't help much. For example it's no longer simple to figure out when the unit will act when hasted or slowed.
On the flipside, HOMM1 had a blissfully stupid magic system. If you had Knowledge 6, you could cast 6 Magic Arrows, 6 Blind spells, 6 Fireballs, and 6 Armageddons. And because high level spells had no secondary skill requirements, mopping up developed enemies was a terrible slog. On the upside, it was EASY to predict how many spells you could cast. Also, shrines found outside towns vere VERY interesting and sometimes of strategic importance - because all spells would be eventually forgotten with use.
If we're talking multiplayer balance, the elephant in the room is that HOMM was always a poor multiplayer game. It just wasn't designed as such. Long, sequential turns. Time-consuming battles that magnify the waiting problem (you can auto-combat but then you're playing suboptimally and throwing out the main draw of the game). You can play it in multiplayer, but anyone who insists Heroes was a great multiplayer game hasn't played a decent modern board game. Or Age of Wonders III, designed with balance in mind. "Serious" multiplayer rulesets remove several potentially game-breaking spells from the game (Town Portal, Dimension Door, Fly especially), and encourage using auto-combat.
Heroes II on the other hand has separate Mass Haste (level 3) and Mass Slow (4).
Initiative and Speed have an interesting interaction in Heroes IV. Remember how classic HOMM has morale which is basically worse version of Luck, because quite often there's nothing more you want to do with that unit at this time? Morale in HOMM4 made units act before all other units for which Morale didn't trigger.
> The Gizmondo was further overshadowed when Swedish press revealed criminal pasts of several executives, causing their resignations including Tiger Telematics CEO Carl Freer. Director of Gizmondo Europe Stefan Eriksson was involved in a Swedish criminal organisation, the "Uppsalamaffian" (the Uppsala mafia). By February 2006, the company was forced into bankruptcy after amassing US$300 million debt, and the Gizmondo stopped production. Weeks thereafter Eriksson crashed a rare Ferrari Enzo driving at 260km/h in California, and was later jailed and subsequently deported for driving under the influence in connection with the crash and other criminal offenses.
And the ending somehow hit me in the gut hard.
Ah yes, all those technically superior jaggy polygons.